From 8d50f90060f8e53c4a5f2ddda88bed2198981938 Mon Sep 17 00:00:00 2001 From: "DESKTOP-RTLN3BA\\$punk" Date: Mon, 27 Apr 2026 14:04:50 -0700 Subject: [PATCH 01/17] chore: linting --- .../app/agents/new_chat/chat_deepagent.py | 12 +- .../agents/new_chat/middleware/__init__.py | 6 +- .../agents/new_chat/middleware/file_intent.py | 9 +- .../agents/new_chat/middleware/filesystem.py | 36 +++- .../new_chat/middleware/knowledge_search.py | 2 +- .../middleware/local_folder_backend.py | 44 +++-- .../multi_root_local_folder_backend.py | 28 +++- .../new_chat/tools/connected_accounts.py | 12 +- .../agents/new_chat/tools/discord/_auth.py | 3 +- .../new_chat/tools/discord/list_channels.py | 32 +++- .../new_chat/tools/discord/read_messages.py | 32 +++- .../new_chat/tools/discord/send_message.py | 35 +++- .../agents/new_chat/tools/gmail/read_email.py | 21 ++- .../new_chat/tools/gmail/search_emails.py | 45 +++-- .../tools/google_calendar/search_events.py | 54 ++++-- .../app/agents/new_chat/tools/hitl.py | 4 +- .../app/agents/new_chat/tools/luma/_auth.py | 3 +- .../new_chat/tools/luma/create_event.py | 21 ++- .../agents/new_chat/tools/luma/list_events.py | 37 ++-- .../agents/new_chat/tools/luma/read_event.py | 16 +- .../app/agents/new_chat/tools/mcp_client.py | 6 +- .../app/agents/new_chat/tools/mcp_tool.py | 158 +++++++++++------- .../app/agents/new_chat/tools/registry.py | 13 +- .../app/agents/new_chat/tools/teams/_auth.py | 3 +- .../new_chat/tools/teams/list_channels.py | 33 +++- .../new_chat/tools/teams/read_messages.py | 32 ++-- .../new_chat/tools/teams/send_message.py | 24 ++- .../agents/new_chat/tools/tool_response.py | 5 +- .../app/connectors/exceptions.py | 1 - surfsense_backend/app/routes/__init__.py | 4 +- .../app/routes/mcp_oauth_route.py | 130 ++++++++++---- .../app/routes/new_chat_routes.py | 2 +- .../app/routes/oauth_connector_base.py | 29 ++-- .../routes/search_source_connectors_routes.py | 4 +- .../app/services/mcp_oauth/discovery.py | 4 +- .../app/services/mcp_oauth/registry.py | 62 ++++--- .../app/services/obsidian_plugin_indexer.py | 9 +- .../app/tasks/chat/stream_new_chat.py | 41 ++--- surfsense_backend/app/utils/async_retry.py | 9 +- .../app/utils/connector_naming.py | 5 +- .../test_obsidian_plugin_routes.py | 16 +- .../middleware/test_file_intent_middleware.py | 10 +- .../test_filesystem_verification.py | 4 +- .../unit/test_obsidian_plugin_indexer.py | 7 +- .../unit/test_stream_new_chat_contract.py | 1 - .../new-chat/[[...chat_id]]/page.tsx | 23 +-- .../components/DesktopShortcutsContent.tsx | 50 +++--- surfsense_web/app/desktop/login/page.tsx | 11 +- .../assistant-ui/connector-popup.tsx | 50 +++--- .../components/mcp-connect-form.tsx | 14 +- .../components/mcp-config.tsx | 14 +- .../components/teams-config.tsx | 6 +- .../views/connector-edit-view.tsx | 12 +- .../views/indexing-configuration-view.tsx | 5 +- .../tabs/active-connectors-tab.tsx | 6 +- .../views/connector-accounts-list-view.tsx | 129 +++++++------- .../components/assistant-ui/markdown-text.tsx | 7 +- .../components/editor-panel/editor-panel.tsx | 141 +++++++++------- .../editor/plugins/fixed-toolbar-kit.tsx | 3 +- .../components/editor/source-code-editor.tsx | 2 +- .../layout/ui/right-panel/RightPanel.tsx | 8 +- .../ui/sidebar/DesktopLocalTabContent.tsx | 6 +- .../layout/ui/sidebar/DocumentsSidebar.tsx | 62 ++++--- .../ui/sidebar/LocalFilesystemBrowser.tsx | 109 ++++++------ .../layout/ui/tabs/DocumentTabContent.tsx | 4 +- .../components/new-chat/model-selector.tsx | 10 +- .../components/report-panel/report-panel.tsx | 3 +- .../settings/agent-model-manager.tsx | 10 +- .../components/settings/roles-manager.tsx | 32 +++- .../settings/user-settings-dialog.tsx | 17 +- .../tool-ui/generic-hitl-approval.tsx | 4 +- .../tool-ui/google-calendar/create-event.tsx | 9 +- surfsense_web/contracts/enums/toolIcons.tsx | 2 +- surfsense_web/types/window.d.ts | 15 +- 74 files changed, 1135 insertions(+), 693 deletions(-) diff --git a/surfsense_backend/app/agents/new_chat/chat_deepagent.py b/surfsense_backend/app/agents/new_chat/chat_deepagent.py index 73a39ccbf..ddf87cf2a 100644 --- a/surfsense_backend/app/agents/new_chat/chat_deepagent.py +++ b/surfsense_backend/app/agents/new_chat/chat_deepagent.py @@ -50,7 +50,10 @@ from app.agents.new_chat.system_prompt import ( build_configurable_system_prompt, build_surfsense_system_prompt, ) -from app.agents.new_chat.tools.registry import build_tools_async, get_connector_gated_tools +from app.agents.new_chat.tools.registry import ( + build_tools_async, + get_connector_gated_tools, +) from app.db import ChatVisibility from app.services.connector_service import ConnectorService from app.utils.perf import get_perf_logger @@ -294,9 +297,7 @@ async def create_surfsense_deep_agent( } modified_disabled_tools = list(disabled_tools) if disabled_tools else [] - modified_disabled_tools.extend( - get_connector_gated_tools(available_connectors) - ) + modified_disabled_tools.extend(get_connector_gated_tools(available_connectors)) # Remove direct KB search tool; we now pre-seed a scoped filesystem via middleware. if "search_knowledge_base" not in modified_disabled_tools: @@ -328,7 +329,8 @@ async def create_surfsense_deep_agent( meta = getattr(t, "metadata", None) or {} if meta.get("mcp_is_generic") and meta.get("mcp_connector_name"): _mcp_connector_tools.setdefault( - meta["mcp_connector_name"], [], + meta["mcp_connector_name"], + [], ).append(t.name) if _mcp_connector_tools: diff --git a/surfsense_backend/app/agents/new_chat/middleware/__init__.py b/surfsense_backend/app/agents/new_chat/middleware/__init__.py index 5a24b2f9e..6e4542e1a 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/__init__.py +++ b/surfsense_backend/app/agents/new_chat/middleware/__init__.py @@ -3,12 +3,12 @@ from app.agents.new_chat.middleware.dedup_tool_calls import ( DedupHITLToolCallsMiddleware, ) -from app.agents.new_chat.middleware.filesystem import ( - SurfSenseFilesystemMiddleware, -) from app.agents.new_chat.middleware.file_intent import ( FileIntentMiddleware, ) +from app.agents.new_chat.middleware.filesystem import ( + SurfSenseFilesystemMiddleware, +) from app.agents.new_chat.middleware.knowledge_search import ( KnowledgeBaseSearchMiddleware, ) diff --git a/surfsense_backend/app/agents/new_chat/middleware/file_intent.py b/surfsense_backend/app/agents/new_chat/middleware/file_intent.py index 4bf5dcfe4..05cb230ce 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/file_intent.py +++ b/surfsense_backend/app/agents/new_chat/middleware/file_intent.py @@ -213,7 +213,9 @@ def _build_classifier_prompt(*, recent_conversation: str, user_text: str) -> str ) -def _build_recent_conversation(messages: list[BaseMessage], *, max_messages: int = 6) -> str: +def _build_recent_conversation( + messages: list[BaseMessage], *, max_messages: int = 6 +) -> str: rows: list[str] = [] for msg in messages[-max_messages:]: role = "user" if isinstance(msg, HumanMessage) else "assistant" @@ -246,7 +248,9 @@ class FileIntentMiddleware(AgentMiddleware): # type: ignore[type-arg] [HumanMessage(content=prompt)], config={"tags": ["surfsense:internal"]}, ) - payload = json.loads(_extract_json_payload(_extract_text_from_message(response))) + payload = json.loads( + _extract_json_payload(_extract_text_from_message(response)) + ) plan = FileIntentPlan.model_validate(payload) return plan except (json.JSONDecodeError, ValidationError, ValueError) as exc: @@ -317,4 +321,3 @@ class FileIntentMiddleware(AgentMiddleware): # type: ignore[type-arg] insert_at = max(len(new_messages) - 1, 0) new_messages.insert(insert_at, contract_msg) return {"messages": new_messages, "file_operation_contract": contract} - diff --git a/surfsense_backend/app/agents/new_chat/middleware/filesystem.py b/surfsense_backend/app/agents/new_chat/middleware/filesystem.py index 8dfa89ef2..cb50693f1 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/filesystem.py +++ b/surfsense_backend/app/agents/new_chat/middleware/filesystem.py @@ -877,7 +877,9 @@ class SurfSenseFilesystemMiddleware(FilesystemMiddleware): suggested_path = contract.get("suggested_path") if isinstance(suggested_path, str) and suggested_path.strip(): normalized_suggested = self._normalize_absolute_path(suggested_path) - suggested_mount = self._extract_mount_from_path(normalized_suggested, mounts) + suggested_mount = self._extract_mount_from_path( + normalized_suggested, mounts + ) matching_mounts = [ mount @@ -1071,14 +1073,18 @@ class SurfSenseFilesystemMiddleware(FilesystemMiddleware): ] = False, ) -> Command | str: if self._filesystem_mode != FilesystemMode.DESKTOP_LOCAL_FOLDER: - return "Error: move_file is only available in desktop local-folder mode." + return ( + "Error: move_file is only available in desktop local-folder mode." + ) if not source_path.strip() or not destination_path.strip(): return "Error: source_path and destination_path are required." resolved_backend = self._get_backend(runtime) source_target = self._resolve_move_target_path(source_path, runtime) - destination_target = self._resolve_move_target_path(destination_path, runtime) + destination_target = self._resolve_move_target_path( + destination_path, runtime + ) try: validated_source = validate_path(source_target) validated_destination = validate_path(destination_target) @@ -1106,7 +1112,9 @@ class SurfSenseFilesystemMiddleware(FilesystemMiddleware): ], } ) - return f"Moved '{validated_source}' to '{res.path or validated_destination}'" + return ( + f"Moved '{validated_source}' to '{res.path or validated_destination}'" + ) async def async_move_file( source_path: Annotated[ @@ -1125,14 +1133,18 @@ class SurfSenseFilesystemMiddleware(FilesystemMiddleware): ] = False, ) -> Command | str: if self._filesystem_mode != FilesystemMode.DESKTOP_LOCAL_FOLDER: - return "Error: move_file is only available in desktop local-folder mode." + return ( + "Error: move_file is only available in desktop local-folder mode." + ) if not source_path.strip() or not destination_path.strip(): return "Error: source_path and destination_path are required." resolved_backend = self._get_backend(runtime) source_target = self._resolve_move_target_path(source_path, runtime) - destination_target = self._resolve_move_target_path(destination_path, runtime) + destination_target = self._resolve_move_target_path( + destination_path, runtime + ) try: validated_source = validate_path(source_target) validated_destination = validate_path(destination_target) @@ -1160,7 +1172,9 @@ class SurfSenseFilesystemMiddleware(FilesystemMiddleware): ], } ) - return f"Moved '{validated_source}' to '{res.path or validated_destination}'" + return ( + f"Moved '{validated_source}' to '{res.path or validated_destination}'" + ) return StructuredTool.from_function( name="move_file", @@ -1201,7 +1215,9 @@ class SurfSenseFilesystemMiddleware(FilesystemMiddleware): ] = True, ) -> str: if self._filesystem_mode != FilesystemMode.DESKTOP_LOCAL_FOLDER: - return "Error: list_tree is only available in desktop local-folder mode." + return ( + "Error: list_tree is only available in desktop local-folder mode." + ) if max_depth < 0: return "Error: max_depth must be >= 0." if page_size < 1: @@ -1253,7 +1269,9 @@ class SurfSenseFilesystemMiddleware(FilesystemMiddleware): ] = True, ) -> str: if self._filesystem_mode != FilesystemMode.DESKTOP_LOCAL_FOLDER: - return "Error: list_tree is only available in desktop local-folder mode." + return ( + "Error: list_tree is only available in desktop local-folder mode." + ) if max_depth < 0: return "Error: max_depth must be >= 0." if page_size < 1: diff --git a/surfsense_backend/app/agents/new_chat/middleware/knowledge_search.py b/surfsense_backend/app/agents/new_chat/middleware/knowledge_search.py index 51378a013..6df317aaa 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/knowledge_search.py +++ b/surfsense_backend/app/agents/new_chat/middleware/knowledge_search.py @@ -27,8 +27,8 @@ from pydantic import BaseModel, Field, ValidationError from sqlalchemy import select from sqlalchemy.ext.asyncio import AsyncSession -from app.agents.new_chat.utils import parse_date_or_datetime, resolve_date_range from app.agents.new_chat.filesystem_selection import FilesystemMode +from app.agents.new_chat.utils import parse_date_or_datetime, resolve_date_range from app.db import ( NATIVE_TO_LEGACY_DOCTYPE, Chunk, diff --git a/surfsense_backend/app/agents/new_chat/middleware/local_folder_backend.py b/surfsense_backend/app/agents/new_chat/middleware/local_folder_backend.py index 0cee3e007..565fcb48b 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/local_folder_backend.py +++ b/surfsense_backend/app/agents/new_chat/middleware/local_folder_backend.py @@ -120,7 +120,9 @@ class LocalFolderBackend: if not target.exists() or not target.is_dir(): return [] infos: list[FileInfo] = [] - for child in sorted(target.iterdir(), key=lambda p: (not p.is_dir(), p.name.lower())): + for child in sorted( + target.iterdir(), key=lambda p: (not p.is_dir(), p.name.lower()) + ): infos.append( FileInfo( path=self._to_virtual(child, self._root), @@ -317,7 +319,9 @@ class LocalFolderBackend: return WriteResult(error="Error: source and destination paths are the same") with self._acquire_path_locks(source_path, destination_path): if not source.exists(): - return WriteResult(error=f"Error: source path '{source_path}' not found") + return WriteResult( + error=f"Error: source path '{source_path}' not found" + ) if destination.exists(): if not overwrite: return WriteResult( @@ -339,8 +343,12 @@ class LocalFolderBackend: else: source.rename(destination) except OSError as exc: - return WriteResult(error=f"Error: failed to move '{source_path}': {exc}") - return WriteResult(path=self._to_virtual(destination, self._root), files_update=None) + return WriteResult( + error=f"Error: failed to move '{source_path}': {exc}" + ) + return WriteResult( + path=self._to_virtual(destination, self._root), files_update=None + ) async def amove( self, @@ -368,12 +376,16 @@ class LocalFolderBackend: if not path.exists() or not path.is_file(): return EditResult(error=f"Error: File '{file_path}' not found") content = path.read_text(encoding="utf-8", errors="replace") - result = perform_string_replacement(content, old_string, new_string, replace_all) + result = perform_string_replacement( + content, old_string, new_string, replace_all + ) if isinstance(result, str): return EditResult(error=result) updated_content, occurrences = result self._write_text_atomic(path, updated_content) - return EditResult(path=file_path, files_update=None, occurrences=int(occurrences)) + return EditResult( + path=file_path, files_update=None, occurrences=int(occurrences) + ) async def aedit( self, @@ -447,7 +459,9 @@ class LocalFolderBackend: matches: list[GrepMatch] = [] for file_path in self._iter_candidate_files(path, glob): try: - lines = file_path.read_text(encoding="utf-8", errors="replace").splitlines() + lines = file_path.read_text( + encoding="utf-8", errors="replace" + ).splitlines() except Exception: continue for idx, line in enumerate(lines, start=1): @@ -481,12 +495,18 @@ class LocalFolderBackend: FileUploadResponse(path=virtual_path, error=_FILE_NOT_FOUND) ) except IsADirectoryError: - responses.append(FileUploadResponse(path=virtual_path, error=_IS_DIRECTORY)) + responses.append( + FileUploadResponse(path=virtual_path, error=_IS_DIRECTORY) + ) except Exception: - responses.append(FileUploadResponse(path=virtual_path, error=_INVALID_PATH)) + responses.append( + FileUploadResponse(path=virtual_path, error=_INVALID_PATH) + ) return responses - async def aupload_files(self, files: list[tuple[str, bytes]]) -> list[FileUploadResponse]: + async def aupload_files( + self, files: list[tuple[str, bytes]] + ) -> list[FileUploadResponse]: return await asyncio.to_thread(self.upload_files, files) def download_files(self, paths: list[str]) -> list[FileDownloadResponse]: @@ -515,7 +535,9 @@ class LocalFolderBackend: ) except Exception: responses.append( - FileDownloadResponse(path=virtual_path, content=None, error=_INVALID_PATH) + FileDownloadResponse( + path=virtual_path, content=None, error=_INVALID_PATH + ) ) return responses diff --git a/surfsense_backend/app/agents/new_chat/middleware/multi_root_local_folder_backend.py b/surfsense_backend/app/agents/new_chat/middleware/multi_root_local_folder_backend.py index 82914f9ce..93eabe6ff 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/multi_root_local_folder_backend.py +++ b/surfsense_backend/app/agents/new_chat/middleware/multi_root_local_folder_backend.py @@ -127,7 +127,9 @@ class MultiRootLocalFolderBackend: mount, local_path = self._split_mount_path(path) except ValueError: return [] - return self._transform_infos(mount, self._mount_to_backend[mount].ls_info(local_path)) + return self._transform_infos( + mount, self._mount_to_backend[mount].ls_info(local_path) + ) async def als_info(self, path: str) -> list[FileInfo]: return await asyncio.to_thread(self.ls_info, path) @@ -355,7 +357,9 @@ class MultiRootLocalFolderBackend: all_matches.extend( [ GrepMatch( - path=self._prefix_mount_path(mount, self._get_str(match, "path")), + path=self._prefix_mount_path( + mount, self._get_str(match, "path") + ), line=self._get_int(match, "line"), text=self._get_str(match, "text"), ) @@ -394,7 +398,9 @@ class MultiRootLocalFolderBackend: try: mount, local_path = self._split_mount_path(virtual_path) except ValueError: - invalid.append(FileUploadResponse(path=virtual_path, error=_INVALID_PATH)) + invalid.append( + FileUploadResponse(path=virtual_path, error=_INVALID_PATH) + ) continue grouped.setdefault(mount, []).append((local_path, content)) @@ -404,7 +410,9 @@ class MultiRootLocalFolderBackend: responses.extend( [ FileUploadResponse( - path=self._prefix_mount_path(mount, self._get_str(item, "path")), + path=self._prefix_mount_path( + mount, self._get_str(item, "path") + ), error=self._get_str(item, "error") or None, ) for item in result @@ -412,7 +420,9 @@ class MultiRootLocalFolderBackend: ) return responses - async def aupload_files(self, files: list[tuple[str, bytes]]) -> list[FileUploadResponse]: + async def aupload_files( + self, files: list[tuple[str, bytes]] + ) -> list[FileUploadResponse]: return await asyncio.to_thread(self.upload_files, files) def download_files(self, paths: list[str]) -> list[FileDownloadResponse]: @@ -423,7 +433,9 @@ class MultiRootLocalFolderBackend: mount, local_path = self._split_mount_path(virtual_path) except ValueError: invalid.append( - FileDownloadResponse(path=virtual_path, content=None, error=_INVALID_PATH) + FileDownloadResponse( + path=virtual_path, content=None, error=_INVALID_PATH + ) ) continue grouped.setdefault(mount, []).append(local_path) @@ -434,7 +446,9 @@ class MultiRootLocalFolderBackend: responses.extend( [ FileDownloadResponse( - path=self._prefix_mount_path(mount, self._get_str(item, "path")), + path=self._prefix_mount_path( + mount, self._get_str(item, "path") + ), content=self._get_value(item, "content"), error=self._get_str(item, "error") or None, ) diff --git a/surfsense_backend/app/agents/new_chat/tools/connected_accounts.py b/surfsense_backend/app/agents/new_chat/tools/connected_accounts.py index e0b1978e1..5675a42e6 100644 --- a/surfsense_backend/app/agents/new_chat/tools/connected_accounts.py +++ b/surfsense_backend/app/agents/new_chat/tools/connected_accounts.py @@ -57,7 +57,11 @@ def create_get_connected_accounts_tool( async def _run(service: str) -> list[dict[str, Any]]: svc_cfg = MCP_SERVICES.get(service) if not svc_cfg: - return [{"error": f"Unknown service '{service}'. Valid: {', '.join(sorted(MCP_SERVICES.keys()))}"}] + return [ + { + "error": f"Unknown service '{service}'. Valid: {', '.join(sorted(MCP_SERVICES.keys()))}" + } + ] try: connector_type = SearchSourceConnectorType(svc_cfg.connector_type) @@ -74,7 +78,11 @@ def create_get_connected_accounts_tool( connectors = result.scalars().all() if not connectors: - return [{"error": f"No {svc_cfg.name} accounts connected. Ask the user to connect one in settings."}] + return [ + { + "error": f"No {svc_cfg.name} accounts connected. Ask the user to connect one in settings." + } + ] is_multi = len(connectors) > 1 diff --git a/surfsense_backend/app/agents/new_chat/tools/discord/_auth.py b/surfsense_backend/app/agents/new_chat/tools/discord/_auth.py index 1f51e3660..c345f8a5e 100644 --- a/surfsense_backend/app/agents/new_chat/tools/discord/_auth.py +++ b/surfsense_backend/app/agents/new_chat/tools/discord/_auth.py @@ -19,7 +19,8 @@ async def get_discord_connector( select(SearchSourceConnector).filter( SearchSourceConnector.search_space_id == search_space_id, SearchSourceConnector.user_id == user_id, - SearchSourceConnector.connector_type == SearchSourceConnectorType.DISCORD_CONNECTOR, + SearchSourceConnector.connector_type + == SearchSourceConnectorType.DISCORD_CONNECTOR, ) ) return result.scalars().first() diff --git a/surfsense_backend/app/agents/new_chat/tools/discord/list_channels.py b/surfsense_backend/app/agents/new_chat/tools/discord/list_channels.py index a33b88aa0..3cc99ac17 100644 --- a/surfsense_backend/app/agents/new_chat/tools/discord/list_channels.py +++ b/surfsense_backend/app/agents/new_chat/tools/discord/list_channels.py @@ -23,16 +23,24 @@ def create_list_discord_channels_tool( Dictionary with status and a list of channels (id, name). """ if db_session is None or search_space_id is None or user_id is None: - return {"status": "error", "message": "Discord tool not properly configured."} + return { + "status": "error", + "message": "Discord tool not properly configured.", + } try: - connector = await get_discord_connector(db_session, search_space_id, user_id) + connector = await get_discord_connector( + db_session, search_space_id, user_id + ) if not connector: return {"status": "error", "message": "No Discord connector found."} guild_id = get_guild_id(connector) if not guild_id: - return {"status": "error", "message": "No guild ID in Discord connector config."} + return { + "status": "error", + "message": "No guild ID in Discord connector config.", + } token = get_bot_token(connector) @@ -44,9 +52,16 @@ def create_list_discord_channels_tool( ) if resp.status_code == 401: - return {"status": "auth_error", "message": "Discord bot token is invalid.", "connector_type": "discord"} + return { + "status": "auth_error", + "message": "Discord bot token is invalid.", + "connector_type": "discord", + } if resp.status_code != 200: - return {"status": "error", "message": f"Discord API error: {resp.status_code}"} + return { + "status": "error", + "message": f"Discord API error: {resp.status_code}", + } # Type 0 = text channel channels = [ @@ -54,7 +69,12 @@ def create_list_discord_channels_tool( for ch in resp.json() if ch.get("type") == 0 ] - return {"status": "success", "guild_id": guild_id, "channels": channels, "total": len(channels)} + return { + "status": "success", + "guild_id": guild_id, + "channels": channels, + "total": len(channels), + } except Exception as e: from langgraph.errors import GraphInterrupt diff --git a/surfsense_backend/app/agents/new_chat/tools/discord/read_messages.py b/surfsense_backend/app/agents/new_chat/tools/discord/read_messages.py index 852a9297b..d8bf989a1 100644 --- a/surfsense_backend/app/agents/new_chat/tools/discord/read_messages.py +++ b/surfsense_backend/app/agents/new_chat/tools/discord/read_messages.py @@ -31,12 +31,17 @@ def create_read_discord_messages_tool( id, author, content, timestamp. """ if db_session is None or search_space_id is None or user_id is None: - return {"status": "error", "message": "Discord tool not properly configured."} + return { + "status": "error", + "message": "Discord tool not properly configured.", + } limit = min(limit, 50) try: - connector = await get_discord_connector(db_session, search_space_id, user_id) + connector = await get_discord_connector( + db_session, search_space_id, user_id + ) if not connector: return {"status": "error", "message": "No Discord connector found."} @@ -51,11 +56,21 @@ def create_read_discord_messages_tool( ) if resp.status_code == 401: - return {"status": "auth_error", "message": "Discord bot token is invalid.", "connector_type": "discord"} + return { + "status": "auth_error", + "message": "Discord bot token is invalid.", + "connector_type": "discord", + } if resp.status_code == 403: - return {"status": "error", "message": "Bot lacks permission to read this channel."} + return { + "status": "error", + "message": "Bot lacks permission to read this channel.", + } if resp.status_code != 200: - return {"status": "error", "message": f"Discord API error: {resp.status_code}"} + return { + "status": "error", + "message": f"Discord API error: {resp.status_code}", + } messages = [ { @@ -67,7 +82,12 @@ def create_read_discord_messages_tool( for m in resp.json() ] - return {"status": "success", "channel_id": channel_id, "messages": messages, "total": len(messages)} + return { + "status": "success", + "channel_id": channel_id, + "messages": messages, + "total": len(messages), + } except Exception as e: from langgraph.errors import GraphInterrupt diff --git a/surfsense_backend/app/agents/new_chat/tools/discord/send_message.py b/surfsense_backend/app/agents/new_chat/tools/discord/send_message.py index be4e6fdb2..236cd017a 100644 --- a/surfsense_backend/app/agents/new_chat/tools/discord/send_message.py +++ b/surfsense_backend/app/agents/new_chat/tools/discord/send_message.py @@ -35,13 +35,21 @@ def create_send_discord_message_tool( - If status is "rejected", the user explicitly declined. Do NOT retry. """ if db_session is None or search_space_id is None or user_id is None: - return {"status": "error", "message": "Discord tool not properly configured."} + return { + "status": "error", + "message": "Discord tool not properly configured.", + } if len(content) > 2000: - return {"status": "error", "message": "Message exceeds Discord's 2000-character limit."} + return { + "status": "error", + "message": "Message exceeds Discord's 2000-character limit.", + } try: - connector = await get_discord_connector(db_session, search_space_id, user_id) + connector = await get_discord_connector( + db_session, search_space_id, user_id + ) if not connector: return {"status": "error", "message": "No Discord connector found."} @@ -53,7 +61,10 @@ def create_send_discord_message_tool( ) if result.rejected: - return {"status": "rejected", "message": "User declined. Message was not sent."} + return { + "status": "rejected", + "message": "User declined. Message was not sent.", + } final_content = result.params.get("content", content) final_channel = result.params.get("channel_id", channel_id) @@ -72,11 +83,21 @@ def create_send_discord_message_tool( ) if resp.status_code == 401: - return {"status": "auth_error", "message": "Discord bot token is invalid.", "connector_type": "discord"} + return { + "status": "auth_error", + "message": "Discord bot token is invalid.", + "connector_type": "discord", + } if resp.status_code == 403: - return {"status": "error", "message": "Bot lacks permission to send messages in this channel."} + return { + "status": "error", + "message": "Bot lacks permission to send messages in this channel.", + } if resp.status_code not in (200, 201): - return {"status": "error", "message": f"Discord API error: {resp.status_code}"} + return { + "status": "error", + "message": f"Discord API error: {resp.status_code}", + } msg_data = resp.json() return { diff --git a/surfsense_backend/app/agents/new_chat/tools/gmail/read_email.py b/surfsense_backend/app/agents/new_chat/tools/gmail/read_email.py index 9071f129a..deec1627c 100644 --- a/surfsense_backend/app/agents/new_chat/tools/gmail/read_email.py +++ b/surfsense_backend/app/agents/new_chat/tools/gmail/read_email.py @@ -65,12 +65,22 @@ def create_read_gmail_email_tool( detail, error = await gmail.get_message_details(message_id) if error: - if "re-authenticate" in error.lower() or "authentication failed" in error.lower(): - return {"status": "auth_error", "message": error, "connector_type": "gmail"} + if ( + "re-authenticate" in error.lower() + or "authentication failed" in error.lower() + ): + return { + "status": "auth_error", + "message": error, + "connector_type": "gmail", + } return {"status": "error", "message": error} if not detail: - return {"status": "not_found", "message": f"Email with ID '{message_id}' not found."} + return { + "status": "not_found", + "message": f"Email with ID '{message_id}' not found.", + } content = gmail.format_message_to_markdown(detail) @@ -82,6 +92,9 @@ def create_read_gmail_email_tool( if isinstance(e, GraphInterrupt): raise logger.error("Error reading Gmail email: %s", e, exc_info=True) - return {"status": "error", "message": "Failed to read email. Please try again."} + return { + "status": "error", + "message": "Failed to read email. Please try again.", + } return read_gmail_email diff --git a/surfsense_backend/app/agents/new_chat/tools/gmail/search_emails.py b/surfsense_backend/app/agents/new_chat/tools/gmail/search_emails.py index de43f03d0..2e363609e 100644 --- a/surfsense_backend/app/agents/new_chat/tools/gmail/search_emails.py +++ b/surfsense_backend/app/agents/new_chat/tools/gmail/search_emails.py @@ -125,12 +125,24 @@ def create_search_gmail_tool( max_results=max_results, query=query ) if error: - if "re-authenticate" in error.lower() or "authentication failed" in error.lower(): - return {"status": "auth_error", "message": error, "connector_type": "gmail"} + if ( + "re-authenticate" in error.lower() + or "authentication failed" in error.lower() + ): + return { + "status": "auth_error", + "message": error, + "connector_type": "gmail", + } return {"status": "error", "message": error} if not messages_list: - return {"status": "success", "emails": [], "total": 0, "message": "No emails found."} + return { + "status": "success", + "emails": [], + "total": 0, + "message": "No emails found.", + } emails = [] for msg in messages_list: @@ -141,16 +153,18 @@ def create_search_gmail_tool( h["name"].lower(): h["value"] for h in detail.get("payload", {}).get("headers", []) } - emails.append({ - "message_id": detail.get("id"), - "thread_id": detail.get("threadId"), - "subject": headers.get("subject", "No Subject"), - "from": headers.get("from", "Unknown"), - "to": headers.get("to", ""), - "date": headers.get("date", ""), - "snippet": detail.get("snippet", ""), - "labels": detail.get("labelIds", []), - }) + emails.append( + { + "message_id": detail.get("id"), + "thread_id": detail.get("threadId"), + "subject": headers.get("subject", "No Subject"), + "from": headers.get("from", "Unknown"), + "to": headers.get("to", ""), + "date": headers.get("date", ""), + "snippet": detail.get("snippet", ""), + "labels": detail.get("labelIds", []), + } + ) return {"status": "success", "emails": emails, "total": len(emails)} @@ -160,6 +174,9 @@ def create_search_gmail_tool( if isinstance(e, GraphInterrupt): raise logger.error("Error searching Gmail: %s", e, exc_info=True) - return {"status": "error", "message": "Failed to search Gmail. Please try again."} + return { + "status": "error", + "message": "Failed to search Gmail. Please try again.", + } return search_gmail diff --git a/surfsense_backend/app/agents/new_chat/tools/google_calendar/search_events.py b/surfsense_backend/app/agents/new_chat/tools/google_calendar/search_events.py index a622b0efa..dc6adb822 100644 --- a/surfsense_backend/app/agents/new_chat/tools/google_calendar/search_events.py +++ b/surfsense_backend/app/agents/new_chat/tools/google_calendar/search_events.py @@ -39,7 +39,10 @@ def create_search_calendar_events_tool( event_id, summary, start, end, location, attendees. """ if db_session is None or search_space_id is None or user_id is None: - return {"status": "error", "message": "Calendar tool not properly configured."} + return { + "status": "error", + "message": "Calendar tool not properly configured.", + } max_results = min(max_results, 50) @@ -76,10 +79,22 @@ def create_search_calendar_events_tool( ) if error: - if "re-authenticate" in error.lower() or "authentication failed" in error.lower(): - return {"status": "auth_error", "message": error, "connector_type": "google_calendar"} + if ( + "re-authenticate" in error.lower() + or "authentication failed" in error.lower() + ): + return { + "status": "auth_error", + "message": error, + "connector_type": "google_calendar", + } if "no events found" in error.lower(): - return {"status": "success", "events": [], "total": 0, "message": error} + return { + "status": "success", + "events": [], + "total": 0, + "message": error, + } return {"status": "error", "message": error} events = [] @@ -87,19 +102,19 @@ def create_search_calendar_events_tool( start = ev.get("start", {}) end = ev.get("end", {}) attendees_raw = ev.get("attendees", []) - events.append({ - "event_id": ev.get("id"), - "summary": ev.get("summary", "No Title"), - "start": start.get("dateTime") or start.get("date", ""), - "end": end.get("dateTime") or end.get("date", ""), - "location": ev.get("location", ""), - "description": ev.get("description", ""), - "html_link": ev.get("htmlLink", ""), - "attendees": [ - a.get("email", "") for a in attendees_raw[:10] - ], - "status": ev.get("status", ""), - }) + events.append( + { + "event_id": ev.get("id"), + "summary": ev.get("summary", "No Title"), + "start": start.get("dateTime") or start.get("date", ""), + "end": end.get("dateTime") or end.get("date", ""), + "location": ev.get("location", ""), + "description": ev.get("description", ""), + "html_link": ev.get("htmlLink", ""), + "attendees": [a.get("email", "") for a in attendees_raw[:10]], + "status": ev.get("status", ""), + } + ) return {"status": "success", "events": events, "total": len(events)} @@ -109,6 +124,9 @@ def create_search_calendar_events_tool( if isinstance(e, GraphInterrupt): raise logger.error("Error searching calendar events: %s", e, exc_info=True) - return {"status": "error", "message": "Failed to search calendar events. Please try again."} + return { + "status": "error", + "message": "Failed to search calendar events. Please try again.", + } return search_calendar_events diff --git a/surfsense_backend/app/agents/new_chat/tools/hitl.py b/surfsense_backend/app/agents/new_chat/tools/hitl.py index 89f02abf6..8480e57b1 100644 --- a/surfsense_backend/app/agents/new_chat/tools/hitl.py +++ b/surfsense_backend/app/agents/new_chat/tools/hitl.py @@ -130,7 +130,9 @@ def request_approval( try: decision_type, edited_params = _parse_decision(approval) except ValueError: - logger.warning("No approval decision received for %s — rejecting for safety", tool_name) + logger.warning( + "No approval decision received for %s — rejecting for safety", tool_name + ) return HITLResult(rejected=True, decision_type="error", params=params) logger.info("User decision for %s: %s", tool_name, decision_type) diff --git a/surfsense_backend/app/agents/new_chat/tools/luma/_auth.py b/surfsense_backend/app/agents/new_chat/tools/luma/_auth.py index 1d88161d6..37deb1525 100644 --- a/surfsense_backend/app/agents/new_chat/tools/luma/_auth.py +++ b/surfsense_backend/app/agents/new_chat/tools/luma/_auth.py @@ -17,7 +17,8 @@ async def get_luma_connector( select(SearchSourceConnector).filter( SearchSourceConnector.search_space_id == search_space_id, SearchSourceConnector.user_id == user_id, - SearchSourceConnector.connector_type == SearchSourceConnectorType.LUMA_CONNECTOR, + SearchSourceConnector.connector_type + == SearchSourceConnectorType.LUMA_CONNECTOR, ) ) return result.scalars().first() diff --git a/surfsense_backend/app/agents/new_chat/tools/luma/create_event.py b/surfsense_backend/app/agents/new_chat/tools/luma/create_event.py index 2217d29e6..0a24a988f 100644 --- a/surfsense_backend/app/agents/new_chat/tools/luma/create_event.py +++ b/surfsense_backend/app/agents/new_chat/tools/luma/create_event.py @@ -62,7 +62,10 @@ def create_create_luma_event_tool( ) if result.rejected: - return {"status": "rejected", "message": "User declined. Event was not created."} + return { + "status": "rejected", + "message": "User declined. Event was not created.", + } final_name = result.params.get("name", name) final_start = result.params.get("start_at", start_at) @@ -90,11 +93,21 @@ def create_create_luma_event_tool( ) if resp.status_code == 401: - return {"status": "auth_error", "message": "Luma API key is invalid.", "connector_type": "luma"} + return { + "status": "auth_error", + "message": "Luma API key is invalid.", + "connector_type": "luma", + } if resp.status_code == 403: - return {"status": "error", "message": "Luma Plus subscription required to create events via API."} + return { + "status": "error", + "message": "Luma Plus subscription required to create events via API.", + } if resp.status_code not in (200, 201): - return {"status": "error", "message": f"Luma API error: {resp.status_code} — {resp.text[:200]}"} + return { + "status": "error", + "message": f"Luma API error: {resp.status_code} — {resp.text[:200]}", + } data = resp.json() event_id = data.get("api_id") or data.get("event", {}).get("api_id") diff --git a/surfsense_backend/app/agents/new_chat/tools/luma/list_events.py b/surfsense_backend/app/agents/new_chat/tools/luma/list_events.py index cd4721758..aec5ad220 100644 --- a/surfsense_backend/app/agents/new_chat/tools/luma/list_events.py +++ b/surfsense_backend/app/agents/new_chat/tools/luma/list_events.py @@ -46,7 +46,9 @@ def create_list_luma_events_tool( async with httpx.AsyncClient(timeout=20.0) as client: while len(all_entries) < max_results: - params: dict[str, Any] = {"limit": min(100, max_results - len(all_entries))} + params: dict[str, Any] = { + "limit": min(100, max_results - len(all_entries)) + } if cursor: params["cursor"] = cursor @@ -57,9 +59,16 @@ def create_list_luma_events_tool( ) if resp.status_code == 401: - return {"status": "auth_error", "message": "Luma API key is invalid.", "connector_type": "luma"} + return { + "status": "auth_error", + "message": "Luma API key is invalid.", + "connector_type": "luma", + } if resp.status_code != 200: - return {"status": "error", "message": f"Luma API error: {resp.status_code}"} + return { + "status": "error", + "message": f"Luma API error: {resp.status_code}", + } data = resp.json() entries = data.get("entries", []) @@ -76,16 +85,18 @@ def create_list_luma_events_tool( for entry in all_entries[:max_results]: ev = entry.get("event", {}) geo = ev.get("geo_info", {}) - events.append({ - "event_id": entry.get("api_id"), - "name": ev.get("name", "Untitled"), - "start_at": ev.get("start_at", ""), - "end_at": ev.get("end_at", ""), - "timezone": ev.get("timezone", ""), - "location": geo.get("name", ""), - "url": ev.get("url", ""), - "visibility": ev.get("visibility", ""), - }) + events.append( + { + "event_id": entry.get("api_id"), + "name": ev.get("name", "Untitled"), + "start_at": ev.get("start_at", ""), + "end_at": ev.get("end_at", ""), + "timezone": ev.get("timezone", ""), + "location": geo.get("name", ""), + "url": ev.get("url", ""), + "visibility": ev.get("visibility", ""), + } + ) return {"status": "success", "events": events, "total": len(events)} diff --git a/surfsense_backend/app/agents/new_chat/tools/luma/read_event.py b/surfsense_backend/app/agents/new_chat/tools/luma/read_event.py index eb3ac55c6..b37a9d617 100644 --- a/surfsense_backend/app/agents/new_chat/tools/luma/read_event.py +++ b/surfsense_backend/app/agents/new_chat/tools/luma/read_event.py @@ -44,11 +44,21 @@ def create_read_luma_event_tool( ) if resp.status_code == 401: - return {"status": "auth_error", "message": "Luma API key is invalid.", "connector_type": "luma"} + return { + "status": "auth_error", + "message": "Luma API key is invalid.", + "connector_type": "luma", + } if resp.status_code == 404: - return {"status": "not_found", "message": f"Event '{event_id}' not found."} + return { + "status": "not_found", + "message": f"Event '{event_id}' not found.", + } if resp.status_code != 200: - return {"status": "error", "message": f"Luma API error: {resp.status_code}"} + return { + "status": "error", + "message": f"Luma API error: {resp.status_code}", + } data = resp.json() ev = data.get("event", data) diff --git a/surfsense_backend/app/agents/new_chat/tools/mcp_client.py b/surfsense_backend/app/agents/new_chat/tools/mcp_client.py index b46ddbcc5..e28ac8bda 100644 --- a/surfsense_backend/app/agents/new_chat/tools/mcp_client.py +++ b/surfsense_backend/app/agents/new_chat/tools/mcp_client.py @@ -220,10 +220,8 @@ class MCPClient: logger.info("MCP tool '%s' succeeded: %s", tool_name, result_str[:200]) return result_str - except asyncio.TimeoutError: - logger.error( - "MCP tool '%s' timed out after %.0fs", tool_name, timeout - ) + except TimeoutError: + logger.error("MCP tool '%s' timed out after %.0fs", tool_name, timeout) return f"Error: MCP tool '{tool_name}' timed out after {timeout:.0f}s" except RuntimeError as e: if "Invalid structured content" in str(e): diff --git a/surfsense_backend/app/agents/new_chat/tools/mcp_tool.py b/surfsense_backend/app/agents/new_chat/tools/mcp_tool.py index dfee24516..5b96ab374 100644 --- a/surfsense_backend/app/agents/new_chat/tools/mcp_tool.py +++ b/surfsense_backend/app/agents/new_chat/tools/mcp_tool.py @@ -35,7 +35,7 @@ from sqlalchemy.ext.asyncio import AsyncSession from app.agents.new_chat.tools.hitl import request_approval from app.agents.new_chat.tools.mcp_client import MCPClient -from app.db import SearchSourceConnector, SearchSourceConnectorType +from app.db import SearchSourceConnector from app.services.mcp_oauth.registry import MCP_SERVICES, get_service_by_connector_type logger = logging.getLogger(__name__) @@ -105,13 +105,15 @@ def _create_dynamic_input_model_from_schema( description=( "Arguments to pass to this tool as a JSON object. " "Infer sensible key names from the tool name and description " - "(e.g. {\"search\": \"my query\"} for a search tool)." + '(e.g. {"search": "my query"} for a search tool).' ), ), ) model_name = f"{tool_name.replace(' ', '').replace('-', '_')}Input" - model = create_model(model_name, __config__=ConfigDict(extra="allow"), **field_definitions) + model = create_model( + model_name, __config__=ConfigDict(extra="allow"), **field_definitions + ) return model @@ -187,16 +189,23 @@ async def _create_mcp_tool_from_definition_stdio( except Exception as e: last_error = e if attempt < _TOOL_CALL_MAX_RETRIES - 1: - delay = _TOOL_CALL_RETRY_DELAY * (2 ** attempt) + delay = _TOOL_CALL_RETRY_DELAY * (2**attempt) logger.warning( "MCP tool '%s' failed (attempt %d/%d): %s. Retrying in %.1fs...", - tool_name, attempt + 1, _TOOL_CALL_MAX_RETRIES, e, delay, + tool_name, + attempt + 1, + _TOOL_CALL_MAX_RETRIES, + e, + delay, ) await asyncio.sleep(delay) else: logger.error( "MCP tool '%s' failed after %d attempts: %s", - tool_name, _TOOL_CALL_MAX_RETRIES, e, exc_info=True, + tool_name, + _TOOL_CALL_MAX_RETRIES, + e, + exc_info=True, ) return f"Error: MCP tool '{tool_name}' failed after {_TOOL_CALL_MAX_RETRIES} attempts: {last_error!s}" @@ -318,17 +327,22 @@ async def _create_mcp_tool_from_definition_http( try: result_str = await _do_mcp_call(headers, call_kwargs) - logger.debug("MCP HTTP tool '%s' succeeded (len=%d)", exposed_name, len(result_str)) + logger.debug( + "MCP HTTP tool '%s' succeeded (len=%d)", exposed_name, len(result_str) + ) return result_str except Exception as first_err: if not _is_auth_error(first_err) or connector_id is None: - logger.exception("MCP HTTP tool '%s' execution failed: %s", exposed_name, first_err) + logger.exception( + "MCP HTTP tool '%s' execution failed: %s", exposed_name, first_err + ) return f"Error: MCP HTTP tool '{exposed_name}' execution failed: {first_err!s}" logger.warning( "MCP HTTP tool '%s' got 401 — attempting token refresh for connector %s", - exposed_name, connector_id, + exposed_name, + connector_id, ) fresh_headers = await _force_refresh_and_get_headers(connector_id) if fresh_headers is None: @@ -348,7 +362,8 @@ async def _create_mcp_tool_from_definition_http( except Exception as retry_err: logger.exception( "MCP HTTP tool '%s' still failing after token refresh: %s", - exposed_name, retry_err, + exposed_name, + retry_err, ) if _is_auth_error(retry_err): await _mark_connector_auth_expired(connector_id) @@ -393,7 +408,8 @@ async def _load_stdio_mcp_tools( if not command or not isinstance(command, str): logger.warning( "MCP connector %d (name: '%s') missing or invalid command field, skipping", - connector_id, connector_name, + connector_id, + connector_name, ) return tools @@ -401,7 +417,8 @@ async def _load_stdio_mcp_tools( if not isinstance(args, list): logger.warning( "MCP connector %d (name: '%s') has invalid args field (must be list), skipping", - connector_id, connector_name, + connector_id, + connector_name, ) return tools @@ -409,7 +426,8 @@ async def _load_stdio_mcp_tools( if not isinstance(env, dict): logger.warning( "MCP connector %d (name: '%s') has invalid env field (must be dict), skipping", - connector_id, connector_name, + connector_id, + connector_name, ) return tools @@ -420,7 +438,9 @@ async def _load_stdio_mcp_tools( logger.info( "Discovered %d tools from stdio MCP server '%s' (connector %d)", - len(tool_definitions), command, connector_id, + len(tool_definitions), + command, + connector_id, ) for tool_def in tool_definitions: @@ -436,7 +456,9 @@ async def _load_stdio_mcp_tools( except Exception as e: logger.exception( "Failed to create tool '%s' from connector %d: %s", - tool_def.get("name"), connector_id, e, + tool_def.get("name"), + connector_id, + e, ) return tools @@ -468,7 +490,8 @@ async def _load_http_mcp_tools( if not url or not isinstance(url, str): logger.warning( "MCP connector %d (name: '%s') missing or invalid url field, skipping", - connector_id, connector_name, + connector_id, + connector_name, ) return tools @@ -476,7 +499,8 @@ async def _load_http_mcp_tools( if not isinstance(headers, dict): logger.warning( "MCP connector %d (name: '%s') has invalid headers field (must be dict), skipping", - connector_id, connector_name, + connector_id, + connector_name, ) return tools @@ -507,7 +531,9 @@ async def _load_http_mcp_tools( if not _is_auth_error(first_err) or connector_id is None: logger.exception( "Failed to connect to HTTP MCP server at '%s' (connector %d): %s", - url, connector_id, first_err, + url, + connector_id, + first_err, ) return tools @@ -534,7 +560,8 @@ async def _load_http_mcp_tools( except Exception as retry_err: logger.exception( "HTTP MCP discovery for connector %d still failing after refresh: %s", - connector_id, retry_err, + connector_id, + retry_err, ) if _is_auth_error(retry_err): await _mark_connector_auth_expired(connector_id) @@ -543,17 +570,20 @@ async def _load_http_mcp_tools( total_discovered = len(tool_definitions) if allowed_set: - tool_definitions = [ - td for td in tool_definitions if td["name"] in allowed_set - ] + tool_definitions = [td for td in tool_definitions if td["name"] in allowed_set] logger.info( "HTTP MCP server '%s' (connector %d): %d/%d tools after allowlist filter", - url, connector_id, len(tool_definitions), total_discovered, + url, + connector_id, + len(tool_definitions), + total_discovered, ) else: logger.info( "Discovered %d tools from HTTP MCP server '%s' (connector %d) — no allowlist, loading all", - total_discovered, url, connector_id, + total_discovered, + url, + connector_id, ) for tool_def in tool_definitions: @@ -573,7 +603,9 @@ async def _load_http_mcp_tools( except Exception as e: logger.exception( "Failed to create HTTP tool '%s' from connector %d: %s", - tool_def.get("name"), connector_id, e, + tool_def.get("name"), + connector_id, + e, ) return tools @@ -628,7 +660,7 @@ def _inject_oauth_headers( async def _refresh_connector_token( session: AsyncSession, - connector: "SearchSourceConnector", + connector: SearchSourceConnector, ) -> str | None: """Refresh the OAuth token for an MCP connector and persist the result. @@ -692,12 +724,8 @@ async def _refresh_connector_token( updated_oauth = dict(mcp_oauth) updated_oauth["access_token"] = enc.encrypt_token(new_access) if token_json.get("refresh_token"): - updated_oauth["refresh_token"] = enc.encrypt_token( - token_json["refresh_token"] - ) - updated_oauth["expires_at"] = ( - new_expires_at.isoformat() if new_expires_at else None - ) + updated_oauth["refresh_token"] = enc.encrypt_token(token_json["refresh_token"]) + updated_oauth["expires_at"] = new_expires_at.isoformat() if new_expires_at else None updated_cfg = {**cfg, "mcp_oauth": updated_oauth} updated_cfg.pop("auth_expired", None) @@ -713,7 +741,7 @@ async def _refresh_connector_token( async def _maybe_refresh_mcp_oauth_token( session: AsyncSession, - connector: "SearchSourceConnector", + connector: SearchSourceConnector, cfg: dict[str, Any], server_config: dict[str, Any], ) -> dict[str, Any]: @@ -731,10 +759,11 @@ async def _maybe_refresh_mcp_oauth_token( try: expires_at = datetime.fromisoformat(expires_at_str) if expires_at.tzinfo is None: - from datetime import timezone - expires_at = expires_at.replace(tzinfo=timezone.utc) + expires_at = expires_at.replace(tzinfo=UTC) - if datetime.now(UTC) < expires_at - timedelta(seconds=_TOKEN_REFRESH_BUFFER_SECONDS): + if datetime.now(UTC) < expires_at - timedelta( + seconds=_TOKEN_REFRESH_BUFFER_SECONDS + ): return server_config except (ValueError, TypeError): return server_config @@ -744,7 +773,9 @@ async def _maybe_refresh_mcp_oauth_token( if not new_access: return server_config - logger.info("Proactively refreshed MCP OAuth token for connector %s", connector.id) + logger.info( + "Proactively refreshed MCP OAuth token for connector %s", connector.id + ) refreshed_config = dict(server_config) refreshed_config["headers"] = { @@ -920,7 +951,7 @@ async def load_mcp_tools( result = await session.execute( select(SearchSourceConnector).filter( SearchSourceConnector.search_space_id == search_space_id, - cast(SearchSourceConnector.config, JSONB).has_key("server_config"), # noqa: W601 + cast(SearchSourceConnector.config, JSONB).has_key("server_config"), ), ) @@ -956,13 +987,17 @@ async def load_mcp_tools( if not server_config or not isinstance(server_config, dict): logger.warning( "MCP connector %d (name: '%s') has invalid or missing server_config, skipping", - connector.id, connector.name, + connector.id, + connector.name, ) continue if cfg.get("mcp_oauth"): server_config = await _maybe_refresh_mcp_oauth_token( - session, connector, cfg, server_config, + session, + connector, + cfg, + server_config, ) cfg = connector.config or {} server_config = _inject_oauth_headers(cfg, server_config) @@ -995,22 +1030,25 @@ async def load_mcp_tools( if service_key: tool_name_prefix = f"{service_key}_{connector.id}" - discovery_tasks.append({ - "connector_id": connector.id, - "connector_name": connector.name, - "server_config": server_config, - "trusted_tools": trusted_tools, - "allowed_tools": allowed_tools, - "readonly_tools": readonly_tools, - "tool_name_prefix": tool_name_prefix, - "transport": server_config.get("transport", "stdio"), - "is_generic_mcp": svc_cfg is None, - }) + discovery_tasks.append( + { + "connector_id": connector.id, + "connector_name": connector.name, + "server_config": server_config, + "trusted_tools": trusted_tools, + "allowed_tools": allowed_tools, + "readonly_tools": readonly_tools, + "tool_name_prefix": tool_name_prefix, + "transport": server_config.get("transport", "stdio"), + "is_generic_mcp": svc_cfg is None, + } + ) except Exception as e: logger.exception( "Failed to prepare MCP connector %d: %s", - connector.id, e, + connector.id, + e, ) async def _discover_one(task: dict[str, Any]) -> list[StructuredTool]: @@ -1039,23 +1077,23 @@ async def load_mcp_tools( ), timeout=_MCP_DISCOVERY_TIMEOUT_SECONDS, ) - except asyncio.TimeoutError: + except TimeoutError: logger.error( "MCP connector %d timed out after %ds during discovery", - task["connector_id"], _MCP_DISCOVERY_TIMEOUT_SECONDS, + task["connector_id"], + _MCP_DISCOVERY_TIMEOUT_SECONDS, ) return [] except Exception as e: logger.exception( "Failed to load tools from MCP connector %d: %s", - task["connector_id"], e, + task["connector_id"], + e, ) return [] results = await asyncio.gather(*[_discover_one(t) for t in discovery_tasks]) - tools: list[StructuredTool] = [ - tool for sublist in results for tool in sublist - ] + tools: list[StructuredTool] = [tool for sublist in results for tool in sublist] _mcp_tools_cache[search_space_id] = (now, tools) @@ -1063,7 +1101,9 @@ async def load_mcp_tools( oldest_key = min(_mcp_tools_cache, key=lambda k: _mcp_tools_cache[k][0]) del _mcp_tools_cache[oldest_key] - logger.info("Loaded %d MCP tools for search space %d", len(tools), search_space_id) + logger.info( + "Loaded %d MCP tools for search space %d", len(tools), search_space_id + ) return tools except Exception as e: diff --git a/surfsense_backend/app/agents/new_chat/tools/registry.py b/surfsense_backend/app/agents/new_chat/tools/registry.py index 85c89b114..3ac8677b9 100644 --- a/surfsense_backend/app/agents/new_chat/tools/registry.py +++ b/surfsense_backend/app/agents/new_chat/tools/registry.py @@ -50,6 +50,7 @@ from .confluence import ( create_delete_confluence_page_tool, create_update_confluence_page_tool, ) +from .connected_accounts import create_get_connected_accounts_tool from .discord import ( create_list_discord_channels_tool, create_read_discord_messages_tool, @@ -78,7 +79,6 @@ from .google_drive import ( create_create_google_drive_file_tool, create_delete_google_drive_file_tool, ) -from .connected_accounts import create_get_connected_accounts_tool from .luma import ( create_create_luma_event_tool, create_list_luma_events_tool, @@ -675,10 +675,7 @@ def get_connector_gated_tools( available_connectors: list[str] | None, ) -> list[str]: """Return tool names to disable""" - if available_connectors is None: - available = set() - else: - available = set(available_connectors) + available = set() if available_connectors is None else set(available_connectors) disabled: list[str] = [] for tool_def in BUILTIN_TOOLS: @@ -829,14 +826,16 @@ async def build_tools_async( tools.extend(mcp_tools) logging.info( "Registered %d MCP tools: %s", - len(mcp_tools), [t.name for t in mcp_tools], + len(mcp_tools), + [t.name for t in mcp_tools], ) except Exception as e: logging.exception("Failed to load MCP tools: %s", e) logging.info( "Total tools for agent: %d — %s", - len(tools), [t.name for t in tools], + len(tools), + [t.name for t in tools], ) return tools diff --git a/surfsense_backend/app/agents/new_chat/tools/teams/_auth.py b/surfsense_backend/app/agents/new_chat/tools/teams/_auth.py index f24f5502e..4345bb476 100644 --- a/surfsense_backend/app/agents/new_chat/tools/teams/_auth.py +++ b/surfsense_backend/app/agents/new_chat/tools/teams/_auth.py @@ -17,7 +17,8 @@ async def get_teams_connector( select(SearchSourceConnector).filter( SearchSourceConnector.search_space_id == search_space_id, SearchSourceConnector.user_id == user_id, - SearchSourceConnector.connector_type == SearchSourceConnectorType.TEAMS_CONNECTOR, + SearchSourceConnector.connector_type + == SearchSourceConnectorType.TEAMS_CONNECTOR, ) ) return result.scalars().first() diff --git a/surfsense_backend/app/agents/new_chat/tools/teams/list_channels.py b/surfsense_backend/app/agents/new_chat/tools/teams/list_channels.py index a676595c1..d7b000853 100644 --- a/surfsense_backend/app/agents/new_chat/tools/teams/list_channels.py +++ b/surfsense_backend/app/agents/new_chat/tools/teams/list_channels.py @@ -35,12 +35,21 @@ def create_list_teams_channels_tool( headers = {"Authorization": f"Bearer {token}"} async with httpx.AsyncClient(timeout=20.0) as client: - teams_resp = await client.get(f"{GRAPH_API}/me/joinedTeams", headers=headers) + teams_resp = await client.get( + f"{GRAPH_API}/me/joinedTeams", headers=headers + ) if teams_resp.status_code == 401: - return {"status": "auth_error", "message": "Teams token expired. Please re-authenticate.", "connector_type": "teams"} + return { + "status": "auth_error", + "message": "Teams token expired. Please re-authenticate.", + "connector_type": "teams", + } if teams_resp.status_code != 200: - return {"status": "error", "message": f"Graph API error: {teams_resp.status_code}"} + return { + "status": "error", + "message": f"Graph API error: {teams_resp.status_code}", + } teams_data = teams_resp.json().get("value", []) result_teams = [] @@ -58,13 +67,19 @@ def create_list_teams_channels_tool( {"id": ch["id"], "name": ch.get("displayName", "")} for ch in ch_resp.json().get("value", []) ] - result_teams.append({ - "team_id": team_id, - "team_name": team.get("displayName", ""), - "channels": channels, - }) + result_teams.append( + { + "team_id": team_id, + "team_name": team.get("displayName", ""), + "channels": channels, + } + ) - return {"status": "success", "teams": result_teams, "total_teams": len(result_teams)} + return { + "status": "success", + "teams": result_teams, + "total_teams": len(result_teams), + } except Exception as e: from langgraph.errors import GraphInterrupt diff --git a/surfsense_backend/app/agents/new_chat/tools/teams/read_messages.py b/surfsense_backend/app/agents/new_chat/tools/teams/read_messages.py index 90896cb95..d24a7e4d3 100644 --- a/surfsense_backend/app/agents/new_chat/tools/teams/read_messages.py +++ b/surfsense_backend/app/agents/new_chat/tools/teams/read_messages.py @@ -52,11 +52,21 @@ def create_read_teams_messages_tool( ) if resp.status_code == 401: - return {"status": "auth_error", "message": "Teams token expired. Please re-authenticate.", "connector_type": "teams"} + return { + "status": "auth_error", + "message": "Teams token expired. Please re-authenticate.", + "connector_type": "teams", + } if resp.status_code == 403: - return {"status": "error", "message": "Insufficient permissions to read this channel."} + return { + "status": "error", + "message": "Insufficient permissions to read this channel.", + } if resp.status_code != 200: - return {"status": "error", "message": f"Graph API error: {resp.status_code}"} + return { + "status": "error", + "message": f"Graph API error: {resp.status_code}", + } raw_msgs = resp.json().get("value", []) messages = [] @@ -64,13 +74,15 @@ def create_read_teams_messages_tool( sender = m.get("from", {}) user_info = sender.get("user", {}) if sender else {} body = m.get("body", {}) - messages.append({ - "id": m.get("id"), - "sender": user_info.get("displayName", "Unknown"), - "content": body.get("content", ""), - "content_type": body.get("contentType", "text"), - "timestamp": m.get("createdDateTime", ""), - }) + messages.append( + { + "id": m.get("id"), + "sender": user_info.get("displayName", "Unknown"), + "content": body.get("content", ""), + "content_type": body.get("contentType", "text"), + "timestamp": m.get("createdDateTime", ""), + } + ) return { "status": "success", diff --git a/surfsense_backend/app/agents/new_chat/tools/teams/send_message.py b/surfsense_backend/app/agents/new_chat/tools/teams/send_message.py index ba3a515d9..fd8d00870 100644 --- a/surfsense_backend/app/agents/new_chat/tools/teams/send_message.py +++ b/surfsense_backend/app/agents/new_chat/tools/teams/send_message.py @@ -50,12 +50,19 @@ def create_send_teams_message_tool( result = request_approval( action_type="teams_send_message", tool_name="send_teams_message", - params={"team_id": team_id, "channel_id": channel_id, "content": content}, + params={ + "team_id": team_id, + "channel_id": channel_id, + "content": content, + }, context={"connector_id": connector.id}, ) if result.rejected: - return {"status": "rejected", "message": "User declined. Message was not sent."} + return { + "status": "rejected", + "message": "User declined. Message was not sent.", + } final_content = result.params.get("content", content) final_team = result.params.get("team_id", team_id) @@ -74,20 +81,27 @@ def create_send_teams_message_tool( ) if resp.status_code == 401: - return {"status": "auth_error", "message": "Teams token expired. Please re-authenticate.", "connector_type": "teams"} + return { + "status": "auth_error", + "message": "Teams token expired. Please re-authenticate.", + "connector_type": "teams", + } if resp.status_code == 403: return { "status": "insufficient_permissions", "message": "Missing ChannelMessage.Send permission. Please re-authenticate with updated scopes.", } if resp.status_code not in (200, 201): - return {"status": "error", "message": f"Graph API error: {resp.status_code} — {resp.text[:200]}"} + return { + "status": "error", + "message": f"Graph API error: {resp.status_code} — {resp.text[:200]}", + } msg_data = resp.json() return { "status": "success", "message_id": msg_data.get("id"), - "message": f"Message sent to Teams channel.", + "message": "Message sent to Teams channel.", } except Exception as e: diff --git a/surfsense_backend/app/agents/new_chat/tools/tool_response.py b/surfsense_backend/app/agents/new_chat/tools/tool_response.py index 5fb1864b7..8644ada5c 100644 --- a/surfsense_backend/app/agents/new_chat/tools/tool_response.py +++ b/surfsense_backend/app/agents/new_chat/tools/tool_response.py @@ -6,7 +6,6 @@ from typing import Any class ToolResponse: - @staticmethod def success(message: str, **data: Any) -> dict[str, Any]: return {"status": "success", "message": message, **data} @@ -31,9 +30,7 @@ class ToolResponse: return {"status": "rejected", "message": message} @staticmethod - def not_found( - resource: str, identifier: str, **data: Any - ) -> dict[str, Any]: + def not_found(resource: str, identifier: str, **data: Any) -> dict[str, Any]: return { "status": "not_found", "error": f"{resource} '{identifier}' was not found.", diff --git a/surfsense_backend/app/connectors/exceptions.py b/surfsense_backend/app/connectors/exceptions.py index 32a1e7bdc..027adbb87 100644 --- a/surfsense_backend/app/connectors/exceptions.py +++ b/surfsense_backend/app/connectors/exceptions.py @@ -13,7 +13,6 @@ from typing import Any class ConnectorError(Exception): - def __init__( self, message: str, diff --git a/surfsense_backend/app/routes/__init__.py b/surfsense_backend/app/routes/__init__.py index 8df930f30..de4e05423 100644 --- a/surfsense_backend/app/routes/__init__.py +++ b/surfsense_backend/app/routes/__init__.py @@ -98,7 +98,9 @@ router.include_router(logs_router) router.include_router(circleback_webhook_router) # Circleback meeting webhooks router.include_router(surfsense_docs_router) # Surfsense documentation for citations router.include_router(notifications_router) # Notifications with Zero sync -router.include_router(mcp_oauth_router) # MCP OAuth 2.1 for Linear, Jira, ClickUp, Slack, Airtable +router.include_router( + mcp_oauth_router +) # MCP OAuth 2.1 for Linear, Jira, ClickUp, Slack, Airtable router.include_router(composio_router) # Composio OAuth and toolkit management router.include_router(public_chat_router) # Public chat sharing and cloning router.include_router(incentive_tasks_router) # Incentive tasks for earning free pages diff --git a/surfsense_backend/app/routes/mcp_oauth_route.py b/surfsense_backend/app/routes/mcp_oauth_route.py index e14be83d0..1abc1f1ec 100644 --- a/surfsense_backend/app/routes/mcp_oauth_route.py +++ b/surfsense_backend/app/routes/mcp_oauth_route.py @@ -29,7 +29,11 @@ from app.db import ( ) from app.users import current_active_user from app.utils.connector_naming import generate_unique_connector_name -from app.utils.oauth_security import OAuthStateManager, TokenEncryption, generate_pkce_pair +from app.utils.oauth_security import ( + OAuthStateManager, + TokenEncryption, + generate_pkce_pair, +) logger = logging.getLogger(__name__) @@ -37,7 +41,9 @@ router = APIRouter() async def _fetch_account_metadata( - service_key: str, access_token: str, token_json: dict[str, Any], + service_key: str, + access_token: str, + token_json: dict[str, Any], ) -> dict[str, Any]: """Fetch display-friendly account metadata after a successful token exchange. @@ -86,7 +92,8 @@ async def _fetch_account_metadata( meta["display_name"] = whoami.get("email", "Airtable") else: logger.warning( - "Airtable whoami API returned %d (non-blocking)", resp.status_code, + "Airtable whoami API returned %d (non-blocking)", + resp.status_code, ) except Exception: @@ -98,6 +105,7 @@ async def _fetch_account_metadata( return meta + _state_manager: OAuthStateManager | None = None _token_encryption: TokenEncryption | None = None @@ -151,6 +159,7 @@ def _frontend_redirect( # /add — start MCP OAuth flow # --------------------------------------------------------------------------- + @router.get("/auth/mcp/{service}/connector/add") async def connect_mcp_service( service: str, @@ -170,9 +179,12 @@ async def connect_mcp_service( ) metadata = await discover_oauth_metadata( - svc.mcp_url, origin_override=svc.oauth_discovery_origin, + svc.mcp_url, + origin_override=svc.oauth_discovery_origin, + ) + auth_endpoint = svc.auth_endpoint_override or metadata.get( + "authorization_endpoint" ) - auth_endpoint = svc.auth_endpoint_override or metadata.get("authorization_endpoint") token_endpoint = svc.token_endpoint_override or metadata.get("token_endpoint") registration_endpoint = metadata.get("registration_endpoint") @@ -236,7 +248,9 @@ async def connect_mcp_service( logger.info( "Generated %s MCP OAuth URL for user %s, space %s", - svc.name, user.id, space_id, + svc.name, + user.id, + space_id, ) return {"auth_url": auth_url} @@ -245,7 +259,8 @@ async def connect_mcp_service( except Exception as e: logger.error("Failed to initiate %s MCP OAuth: %s", service, e, exc_info=True) raise HTTPException( - status_code=500, detail=f"Failed to initiate {service} MCP OAuth.", + status_code=500, + detail=f"Failed to initiate {service} MCP OAuth.", ) from e @@ -253,6 +268,7 @@ async def connect_mcp_service( # /callback — handle OAuth redirect # --------------------------------------------------------------------------- + @router.get("/auth/mcp/{service}/connector/callback") async def mcp_oauth_callback( service: str, @@ -271,7 +287,9 @@ async def mcp_oauth_callback( except Exception: pass return _frontend_redirect( - space_id, error=f"{service}_mcp_oauth_denied", service=service, + space_id, + error=f"{service}_mcp_oauth_denied", + service=service, ) if not code: @@ -337,9 +355,7 @@ async def mcp_oauth_callback( expires_at = None if expires_in: - expires_at = datetime.now(UTC) + timedelta( - seconds=int(expires_in) - ) + expires_at = datetime.now(UTC) + timedelta(seconds=int(expires_in)) connector_config = { "server_config": { @@ -349,10 +365,14 @@ async def mcp_oauth_callback( "mcp_service": svc_key, "mcp_oauth": { "client_id": client_id, - "client_secret": enc.encrypt_token(client_secret) if client_secret else "", + "client_secret": enc.encrypt_token(client_secret) + if client_secret + else "", "token_endpoint": token_endpoint, "access_token": enc.encrypt_token(access_token), - "refresh_token": enc.encrypt_token(refresh_token) if refresh_token else None, + "refresh_token": enc.encrypt_token(refresh_token) + if refresh_token + else None, "expires_at": expires_at.isoformat() if expires_at else None, "scope": scope, }, @@ -361,15 +381,27 @@ async def mcp_oauth_callback( account_meta = await _fetch_account_metadata(svc_key, access_token, token_json) if account_meta: - _SAFE_META_KEYS = {"display_name", "team_id", "team_name", "user_id", "user_email", - "workspace_id", "workspace_name", "organization_name", - "organization_url_key", "cloud_id", "site_name", "base_url"} + safe_meta_keys = { + "display_name", + "team_id", + "team_name", + "user_id", + "user_email", + "workspace_id", + "workspace_name", + "organization_name", + "organization_url_key", + "cloud_id", + "site_name", + "base_url", + } for k, v in account_meta.items(): - if k in _SAFE_META_KEYS: + if k in safe_meta_keys: connector_config[k] = v logger.info( "Stored account metadata for %s: display_name=%s", - svc_key, account_meta.get("display_name", ""), + svc_key, + account_meta.get("display_name", ""), ) # ---- Re-auth path ---- @@ -400,15 +432,24 @@ async def mcp_oauth_callback( logger.info( "Re-authenticated %s MCP connector %s for user %s", - svc.name, db_connector.id, user_id, + svc.name, + db_connector.id, + user_id, ) reauth_return_url = data.get("return_url") - if reauth_return_url and reauth_return_url.startswith("/") and not reauth_return_url.startswith("//"): + if ( + reauth_return_url + and reauth_return_url.startswith("/") + and not reauth_return_url.startswith("//") + ): return RedirectResponse( url=f"{config.NEXT_FRONTEND_URL}{reauth_return_url}" ) return _frontend_redirect( - space_id, success=True, connector_id=db_connector.id, service=service, + space_id, + success=True, + connector_id=db_connector.id, + service=service, ) # ---- New connector path ---- @@ -436,24 +477,34 @@ async def mcp_oauth_callback( except IntegrityError as e: await session.rollback() raise HTTPException( - status_code=409, detail="A connector for this service already exists.", + status_code=409, + detail="A connector for this service already exists.", ) from e _invalidate_cache(space_id) logger.info( "Created %s MCP connector %s for user %s in space %s", - svc.name, new_connector.id, user_id, space_id, + svc.name, + new_connector.id, + user_id, + space_id, ) return _frontend_redirect( - space_id, success=True, connector_id=new_connector.id, service=service, + space_id, + success=True, + connector_id=new_connector.id, + service=service, ) except HTTPException: raise except Exception as e: logger.error( - "Failed to complete %s MCP OAuth: %s", service, e, exc_info=True, + "Failed to complete %s MCP OAuth: %s", + service, + e, + exc_info=True, ) raise HTTPException( status_code=500, @@ -465,6 +516,7 @@ async def mcp_oauth_callback( # /reauth — re-authenticate an existing MCP connector # --------------------------------------------------------------------------- + @router.get("/auth/mcp/{service}/connector/reauth") async def reauth_mcp_service( service: str, @@ -491,7 +543,8 @@ async def reauth_mcp_service( ) if not result.scalars().first(): raise HTTPException( - status_code=404, detail="Connector not found or access denied", + status_code=404, + detail="Connector not found or access denied", ) try: @@ -501,9 +554,12 @@ async def reauth_mcp_service( ) metadata = await discover_oauth_metadata( - svc.mcp_url, origin_override=svc.oauth_discovery_origin, + svc.mcp_url, + origin_override=svc.oauth_discovery_origin, + ) + auth_endpoint = svc.auth_endpoint_override or metadata.get( + "authorization_endpoint" ) - auth_endpoint = svc.auth_endpoint_override or metadata.get("authorization_endpoint") token_endpoint = svc.token_endpoint_override or metadata.get("token_endpoint") registration_endpoint = metadata.get("registration_endpoint") @@ -545,7 +601,9 @@ async def reauth_mcp_service( "service": service, "code_verifier": verifier, "mcp_client_id": client_id, - "mcp_client_secret": enc.encrypt_token(client_secret) if client_secret else "", + "mcp_client_secret": enc.encrypt_token(client_secret) + if client_secret + else "", "mcp_token_endpoint": token_endpoint, "mcp_url": svc.mcp_url, "connector_id": connector_id, @@ -554,7 +612,9 @@ async def reauth_mcp_service( extra["return_url"] = return_url state = _get_state_manager().generate_secure_state( - space_id, user.id, **extra, + space_id, + user.id, + **extra, ) auth_params: dict[str, str] = { @@ -572,7 +632,9 @@ async def reauth_mcp_service( logger.info( "Initiating %s MCP re-auth for user %s, connector %s", - svc.name, user.id, connector_id, + svc.name, + user.id, + connector_id, ) return {"auth_url": auth_url} @@ -580,7 +642,10 @@ async def reauth_mcp_service( raise except Exception as e: logger.error( - "Failed to initiate %s MCP re-auth: %s", service, e, exc_info=True, + "Failed to initiate %s MCP re-auth: %s", + service, + e, + exc_info=True, ) raise HTTPException( status_code=500, @@ -592,6 +657,7 @@ async def reauth_mcp_service( # Helpers # --------------------------------------------------------------------------- + def _invalidate_cache(space_id: int) -> None: try: from app.agents.new_chat.tools.mcp_tool import invalidate_mcp_tools_cache diff --git a/surfsense_backend/app/routes/new_chat_routes.py b/surfsense_backend/app/routes/new_chat_routes.py index 85a8658ec..091e87737 100644 --- a/surfsense_backend/app/routes/new_chat_routes.py +++ b/surfsense_backend/app/routes/new_chat_routes.py @@ -24,9 +24,9 @@ from sqlalchemy.orm import selectinload from app.agents.new_chat.filesystem_selection import ( ClientPlatform, - LocalFilesystemMount, FilesystemMode, FilesystemSelection, + LocalFilesystemMount, ) from app.config import config from app.db import ( diff --git a/surfsense_backend/app/routes/oauth_connector_base.py b/surfsense_backend/app/routes/oauth_connector_base.py index 0638e8f34..5b75d8519 100644 --- a/surfsense_backend/app/routes/oauth_connector_base.py +++ b/surfsense_backend/app/routes/oauth_connector_base.py @@ -9,6 +9,7 @@ Call ``build_router()`` to get a FastAPI ``APIRouter`` with ``/connector/add``, from __future__ import annotations import base64 +import contextlib import logging from datetime import UTC, datetime, timedelta from typing import Any @@ -41,7 +42,6 @@ logger = logging.getLogger(__name__) class OAuthConnectorRoute: - def __init__( self, *, @@ -244,10 +244,8 @@ class OAuthConnectorRoute: if resp.status_code != 200: detail = resp.text - try: + with contextlib.suppress(Exception): detail = resp.json().get("error_description", detail) - except Exception: - pass raise HTTPException( status_code=400, detail=f"Token exchange failed: {detail}" ) @@ -430,7 +428,11 @@ class OAuthConnectorRoute: state_mgr = oauth._get_state_manager() extra: dict[str, Any] = {"connector_id": connector_id} - if return_url and return_url.startswith("/") and not return_url.startswith("//"): + if ( + return_url + and return_url.startswith("/") + and not return_url.startswith("//") + ): extra["return_url"] = return_url auth_params: dict[str, str] = { @@ -450,9 +452,7 @@ class OAuthConnectorRoute: auth_params.update(oauth.extra_auth_params) - state_encoded = state_mgr.generate_secure_state( - space_id, user.id, **extra - ) + state_encoded = state_mgr.generate_secure_state(space_id, user.id, **extra) auth_params["state"] = state_encoded auth_url = f"{oauth.authorize_url}?{urlencode(auth_params)}" @@ -489,9 +489,7 @@ class OAuthConnectorRoute: status_code=400, detail="Missing authorization code" ) if not state: - raise HTTPException( - status_code=400, detail="Missing state parameter" - ) + raise HTTPException(status_code=400, detail="Missing state parameter") state_mgr = oauth._get_state_manager() try: @@ -552,7 +550,11 @@ class OAuthConnectorRoute: db_connector.id, user_id, ) - if reauth_return_url and reauth_return_url.startswith("/") and not reauth_return_url.startswith("//"): + if ( + reauth_return_url + and reauth_return_url.startswith("/") + and not reauth_return_url.startswith("//") + ): return RedirectResponse( url=f"{config.NEXT_FRONTEND_URL}{reauth_return_url}" ) @@ -603,7 +605,8 @@ class OAuthConnectorRoute: except IntegrityError as e: await session.rollback() raise HTTPException( - status_code=409, detail="A connector for this service already exists." + status_code=409, + detail="A connector for this service already exists.", ) from e logger.info( diff --git a/surfsense_backend/app/routes/search_source_connectors_routes.py b/surfsense_backend/app/routes/search_source_connectors_routes.py index d42a7fa1a..9037d275a 100644 --- a/surfsense_backend/app/routes/search_source_connectors_routes.py +++ b/surfsense_backend/app/routes/search_source_connectors_routes.py @@ -3092,7 +3092,7 @@ async def trust_mcp_tool( select(SearchSourceConnector).filter( SearchSourceConnector.id == connector_id, SearchSourceConnector.user_id == user.id, - cast(SearchSourceConnector.config, PG_JSONB).has_key("server_config"), # noqa: W601 + cast(SearchSourceConnector.config, PG_JSONB).has_key("server_config"), ) ) connector = result.scalars().first() @@ -3147,7 +3147,7 @@ async def untrust_mcp_tool( select(SearchSourceConnector).filter( SearchSourceConnector.id == connector_id, SearchSourceConnector.user_id == user.id, - cast(SearchSourceConnector.config, PG_JSONB).has_key("server_config"), # noqa: W601 + cast(SearchSourceConnector.config, PG_JSONB).has_key("server_config"), ) ) connector = result.scalars().first() diff --git a/surfsense_backend/app/services/mcp_oauth/discovery.py b/surfsense_backend/app/services/mcp_oauth/discovery.py index b0f3fef2a..dc21443bc 100644 --- a/surfsense_backend/app/services/mcp_oauth/discovery.py +++ b/surfsense_backend/app/services/mcp_oauth/discovery.py @@ -55,7 +55,9 @@ async def register_client( async with httpx.AsyncClient(follow_redirects=True) as client: resp = await client.post( - registration_endpoint, json=payload, timeout=timeout, + registration_endpoint, + json=payload, + timeout=timeout, ) resp.raise_for_status() return resp.json() diff --git a/surfsense_backend/app/services/mcp_oauth/registry.py b/surfsense_backend/app/services/mcp_oauth/registry.py index 49bc74d3d..835d70184 100644 --- a/surfsense_backend/app/services/mcp_oauth/registry.py +++ b/surfsense_backend/app/services/mcp_oauth/registry.py @@ -70,12 +70,14 @@ MCP_SERVICES: dict[str, MCPServiceConfig] = { "createJiraIssue", "editJiraIssue", ], - readonly_tools=frozenset({ - "getAccessibleAtlassianResources", - "searchJiraIssuesUsingJql", - "getVisibleJiraProjects", - "getJiraProjectIssueTypesMetadata", - }), + readonly_tools=frozenset( + { + "getAccessibleAtlassianResources", + "searchJiraIssuesUsingJql", + "getVisibleJiraProjects", + "getJiraProjectIssueTypesMetadata", + } + ), account_metadata_keys=["cloud_id", "site_name", "base_url"], ), "clickup": MCPServiceConfig( @@ -99,15 +101,23 @@ MCP_SERVICES: dict[str, MCPServiceConfig] = { auth_endpoint_override="https://slack.com/oauth/v2_user/authorize", token_endpoint_override="https://slack.com/api/oauth.v2.user.access", scopes=[ - "search:read.public", "search:read.private", "search:read.mpim", "search:read.im", - "channels:history", "groups:history", "mpim:history", "im:history", + "search:read.public", + "search:read.private", + "search:read.mpim", + "search:read.im", + "channels:history", + "groups:history", + "mpim:history", + "im:history", ], allowed_tools=[ "slack_search_channels", "slack_read_channel", "slack_read_thread", ], - readonly_tools=frozenset({"slack_search_channels", "slack_read_channel", "slack_read_thread"}), + readonly_tools=frozenset( + {"slack_search_channels", "slack_read_channel", "slack_read_thread"} + ), # TODO: oauth.v2.user.access only returns team.id, not team.name. # To populate team_name, either add "team:read" scope and call # GET /api/team.info during OAuth callback, or switch to oauth.v2.access. @@ -127,7 +137,9 @@ MCP_SERVICES: dict[str, MCPServiceConfig] = { "list_tables_for_base", "list_records_for_table", ], - readonly_tools=frozenset({"list_bases", "list_tables_for_base", "list_records_for_table"}), + readonly_tools=frozenset( + {"list_bases", "list_tables_for_base", "list_records_for_table"} + ), account_metadata_keys=["user_id", "user_email"], ), } @@ -136,20 +148,22 @@ _CONNECTOR_TYPE_TO_SERVICE: dict[str, MCPServiceConfig] = { svc.connector_type: svc for svc in MCP_SERVICES.values() } -LIVE_CONNECTOR_TYPES: frozenset[SearchSourceConnectorType] = frozenset({ - SearchSourceConnectorType.SLACK_CONNECTOR, - SearchSourceConnectorType.TEAMS_CONNECTOR, - SearchSourceConnectorType.LINEAR_CONNECTOR, - SearchSourceConnectorType.JIRA_CONNECTOR, - SearchSourceConnectorType.CLICKUP_CONNECTOR, - SearchSourceConnectorType.GOOGLE_CALENDAR_CONNECTOR, - SearchSourceConnectorType.COMPOSIO_GOOGLE_CALENDAR_CONNECTOR, - SearchSourceConnectorType.AIRTABLE_CONNECTOR, - SearchSourceConnectorType.GOOGLE_GMAIL_CONNECTOR, - SearchSourceConnectorType.COMPOSIO_GMAIL_CONNECTOR, - SearchSourceConnectorType.DISCORD_CONNECTOR, - SearchSourceConnectorType.LUMA_CONNECTOR, -}) +LIVE_CONNECTOR_TYPES: frozenset[SearchSourceConnectorType] = frozenset( + { + SearchSourceConnectorType.SLACK_CONNECTOR, + SearchSourceConnectorType.TEAMS_CONNECTOR, + SearchSourceConnectorType.LINEAR_CONNECTOR, + SearchSourceConnectorType.JIRA_CONNECTOR, + SearchSourceConnectorType.CLICKUP_CONNECTOR, + SearchSourceConnectorType.GOOGLE_CALENDAR_CONNECTOR, + SearchSourceConnectorType.COMPOSIO_GOOGLE_CALENDAR_CONNECTOR, + SearchSourceConnectorType.AIRTABLE_CONNECTOR, + SearchSourceConnectorType.GOOGLE_GMAIL_CONNECTOR, + SearchSourceConnectorType.COMPOSIO_GMAIL_CONNECTOR, + SearchSourceConnectorType.DISCORD_CONNECTOR, + SearchSourceConnectorType.LUMA_CONNECTOR, + } +) def get_service(key: str) -> MCPServiceConfig | None: diff --git a/surfsense_backend/app/services/obsidian_plugin_indexer.py b/surfsense_backend/app/services/obsidian_plugin_indexer.py index 8fbdad269..0fc4f30f4 100644 --- a/surfsense_backend/app/services/obsidian_plugin_indexer.py +++ b/surfsense_backend/app/services/obsidian_plugin_indexer.py @@ -156,7 +156,9 @@ async def _extract_binary_attachment_markdown( try: raw_bytes = base64.b64decode(payload.binary_base64, validate=True) except Exception: - logger.warning("obsidian attachment payload had invalid base64: %s", payload.path) + logger.warning( + "obsidian attachment payload had invalid base64: %s", payload.path + ) return "", {"attachment_extraction_status": "invalid_binary_payload"} suffix = f".{payload.extension.lstrip('.')}" @@ -180,7 +182,10 @@ async def _extract_binary_attachment_markdown( return result.markdown_content, metadata except Exception as exc: logger.warning( - "obsidian attachment ETL failed for %s: %s", payload.path, exc, exc_info=True + "obsidian attachment ETL failed for %s: %s", + payload.path, + exc, + exc_info=True, ) return "", { "attachment_extraction_status": "etl_failed", diff --git a/surfsense_backend/app/tasks/chat/stream_new_chat.py b/surfsense_backend/app/tasks/chat/stream_new_chat.py index 5a6117808..7239c57a5 100644 --- a/surfsense_backend/app/tasks/chat/stream_new_chat.py +++ b/surfsense_backend/app/tasks/chat/stream_new_chat.py @@ -31,7 +31,6 @@ from sqlalchemy.orm import selectinload from app.agents.new_chat.chat_deepagent import create_surfsense_deep_agent from app.agents.new_chat.checkpointer import get_checkpointer from app.agents.new_chat.filesystem_selection import FilesystemSelection -from app.config import config from app.agents.new_chat.llm_config import ( AgentConfig, create_chat_litellm_from_agent_config, @@ -182,9 +181,9 @@ def _tool_output_has_error(tool_output: Any) -> bool: if tool_output.get("error"): return True result = tool_output.get("result") - if isinstance(result, str) and result.strip().lower().startswith("error:"): - return True - return False + return bool( + isinstance(result, str) and result.strip().lower().startswith("error:") + ) if isinstance(tool_output, str): return tool_output.strip().lower().startswith("error:") return False @@ -230,7 +229,9 @@ def _log_file_contract(stage: str, result: StreamResult, **extra: Any) -> None: "stage": stage, "request_id": result.request_id or "unknown", "turn_id": result.turn_id or "unknown", - "chat_id": result.turn_id.split(":", 1)[0] if ":" in result.turn_id else "unknown", + "chat_id": result.turn_id.split(":", 1)[0] + if ":" in result.turn_id + else "unknown", "filesystem_mode": result.filesystem_mode, "client_platform": result.client_platform, "intent_detected": result.intent_detected, @@ -242,7 +243,9 @@ def _log_file_contract(stage: str, result: StreamResult, **extra: Any) -> None: "commit_gate_reason": result.commit_gate_reason or None, } payload.update(extra) - _perf_log.info("[file_operation_contract] %s", json.dumps(payload, ensure_ascii=False)) + _perf_log.info( + "[file_operation_contract] %s", json.dumps(payload, ensure_ascii=False) + ) async def _stream_agent_events( @@ -1289,7 +1292,8 @@ async def _stream_agent_events( result.intent_detected = intent_value if ( isinstance(intent_value, str) - and intent_value in ( + and intent_value + in ( "chat_only", "file_write", "file_read", @@ -1308,18 +1312,17 @@ async def _stream_agent_events( result.commit_gate_passed, result.commit_gate_reason = ( _evaluate_file_contract_outcome(result) ) - if not result.commit_gate_passed: - if _contract_enforcement_active(result): - gate_notice = ( - "I could not complete the requested file write because no successful " - "write_file/edit_file operation was confirmed." - ) - gate_text_id = streaming_service.generate_text_id() - yield streaming_service.format_text_start(gate_text_id) - yield streaming_service.format_text_delta(gate_text_id, gate_notice) - yield streaming_service.format_text_end(gate_text_id) - yield streaming_service.format_terminal_info(gate_notice, "error") - accumulated_text = gate_notice + if not result.commit_gate_passed and _contract_enforcement_active(result): + gate_notice = ( + "I could not complete the requested file write because no successful " + "write_file/edit_file operation was confirmed." + ) + gate_text_id = streaming_service.generate_text_id() + yield streaming_service.format_text_start(gate_text_id) + yield streaming_service.format_text_delta(gate_text_id, gate_notice) + yield streaming_service.format_text_end(gate_text_id) + yield streaming_service.format_terminal_info(gate_notice, "error") + accumulated_text = gate_notice else: result.commit_gate_passed = True result.commit_gate_reason = "" diff --git a/surfsense_backend/app/utils/async_retry.py b/surfsense_backend/app/utils/async_retry.py index c3bdd5386..a56f6550a 100644 --- a/surfsense_backend/app/utils/async_retry.py +++ b/surfsense_backend/app/utils/async_retry.py @@ -2,6 +2,7 @@ from __future__ import annotations +import contextlib import logging from collections.abc import Callable from typing import TypeVar @@ -32,9 +33,7 @@ F = TypeVar("F", bound=Callable) def _is_retryable(exc: BaseException) -> bool: if isinstance(exc, ConnectorError): return exc.retryable - if isinstance(exc, (httpx.TimeoutException, httpx.ConnectError)): - return True - return False + return bool(isinstance(exc, (httpx.TimeoutException, httpx.ConnectError))) def build_retry( @@ -86,10 +85,8 @@ def raise_for_status( retry_after_raw = response.headers.get("Retry-After") retry_after: float | None = None if retry_after_raw: - try: + with contextlib.suppress(ValueError, TypeError): retry_after = float(retry_after_raw) - except (ValueError, TypeError): - pass raise ConnectorRateLimitError( f"{service} rate limited (429)", service=service, diff --git a/surfsense_backend/app/utils/connector_naming.py b/surfsense_backend/app/utils/connector_naming.py index 889bf1464..99c8243a5 100644 --- a/surfsense_backend/app/utils/connector_naming.py +++ b/surfsense_backend/app/utils/connector_naming.py @@ -233,7 +233,10 @@ async def generate_unique_connector_name( if identifier: name = f"{base} - {identifier}" return await ensure_unique_connector_name( - session, name, search_space_id, user_id, + session, + name, + search_space_id, + user_id, ) count = await count_connectors_of_type( diff --git a/surfsense_backend/tests/integration/test_obsidian_plugin_routes.py b/surfsense_backend/tests/integration/test_obsidian_plugin_routes.py index 41779a570..22f6c6de5 100644 --- a/surfsense_backend/tests/integration/test_obsidian_plugin_routes.py +++ b/surfsense_backend/tests/integration/test_obsidian_plugin_routes.py @@ -499,7 +499,9 @@ class TestWireContractSmoke: "app.routes.obsidian_plugin_routes.upsert_note", new=AsyncMock(return_value=fake_doc), ) as upsert_mock, - patch("app.routes.obsidian_plugin_routes._queue_obsidian_attachment") as queue_mock, + patch( + "app.routes.obsidian_plugin_routes._queue_obsidian_attachment" + ) as queue_mock, ): sync_resp = await obsidian_sync( SyncBatchRequest( @@ -548,7 +550,9 @@ class TestWireContractSmoke: "app.routes.obsidian_plugin_routes.upsert_note", new=AsyncMock(return_value=fake_doc), ), - patch("app.routes.obsidian_plugin_routes._queue_obsidian_attachment") as queue_mock, + patch( + "app.routes.obsidian_plugin_routes._queue_obsidian_attachment" + ) as queue_mock, ): sync_resp = await obsidian_sync( SyncBatchRequest( @@ -600,7 +604,9 @@ class TestWireContractSmoke: "app.routes.obsidian_plugin_routes.upsert_note", new=AsyncMock(return_value=fake_doc), ), - patch("app.routes.obsidian_plugin_routes._queue_obsidian_attachment") as queue_mock, + patch( + "app.routes.obsidian_plugin_routes._queue_obsidian_attachment" + ) as queue_mock, ): sync_resp = await obsidian_sync( SyncBatchRequest( @@ -619,7 +625,5 @@ class TestWireContractSmoke: items_by_path = {it.path: it for it in sync_resp.items} assert items_by_path["ok.md"].status == "ok" assert items_by_path["image.png"].status == "error" - assert "does not match extension" in ( - items_by_path["image.png"].error or "" - ) + assert "does not match extension" in (items_by_path["image.png"].error or "") queue_mock.assert_not_called() diff --git a/surfsense_backend/tests/unit/middleware/test_file_intent_middleware.py b/surfsense_backend/tests/unit/middleware/test_file_intent_middleware.py index 673331b0a..7fd3fe4a7 100644 --- a/surfsense_backend/tests/unit/middleware/test_file_intent_middleware.py +++ b/surfsense_backend/tests/unit/middleware/test_file_intent_middleware.py @@ -45,9 +45,7 @@ async def test_file_write_intent_injects_contract_message(): @pytest.mark.asyncio async def test_non_write_intent_does_not_inject_contract_message(): - llm = _FakeLLM( - '{"intent":"file_read","confidence":0.88,"suggested_filename":null}' - ) + llm = _FakeLLM('{"intent":"file_read","confidence":0.88,"suggested_filename":null}') middleware = FileIntentMiddleware(llm=llm) original_messages = [HumanMessage(content="Read /notes.md")] state = {"messages": original_messages, "turn_id": "abc:def"} @@ -55,7 +53,10 @@ async def test_non_write_intent_does_not_inject_contract_message(): result = await middleware.abefore_agent(state, runtime=None) # type: ignore[arg-type] assert result is not None - assert result["file_operation_contract"]["intent"] == FileOperationIntent.FILE_READ.value + assert ( + result["file_operation_contract"]["intent"] + == FileOperationIntent.FILE_READ.value + ) assert "messages" not in result @@ -211,4 +212,3 @@ def test_fallback_path_keeps_posix_style_absolute_path_for_linux_and_macos() -> ) assert resolved == "/var/log/surfsense/notes.md" - diff --git a/surfsense_backend/tests/unit/middleware/test_filesystem_verification.py b/surfsense_backend/tests/unit/middleware/test_filesystem_verification.py index d00365032..cca15e789 100644 --- a/surfsense_backend/tests/unit/middleware/test_filesystem_verification.py +++ b/surfsense_backend/tests/unit/middleware/test_filesystem_verification.py @@ -2,11 +2,11 @@ from pathlib import Path import pytest +from app.agents.new_chat.filesystem_selection import FilesystemMode +from app.agents.new_chat.middleware.filesystem import SurfSenseFilesystemMiddleware from app.agents.new_chat.middleware.multi_root_local_folder_backend import ( MultiRootLocalFolderBackend, ) -from app.agents.new_chat.filesystem_selection import FilesystemMode -from app.agents.new_chat.middleware.filesystem import SurfSenseFilesystemMiddleware pytestmark = pytest.mark.unit diff --git a/surfsense_backend/tests/unit/test_obsidian_plugin_indexer.py b/surfsense_backend/tests/unit/test_obsidian_plugin_indexer.py index 7ab3c52e0..20795c739 100644 --- a/surfsense_backend/tests/unit/test_obsidian_plugin_indexer.py +++ b/surfsense_backend/tests/unit/test_obsidian_plugin_indexer.py @@ -15,7 +15,6 @@ from app.services.obsidian_plugin_indexer import ( _require_extracted_attachment_content, ) - _FAKE_PNG_B64 = base64.b64encode(b"\x89PNG\r\n\x1a\n").decode("ascii") @@ -102,9 +101,7 @@ async def test_extract_binary_attachment_markdown_uses_etl(monkeypatch) -> None: mime_type="application/pdf", ) - async def _fake_run_etl_extract( # noqa: ANN001 - *, file_path, filename, vision_llm - ): + async def _fake_run_etl_extract(*, file_path, filename, vision_llm): assert filename == "spec.pdf" assert file_path assert vision_llm is None @@ -216,7 +213,7 @@ def test_note_payload_rejects_markdown_with_binary_fields() -> None: def test_require_extracted_attachment_content_rejects_empty_content() -> None: with pytest.raises( - RuntimeError, match="Attachment extraction failed for assets/img.png" + RuntimeError, match=r"Attachment extraction failed for assets/img\.png" ): _require_extracted_attachment_content( content=" ", diff --git a/surfsense_backend/tests/unit/test_stream_new_chat_contract.py b/surfsense_backend/tests/unit/test_stream_new_chat_contract.py index f4adc3d73..034aa484c 100644 --- a/surfsense_backend/tests/unit/test_stream_new_chat_contract.py +++ b/surfsense_backend/tests/unit/test_stream_new_chat_contract.py @@ -45,4 +45,3 @@ def test_contract_enforcement_local_only(): result.filesystem_mode = "cloud" assert not _contract_enforcement_active(result) - diff --git a/surfsense_web/app/dashboard/[search_space_id]/new-chat/[[...chat_id]]/page.tsx b/surfsense_web/app/dashboard/[search_space_id]/new-chat/[[...chat_id]]/page.tsx index 06f3bf79f..9f569398e 100644 --- a/surfsense_web/app/dashboard/[search_space_id]/new-chat/[[...chat_id]]/page.tsx +++ b/surfsense_web/app/dashboard/[search_space_id]/new-chat/[[...chat_id]]/page.tsx @@ -45,8 +45,8 @@ import { } from "@/components/assistant-ui/token-usage-context"; import { useChatSessionStateSync } from "@/hooks/use-chat-session-state"; import { useMessagesSync } from "@/hooks/use-messages-sync"; -import { documentsApiService } from "@/lib/apis/documents-api.service"; import { getAgentFilesystemSelection } from "@/lib/agent-filesystem"; +import { documentsApiService } from "@/lib/apis/documents-api.service"; import { getBearerToken } from "@/lib/auth-utils"; import { convertToThreadMessage } from "@/lib/chat/message-utils"; import { @@ -661,8 +661,7 @@ export default function NewChatPage() { const selection = await getAgentFilesystemSelection(searchSpaceId); if ( selection.filesystem_mode === "desktop_local_folder" && - (!selection.local_filesystem_mounts || - selection.local_filesystem_mounts.length === 0) + (!selection.local_filesystem_mounts || selection.local_filesystem_mounts.length === 0) ) { toast.error("Select a local folder before using Local Folder mode."); return; @@ -842,14 +841,7 @@ export default function NewChatPage() { }); } else { const tcId = `interrupt-${action.name}`; - addToolCall( - contentPartsState, - toolsWithUI, - tcId, - action.name, - action.args, - true - ); + addToolCall(contentPartsState, toolsWithUI, tcId, action.name, action.args, true); updateToolCall(contentPartsState, tcId, { result: { __interrupt__: true, ...interruptData }, }); @@ -1189,14 +1181,7 @@ export default function NewChatPage() { }); } else { const tcId = `interrupt-${action.name}`; - addToolCall( - contentPartsState, - toolsWithUI, - tcId, - action.name, - action.args, - true - ); + addToolCall(contentPartsState, toolsWithUI, tcId, action.name, action.args, true); updateToolCall(contentPartsState, tcId, { result: { __interrupt__: true, diff --git a/surfsense_web/app/dashboard/[search_space_id]/user-settings/components/DesktopShortcutsContent.tsx b/surfsense_web/app/dashboard/[search_space_id]/user-settings/components/DesktopShortcutsContent.tsx index 6207457c4..12a7d00f0 100644 --- a/surfsense_web/app/dashboard/[search_space_id]/user-settings/components/DesktopShortcutsContent.tsx +++ b/surfsense_web/app/dashboard/[search_space_id]/user-settings/components/DesktopShortcutsContent.tsx @@ -111,9 +111,7 @@ function HotkeyRow({ } > {recording ? ( - - Press hotkeys... - + Press hotkeys... ) : ( )} @@ -155,7 +153,9 @@ export function DesktopShortcutsContent() { if (!api) { return (
-

Hotkeys are only available in the SurfSense desktop app.

+

+ Hotkeys are only available in the SurfSense desktop app. +

); } @@ -178,28 +178,26 @@ export function DesktopShortcutsContent() { updateShortcut(key, DEFAULT_SHORTCUTS[key]); }; - return ( - shortcutsLoaded ? ( -
-
- {HOTKEY_ROWS.map((row) => ( - updateShortcut(row.key, accel)} - onReset={() => resetShortcut(row.key)} - /> - ))} -
+ return shortcutsLoaded ? ( +
+
+ {HOTKEY_ROWS.map((row) => ( + updateShortcut(row.key, accel)} + onReset={() => resetShortcut(row.key)} + /> + ))}
- ) : ( -
- -
- ) +
+ ) : ( +
+ +
); } diff --git a/surfsense_web/app/desktop/login/page.tsx b/surfsense_web/app/desktop/login/page.tsx index 451143949..c64eb65f8 100644 --- a/surfsense_web/app/desktop/login/page.tsx +++ b/surfsense_web/app/desktop/login/page.tsx @@ -24,7 +24,12 @@ const isGoogleAuth = AUTH_TYPE === "GOOGLE"; type ShortcutKey = "generalAssist" | "quickAsk" | "autocomplete"; type ShortcutMap = typeof DEFAULT_SHORTCUTS; -const HOTKEY_ROWS: Array<{ key: ShortcutKey; label: string; description: string; icon: React.ElementType }> = [ +const HOTKEY_ROWS: Array<{ + key: ShortcutKey; + label: string; + description: string; + icon: React.ElementType; +}> = [ { key: "generalAssist", label: "General Assist", @@ -369,7 +374,9 @@ export default function DesktopLoginPage() { )} diff --git a/surfsense_web/components/assistant-ui/connector-popup.tsx b/surfsense_web/components/assistant-ui/connector-popup.tsx index 66333a9ef..32943142a 100644 --- a/surfsense_web/components/assistant-ui/connector-popup.tsx +++ b/surfsense_web/components/assistant-ui/connector-popup.tsx @@ -123,9 +123,9 @@ export const ConnectorIndicator = forwardRef ) : viewingMCPList ? ( - handleDisconnectFromList(connector, () => refreshConnectors())} - onAddAccount={handleAddNewMCPFromList} - addButtonText="Add New MCP Server" - /> + + handleDisconnectFromList(connector, () => refreshConnectors()) + } + onAddAccount={handleAddNewMCPFromList} + addButtonText="Add New MCP Server" + /> ) : viewingAccountsType ? ( - handleDisconnectFromList(connector, () => refreshConnectors())} - onAddAccount={() => { + + handleDisconnectFromList(connector, () => refreshConnectors()) + } + onAddAccount={() => { // Check both OAUTH_CONNECTORS and COMPOSIO_CONNECTORS const oauthConnector = OAUTH_CONNECTORS.find( diff --git a/surfsense_web/components/assistant-ui/connector-popup/connect-forms/components/mcp-connect-form.tsx b/surfsense_web/components/assistant-ui/connector-popup/connect-forms/components/mcp-connect-form.tsx index fc9812240..d9a740af2 100644 --- a/surfsense_web/components/assistant-ui/connector-popup/connect-forms/components/mcp-connect-form.tsx +++ b/surfsense_web/components/assistant-ui/connector-popup/connect-forms/components/mcp-connect-form.tsx @@ -213,13 +213,13 @@ export const MCPConnectForm: FC = ({ onSubmit, isSubmitting }) className="w-full h-8 text-[13px] px-3 rounded-lg font-medium bg-white text-slate-700 hover:bg-slate-50 border-0 shadow-xs dark:bg-secondary dark:text-secondary-foreground dark:hover:bg-secondary/80" > {isTesting ? ( - <> - - Testing Connection... - - ) : ( - "Test Connection" - )} + <> + + Testing Connection... + + ) : ( + "Test Connection" + )}
diff --git a/surfsense_web/components/assistant-ui/connector-popup/connector-configs/components/mcp-config.tsx b/surfsense_web/components/assistant-ui/connector-popup/connector-configs/components/mcp-config.tsx index d6f60e824..97b5de675 100644 --- a/surfsense_web/components/assistant-ui/connector-popup/connector-configs/components/mcp-config.tsx +++ b/surfsense_web/components/assistant-ui/connector-popup/connector-configs/components/mcp-config.tsx @@ -218,13 +218,13 @@ export const MCPConfig: FC = ({ connector, onConfigChange, onNam className="w-full h-8 text-[13px] px-3 rounded-lg font-medium bg-white text-slate-700 hover:bg-slate-50 border-0 shadow-xs dark:bg-secondary dark:text-secondary-foreground dark:hover:bg-secondary/80" > {isTesting ? ( - <> - - Testing Connection... - - ) : ( - "Test Connection" - )} + <> + + Testing Connection... + + ) : ( + "Test Connection" + )} diff --git a/surfsense_web/components/assistant-ui/connector-popup/connector-configs/components/teams-config.tsx b/surfsense_web/components/assistant-ui/connector-popup/connector-configs/components/teams-config.tsx index e96ddfd29..06ce21dae 100644 --- a/surfsense_web/components/assistant-ui/connector-popup/connector-configs/components/teams-config.tsx +++ b/surfsense_web/components/assistant-ui/connector-popup/connector-configs/components/teams-config.tsx @@ -18,9 +18,9 @@ export const TeamsConfig: FC = () => {

Microsoft Teams Access

- Your agent can search and read messages from Teams channels you have access to, - and send messages on your behalf. Make sure you're a member of the teams - you want to interact with. + Your agent can search and read messages from Teams channels you have access to, and send + messages on your behalf. Make sure you're a member of the teams you want to interact + with.

diff --git a/surfsense_web/components/assistant-ui/connector-popup/connector-configs/views/connector-edit-view.tsx b/surfsense_web/components/assistant-ui/connector-popup/connector-configs/views/connector-edit-view.tsx index b2b40dfd6..c104f140a 100644 --- a/surfsense_web/components/assistant-ui/connector-popup/connector-configs/views/connector-edit-view.tsx +++ b/surfsense_web/components/assistant-ui/connector-popup/connector-configs/views/connector-edit-view.tsx @@ -16,7 +16,7 @@ import { DateRangeSelector } from "../../components/date-range-selector"; import { PeriodicSyncConfig } from "../../components/periodic-sync-config"; import { SummaryConfig } from "../../components/summary-config"; import { VisionLLMConfig } from "../../components/vision-llm-config"; -import { LIVE_CONNECTOR_TYPES, getReauthEndpoint } from "../../constants/connector-constants"; +import { getReauthEndpoint, LIVE_CONNECTOR_TYPES } from "../../constants/connector-constants"; import { getConnectorDisplayName } from "../../tabs/all-connectors-tab"; import { MCPServiceConfig } from "../components/mcp-service-config"; import { getConnectorConfigComponent } from "../index"; @@ -314,8 +314,7 @@ export const ConnectorEditView: FC = ({ {connector.is_indexable && (() => { - const isGoogleDrive = - connector.connector_type === "GOOGLE_DRIVE_CONNECTOR"; + const isGoogleDrive = connector.connector_type === "GOOGLE_DRIVE_CONNECTOR"; const isComposioGoogleDrive = connector.connector_type === "COMPOSIO_GOOGLE_DRIVE_CONNECTOR"; const requiresFolderSelection = isGoogleDrive || isComposioGoogleDrive; @@ -327,8 +326,7 @@ export const ConnectorEditView: FC = ({ (connector.config?.selected_files as | Array<{ id: string; name: string }> | undefined) || []; - const hasItemsSelected = - selectedFolders.length > 0 || selectedFiles.length > 0; + const hasItemsSelected = selectedFolders.length > 0 || selectedFiles.length > 0; const isDisabled = requiresFolderSelection && !hasItemsSelected; return ( @@ -380,8 +378,8 @@ export const ConnectorEditView: FC = ({ {/* Fixed Footer - Action buttons */}
- {showDisconnectConfirm ? ( -
+ {showDisconnectConfirm ? ( +
{isLive ? "Your agent will lose access to this service." diff --git a/surfsense_web/components/assistant-ui/connector-popup/connector-configs/views/indexing-configuration-view.tsx b/surfsense_web/components/assistant-ui/connector-popup/connector-configs/views/indexing-configuration-view.tsx index 690333523..982b0be11 100644 --- a/surfsense_web/components/assistant-ui/connector-popup/connector-configs/views/indexing-configuration-view.tsx +++ b/surfsense_web/components/assistant-ui/connector-popup/connector-configs/views/indexing-configuration-view.tsx @@ -12,7 +12,10 @@ import { DateRangeSelector } from "../../components/date-range-selector"; import { PeriodicSyncConfig } from "../../components/periodic-sync-config"; import { SummaryConfig } from "../../components/summary-config"; import { VisionLLMConfig } from "../../components/vision-llm-config"; -import { LIVE_CONNECTOR_TYPES, type IndexingConfigState } from "../../constants/connector-constants"; +import { + type IndexingConfigState, + LIVE_CONNECTOR_TYPES, +} from "../../constants/connector-constants"; import { getConnectorDisplayName } from "../../tabs/all-connectors-tab"; import { getConnectorConfigComponent } from "../index"; diff --git a/surfsense_web/components/assistant-ui/connector-popup/tabs/active-connectors-tab.tsx b/surfsense_web/components/assistant-ui/connector-popup/tabs/active-connectors-tab.tsx index fe9aab14f..755086ba5 100644 --- a/surfsense_web/components/assistant-ui/connector-popup/tabs/active-connectors-tab.tsx +++ b/surfsense_web/components/assistant-ui/connector-popup/tabs/active-connectors-tab.tsx @@ -9,7 +9,11 @@ import { getConnectorIcon } from "@/contracts/enums/connectorIcons"; import type { SearchSourceConnector } from "@/contracts/types/connector.types"; import { getDocumentTypeLabel } from "@/lib/documents/document-type-labels"; import { cn } from "@/lib/utils"; -import { COMPOSIO_CONNECTORS, LIVE_CONNECTOR_TYPES, OAUTH_CONNECTORS } from "../constants/connector-constants"; +import { + COMPOSIO_CONNECTORS, + LIVE_CONNECTOR_TYPES, + OAUTH_CONNECTORS, +} from "../constants/connector-constants"; import { getDocumentCountForConnector } from "../utils/connector-document-mapping"; import { getConnectorDisplayName } from "./all-connectors-tab"; diff --git a/surfsense_web/components/assistant-ui/connector-popup/views/connector-accounts-list-view.tsx b/surfsense_web/components/assistant-ui/connector-popup/views/connector-accounts-list-view.tsx index b3c087599..8aee7e005 100644 --- a/surfsense_web/components/assistant-ui/connector-popup/views/connector-accounts-list-view.tsx +++ b/surfsense_web/components/assistant-ui/connector-popup/views/connector-accounts-list-view.tsx @@ -13,7 +13,7 @@ import type { SearchSourceConnector } from "@/contracts/types/connector.types"; import { authenticatedFetch } from "@/lib/auth-utils"; import { formatRelativeDate } from "@/lib/format-date"; import { cn } from "@/lib/utils"; -import { LIVE_CONNECTOR_TYPES, getReauthEndpoint } from "../constants/connector-constants"; +import { getReauthEndpoint, LIVE_CONNECTOR_TYPES } from "../constants/connector-constants"; import { useConnectorStatus } from "../hooks/use-connector-status"; import { getConnectorDisplayName } from "../tabs/all-connectors-tab"; @@ -182,11 +182,14 @@ export const ConnectorAccountsListView: FC = ({
) : (
- {typeConnectors.map((connector) => { - const isIndexing = indexingConnectorIds.has(connector.id); - const connectorReauthEndpoint = getReauthEndpoint(connector); - const isAuthExpired = !!connectorReauthEndpoint && connector.config?.auth_expired === true; - const isLive = LIVE_CONNECTOR_TYPES.has(connector.connector_type) || Boolean(connector.config?.server_config); + {typeConnectors.map((connector) => { + const isIndexing = indexingConnectorIds.has(connector.id); + const connectorReauthEndpoint = getReauthEndpoint(connector); + const isAuthExpired = + !!connectorReauthEndpoint && connector.config?.auth_expired === true; + const isLive = + LIVE_CONNECTOR_TYPES.has(connector.connector_type) || + Boolean(connector.config?.server_config); return (
= ({

) : null}
- {isAuthExpired ? ( - - ) : isLive && onDisconnect ? ( - confirmDisconnectId === connector.id ? ( -
+ {isAuthExpired ? ( + + ) : isLive && onDisconnect ? ( + confirmDisconnectId === connector.id ? ( +
+ + +
+ ) : ( - -
+ ) ) : ( - ) - ) : ( - - )} + )}
); })} diff --git a/surfsense_web/components/assistant-ui/markdown-text.tsx b/surfsense_web/components/assistant-ui/markdown-text.tsx index 2707e8956..8bb228580 100644 --- a/surfsense_web/components/assistant-ui/markdown-text.tsx +++ b/surfsense_web/components/assistant-ui/markdown-text.tsx @@ -20,7 +20,6 @@ import { openEditorPanelAtom } from "@/atoms/editor/editor-panel.atom"; import { ImagePreview, ImageRoot, ImageZoom } from "@/components/assistant-ui/image"; import "katex/dist/katex.min.css"; import { InlineCitation, UrlCitation } from "@/components/assistant-ui/inline-citation"; -import { useElectronAPI } from "@/hooks/use-platform"; import { Skeleton } from "@/components/ui/skeleton"; import { Table, @@ -30,6 +29,7 @@ import { TableHeader, TableRow, } from "@/components/ui/table"; +import { useElectronAPI } from "@/hooks/use-platform"; import { cn } from "@/lib/utils"; function MarkdownCodeBlockSkeleton() { @@ -493,10 +493,7 @@ const defaultComponents = memoizeMarkdownComponents({ const mounts = (await electronAPI.getAgentFilesystemMounts( resolvedSearchSpaceId )) as AgentFilesystemMount[]; - resolvedLocalPath = normalizeLocalVirtualPathForEditor( - inlineValue, - mounts - ); + resolvedLocalPath = normalizeLocalVirtualPathForEditor(inlineValue, mounts); } catch { // Fall back to the raw inline path if mount lookup fails. } diff --git a/surfsense_web/components/editor-panel/editor-panel.tsx b/surfsense_web/components/editor-panel/editor-panel.tsx index 2fa980d27..3b69ae6e0 100644 --- a/surfsense_web/components/editor-panel/editor-panel.tsx +++ b/surfsense_web/components/editor-panel/editor-panel.tsx @@ -248,7 +248,15 @@ export function EditorPanelContent({ doFetch().catch(() => {}); return () => controller.abort(); - }, [documentId, electronAPI, isLocalFileMode, localFilePath, resolveLocalVirtualPath, searchSpaceId, title]); + }, [ + documentId, + electronAPI, + isLocalFileMode, + localFilePath, + resolveLocalVirtualPath, + searchSpaceId, + title, + ]); useEffect(() => { return () => { @@ -282,69 +290,77 @@ export function EditorPanelContent({ } }, [editorDoc?.source_markdown]); - const handleSave = useCallback(async (_options?: { silent?: boolean }) => { - setSaving(true); - try { - if (isLocalFileMode) { - if (!localFilePath) { - throw new Error("Missing local file path"); + const handleSave = useCallback( + async (_options?: { silent?: boolean }) => { + setSaving(true); + try { + if (isLocalFileMode) { + if (!localFilePath) { + throw new Error("Missing local file path"); + } + if (!electronAPI?.writeAgentLocalFileText) { + throw new Error("Local file editor is available only in desktop mode."); + } + const resolvedLocalPath = await resolveLocalVirtualPath(localFilePath); + const contentToSave = markdownRef.current; + const writeResult = await electronAPI.writeAgentLocalFileText( + resolvedLocalPath, + contentToSave, + searchSpaceId + ); + if (!writeResult.ok) { + throw new Error(writeResult.error || "Failed to save local file"); + } + setEditorDoc((prev) => (prev ? { ...prev, source_markdown: contentToSave } : prev)); + setEditedMarkdown(markdownRef.current === contentToSave ? null : markdownRef.current); + return true; } - if (!electronAPI?.writeAgentLocalFileText) { - throw new Error("Local file editor is available only in desktop mode."); + if (!searchSpaceId || !documentId) { + throw new Error("Missing document context"); } - const resolvedLocalPath = await resolveLocalVirtualPath(localFilePath); - const contentToSave = markdownRef.current; - const writeResult = await electronAPI.writeAgentLocalFileText( - resolvedLocalPath, - contentToSave, - searchSpaceId + const token = getBearerToken(); + if (!token) { + toast.error("Please login to save"); + redirectToLogin(); + return; + } + const response = await authenticatedFetch( + `${process.env.NEXT_PUBLIC_FASTAPI_BACKEND_URL}/api/v1/search-spaces/${searchSpaceId}/documents/${documentId}/save`, + { + method: "POST", + headers: { "Content-Type": "application/json" }, + body: JSON.stringify({ source_markdown: markdownRef.current }), + } ); - if (!writeResult.ok) { - throw new Error(writeResult.error || "Failed to save local file"); + + if (!response.ok) { + const errorData = await response + .json() + .catch(() => ({ detail: "Failed to save document" })); + throw new Error(errorData.detail || "Failed to save document"); } - setEditorDoc((prev) => - prev ? { ...prev, source_markdown: contentToSave } : prev - ); - setEditedMarkdown(markdownRef.current === contentToSave ? null : markdownRef.current); + + setEditorDoc((prev) => (prev ? { ...prev, source_markdown: markdownRef.current } : prev)); + setEditedMarkdown(null); + toast.success("Document saved! Reindexing in background..."); return true; + } catch (err) { + console.error("Error saving document:", err); + toast.error(err instanceof Error ? err.message : "Failed to save document"); + return false; + } finally { + setSaving(false); } - if (!searchSpaceId || !documentId) { - throw new Error("Missing document context"); - } - const token = getBearerToken(); - if (!token) { - toast.error("Please login to save"); - redirectToLogin(); - return; - } - const response = await authenticatedFetch( - `${process.env.NEXT_PUBLIC_FASTAPI_BACKEND_URL}/api/v1/search-spaces/${searchSpaceId}/documents/${documentId}/save`, - { - method: "POST", - headers: { "Content-Type": "application/json" }, - body: JSON.stringify({ source_markdown: markdownRef.current }), - } - ); - - if (!response.ok) { - const errorData = await response - .json() - .catch(() => ({ detail: "Failed to save document" })); - throw new Error(errorData.detail || "Failed to save document"); - } - - setEditorDoc((prev) => (prev ? { ...prev, source_markdown: markdownRef.current } : prev)); - setEditedMarkdown(null); - toast.success("Document saved! Reindexing in background..."); - return true; - } catch (err) { - console.error("Error saving document:", err); - toast.error(err instanceof Error ? err.message : "Failed to save document"); - return false; - } finally { - setSaving(false); - } - }, [documentId, electronAPI, isLocalFileMode, localFilePath, resolveLocalVirtualPath, searchSpaceId]); + }, + [ + documentId, + electronAPI, + isLocalFileMode, + localFilePath, + resolveLocalVirtualPath, + searchSpaceId, + ] + ); const isEditableType = editorDoc ? (editorRenderMode === "source_code" || @@ -594,9 +610,7 @@ export function EditorPanelContent({ } }} > - + Download .md @@ -626,7 +640,7 @@ export function EditorPanelContent({
) : isEditableType ? ( ; } diff --git a/surfsense_web/components/editor/plugins/fixed-toolbar-kit.tsx b/surfsense_web/components/editor/plugins/fixed-toolbar-kit.tsx index bdda0263d..346fe0378 100644 --- a/surfsense_web/components/editor/plugins/fixed-toolbar-kit.tsx +++ b/surfsense_web/components/editor/plugins/fixed-toolbar-kit.tsx @@ -1,7 +1,6 @@ "use client"; -import { createPlatePlugin } from "platejs/react"; -import { useEditorReadOnly } from "platejs/react"; +import { createPlatePlugin, useEditorReadOnly } from "platejs/react"; import { useEditorSave } from "@/components/editor/editor-save-context"; import { FixedToolbar } from "@/components/ui/fixed-toolbar"; diff --git a/surfsense_web/components/editor/source-code-editor.tsx b/surfsense_web/components/editor/source-code-editor.tsx index dd4b3bd8e..9102dffe9 100644 --- a/surfsense_web/components/editor/source-code-editor.tsx +++ b/surfsense_web/components/editor/source-code-editor.tsx @@ -1,8 +1,8 @@ "use client"; import dynamic from "next/dynamic"; -import { useEffect, useRef } from "react"; import { useTheme } from "next-themes"; +import { useEffect, useRef } from "react"; import { Spinner } from "@/components/ui/spinner"; const MonacoEditor = dynamic(() => import("@monaco-editor/react"), { diff --git a/surfsense_web/components/layout/ui/right-panel/RightPanel.tsx b/surfsense_web/components/layout/ui/right-panel/RightPanel.tsx index c26cc9b23..04bae010c 100644 --- a/surfsense_web/components/layout/ui/right-panel/RightPanel.tsx +++ b/surfsense_web/components/layout/ui/right-panel/RightPanel.tsx @@ -72,9 +72,7 @@ export function RightPanelExpandButton() { const reportOpen = reportState.isOpen && !!reportState.reportId; const editorOpen = editorState.isOpen && - (editorState.kind === "document" - ? !!editorState.documentId - : !!editorState.localFilePath); + (editorState.kind === "document" ? !!editorState.documentId : !!editorState.localFilePath); const hitlEditOpen = hitlEditState.isOpen && !!hitlEditState.onSave; const hasContent = documentsOpen || reportOpen || editorOpen || hitlEditOpen; @@ -116,9 +114,7 @@ export function RightPanel({ documentsPanel }: RightPanelProps) { const reportOpen = reportState.isOpen && !!reportState.reportId; const editorOpen = editorState.isOpen && - (editorState.kind === "document" - ? !!editorState.documentId - : !!editorState.localFilePath); + (editorState.kind === "document" ? !!editorState.documentId : !!editorState.localFilePath); const hitlEditOpen = hitlEditState.isOpen && !!hitlEditState.onSave; useEffect(() => { diff --git a/surfsense_web/components/layout/ui/sidebar/DesktopLocalTabContent.tsx b/surfsense_web/components/layout/ui/sidebar/DesktopLocalTabContent.tsx index dd7520d24..cd8fca331 100644 --- a/surfsense_web/components/layout/ui/sidebar/DesktopLocalTabContent.tsx +++ b/surfsense_web/components/layout/ui/sidebar/DesktopLocalTabContent.tsx @@ -1,11 +1,9 @@ "use client"; -import { Folder, FolderPlus, Search, X } from "lucide-react"; import { useAtom } from "jotai"; +import { Folder, FolderPlus, Search, X } from "lucide-react"; import { useCallback, useMemo, useRef, useState } from "react"; import { localExpandedFolderKeysAtom } from "@/atoms/documents/folder.atoms"; -import { Input } from "@/components/ui/input"; -import { Separator } from "@/components/ui/separator"; import { DropdownMenu, DropdownMenuContent, @@ -14,6 +12,8 @@ import { DropdownMenuSeparator, DropdownMenuTrigger, } from "@/components/ui/dropdown-menu"; +import { Input } from "@/components/ui/input"; +import { Separator } from "@/components/ui/separator"; import { Tooltip, TooltipContent, TooltipTrigger } from "@/components/ui/tooltip"; import { useDebouncedValue } from "@/hooks/use-debounced-value"; import { LocalFilesystemBrowser } from "./LocalFilesystemBrowser"; diff --git a/surfsense_web/components/layout/ui/sidebar/DocumentsSidebar.tsx b/surfsense_web/components/layout/ui/sidebar/DocumentsSidebar.tsx index b9c174d71..0a147f7b7 100644 --- a/surfsense_web/components/layout/ui/sidebar/DocumentsSidebar.tsx +++ b/surfsense_web/components/layout/ui/sidebar/DocumentsSidebar.tsx @@ -71,7 +71,7 @@ import { getConnectorIcon } from "@/contracts/enums/connectorIcons"; import type { DocumentTypeEnum } from "@/contracts/types/document.types"; import { useDebouncedValue } from "@/hooks/use-debounced-value"; import { useMediaQuery } from "@/hooks/use-media-query"; -import { usePlatform, useElectronAPI } from "@/hooks/use-platform"; +import { useElectronAPI, usePlatform } from "@/hooks/use-platform"; import { anonymousChatApiService } from "@/lib/apis/anonymous-chat-api.service"; import { documentsApiService } from "@/lib/apis/documents-api.service"; import { foldersApiService } from "@/lib/apis/folders-api.service"; @@ -208,7 +208,8 @@ function AuthenticatedDocumentsSidebarBase({ const [watchedFolderIds, setWatchedFolderIds] = useState>(new Set()); const [folderWatchOpen, setFolderWatchOpen] = useAtom(folderWatchDialogOpenAtom); const [watchInitialFolder, setWatchInitialFolder] = useAtom(folderWatchInitialFolderAtom); - const isElectron = desktopFeaturesEnabled && typeof window !== "undefined" && !!window.electronAPI; + const isElectron = + desktopFeaturesEnabled && typeof window !== "undefined" && !!window.electronAPI; useEffect(() => { if (!electronAPI?.getAgentFilesystemSettings) return; @@ -250,10 +251,13 @@ function AuthenticatedDocumentsSidebarBase({ .filter((rootPath, index, allPaths) => allPaths.indexOf(rootPath) === index) .slice(0, MAX_LOCAL_FILESYSTEM_ROOTS); if (nextLocalRootPaths.length === localRootPaths.length) return; - const updated = await electronAPI.setAgentFilesystemSettings({ - mode: "desktop_local_folder", - localRootPaths: nextLocalRootPaths, - }, searchSpaceId); + const updated = await electronAPI.setAgentFilesystemSettings( + { + mode: "desktop_local_folder", + localRootPaths: nextLocalRootPaths, + }, + searchSpaceId + ); setFilesystemSettings(updated); }, [electronAPI, localRootPaths, searchSpaceId] @@ -282,10 +286,13 @@ function AuthenticatedDocumentsSidebarBase({ const handleRemoveFilesystemRoot = useCallback( async (rootPathToRemove: string) => { if (!electronAPI?.setAgentFilesystemSettings) return; - const updated = await electronAPI.setAgentFilesystemSettings({ - mode: "desktop_local_folder", - localRootPaths: localRootPaths.filter((rootPath) => rootPath !== rootPathToRemove), - }, searchSpaceId); + const updated = await electronAPI.setAgentFilesystemSettings( + { + mode: "desktop_local_folder", + localRootPaths: localRootPaths.filter((rootPath) => rootPath !== rootPathToRemove), + }, + searchSpaceId + ); setFilesystemSettings(updated); }, [electronAPI, localRootPaths, searchSpaceId] @@ -293,19 +300,25 @@ function AuthenticatedDocumentsSidebarBase({ const handleClearFilesystemRoots = useCallback(async () => { if (!electronAPI?.setAgentFilesystemSettings) return; - const updated = await electronAPI.setAgentFilesystemSettings({ - mode: "desktop_local_folder", - localRootPaths: [], - }, searchSpaceId); + const updated = await electronAPI.setAgentFilesystemSettings( + { + mode: "desktop_local_folder", + localRootPaths: [], + }, + searchSpaceId + ); setFilesystemSettings(updated); }, [electronAPI, searchSpaceId]); const handleFilesystemTabChange = useCallback( async (tab: "cloud" | "local") => { if (!electronAPI?.setAgentFilesystemSettings) return; - const updated = await electronAPI.setAgentFilesystemSettings({ - mode: tab === "cloud" ? "cloud" : "desktop_local_folder", - }, searchSpaceId); + const updated = await electronAPI.setAgentFilesystemSettings( + { + mode: tab === "cloud" ? "cloud" : "desktop_local_folder", + }, + searchSpaceId + ); setFilesystemSettings(updated); }, [electronAPI, searchSpaceId] @@ -552,7 +565,9 @@ function AuthenticatedDocumentsSidebarBase({ if (!electronAPI) return; const watchedFolders = (await electronAPI.getWatchedFolders()) as WatchedFolderEntry[]; - const matched = watchedFolders.find((wf: WatchedFolderEntry) => wf.rootFolderId === folder.id); + const matched = watchedFolders.find( + (wf: WatchedFolderEntry) => wf.rootFolderId === folder.id + ); if (!matched) { toast.error("This folder is not being watched"); return; @@ -582,7 +597,9 @@ function AuthenticatedDocumentsSidebarBase({ if (!electronAPI) return; const watchedFolders = (await electronAPI.getWatchedFolders()) as WatchedFolderEntry[]; - const matched = watchedFolders.find((wf: WatchedFolderEntry) => wf.rootFolderId === folder.id); + const matched = watchedFolders.find( + (wf: WatchedFolderEntry) => wf.rootFolderId === folder.id + ); if (!matched) { toast.error("This folder is not being watched"); return; @@ -1015,7 +1032,8 @@ function AuthenticatedDocumentsSidebarBase({ }, [open, onOpenChange, isMobile, setRightPanelCollapsed]); const showFilesystemTabs = !isMobile && !!electronAPI && !!filesystemSettings; - const currentFilesystemTab = filesystemSettings?.mode === "desktop_local_folder" ? "local" : "cloud"; + const currentFilesystemTab = + filesystemSettings?.mode === "desktop_local_folder" ? "local" : "cloud"; const showCloudSkeleton = currentFilesystemTab === "cloud" && (zeroFoldersResult.type !== "complete" || zeroAllDocsResult.type !== "complete"); @@ -1331,8 +1349,8 @@ function AuthenticatedDocumentsSidebarBase({ Trust this workspace? - Local mode can read and edit files inside the folders you select. Continue only if - you trust this workspace and its contents. + Local mode can read and edit files inside the folders you select. Continue only if you + trust this workspace and its contents. {pendingLocalPath && ( diff --git a/surfsense_web/components/layout/ui/sidebar/LocalFilesystemBrowser.tsx b/surfsense_web/components/layout/ui/sidebar/LocalFilesystemBrowser.tsx index 6bfb1d3f1..19c47d605 100644 --- a/surfsense_web/components/layout/ui/sidebar/LocalFilesystemBrowser.tsx +++ b/surfsense_web/components/layout/ui/sidebar/LocalFilesystemBrowser.tsx @@ -141,7 +141,9 @@ export function LocalFilesystemBrowser({ }: LocalFilesystemBrowserProps) { const electronAPI = useElectronAPI(); const [rootStateMap, setRootStateMap] = useState>({}); - const [internalExpandedFolderKeys, setInternalExpandedFolderKeys] = useState>(new Set()); + const [internalExpandedFolderKeys, setInternalExpandedFolderKeys] = useState>( + new Set() + ); const [mountByRootKey, setMountByRootKey] = useState>(new Map()); const [mountStatus, setMountStatus] = useState("idle"); const [mountRefreshInFlight, setMountRefreshInFlight] = useState(false); @@ -188,10 +190,7 @@ export function LocalFilesystemBrowser({ } for (const { rootKey } of rootsToReload) { const nonce = reloadNonceByRoot[rootKey] ?? 0; - lastLoadedSignatureByRootRef.current.set( - rootKey, - `${searchSpaceId}:${rootKey}:${nonce}` - ); + lastLoadedSignatureByRootRef.current.set(rootKey, `${searchSpaceId}:${rootKey}:${nonce}`); } let cancelled = false; @@ -257,35 +256,37 @@ export function LocalFilesystemBrowser({ return; } - const unsubscribe = electronAPI.onAgentFilesystemTreeDirty((event: { - searchSpaceId: number | null; - reason: "watcher_event" | "safety_poll"; - rootPath: string; - changedPath: string | null; - timestamp: number; - }) => { - if ((event.searchSpaceId ?? null) !== (searchSpaceId ?? null)) { - return; + const unsubscribe = electronAPI.onAgentFilesystemTreeDirty( + (event: { + searchSpaceId: number | null; + reason: "watcher_event" | "safety_poll"; + rootPath: string; + changedPath: string | null; + timestamp: number; + }) => { + if ((event.searchSpaceId ?? null) !== (searchSpaceId ?? null)) { + return; + } + const eventRootKey = normalizeRootPathForLookup(event.rootPath, isWindowsPlatform); + const knownRootKeys = new Set( + rootPaths.map((rootPath) => normalizeRootPathForLookup(rootPath, isWindowsPlatform)) + ); + if (!knownRootKeys.has(eventRootKey)) { + setReloadNonceByRoot((prev) => { + const next = { ...prev }; + for (const rootKey of knownRootKeys) { + next[rootKey] = (prev[rootKey] ?? 0) + 1; + } + return next; + }); + return; + } + setReloadNonceByRoot((prev) => ({ + ...prev, + [eventRootKey]: (prev[eventRootKey] ?? 0) + 1, + })); } - const eventRootKey = normalizeRootPathForLookup(event.rootPath, isWindowsPlatform); - const knownRootKeys = new Set( - rootPaths.map((rootPath) => normalizeRootPathForLookup(rootPath, isWindowsPlatform)) - ); - if (!knownRootKeys.has(eventRootKey)) { - setReloadNonceByRoot((prev) => { - const next = { ...prev }; - for (const rootKey of knownRootKeys) { - next[rootKey] = (prev[rootKey] ?? 0) + 1; - } - return next; - }); - return; - } - setReloadNonceByRoot((prev) => ({ - ...prev, - [eventRootKey]: (prev[eventRootKey] ?? 0) + 1, - })); - }); + ); void electronAPI.startAgentFilesystemTreeWatch({ searchSpaceId, rootPaths, @@ -378,22 +379,25 @@ export function LocalFilesystemBrowser({ }); }, [rootPaths, rootStateMap, searchQuery]); - const toggleFolder = useCallback((folderKey: string) => { - const update = (prev: Set) => { - const next = new Set(prev); - if (next.has(folderKey)) { - next.delete(folderKey); - } else { - next.add(folderKey); + const toggleFolder = useCallback( + (folderKey: string) => { + const update = (prev: Set) => { + const next = new Set(prev); + if (next.has(folderKey)) { + next.delete(folderKey); + } else { + next.add(folderKey); + } + return next; + }; + if (onExpandedFolderKeysChange) { + onExpandedFolderKeysChange(update(effectiveExpandedFolderKeys)); + return; } - return next; - }; - if (onExpandedFolderKeysChange) { - onExpandedFolderKeysChange(update(effectiveExpandedFolderKeys)); - return; - } - setInternalExpandedFolderKeys(update); - }, [effectiveExpandedFolderKeys, onExpandedFolderKeysChange]); + setInternalExpandedFolderKeys(update); + }, + [effectiveExpandedFolderKeys, onExpandedFolderKeysChange] + ); const renderFolder = useCallback( (folder: LocalFolderNode, depth: number, mount: string) => { @@ -436,9 +440,7 @@ export function LocalFilesystemBrowser({ : undefined } className={`flex h-8 w-full items-center gap-1.5 rounded-md px-2 text-left text-sm transition-colors ${ - isOpenable - ? "hover:bg-muted/60" - : "cursor-not-allowed opacity-60" + isOpenable ? "hover:bg-muted/60" : "cursor-not-allowed opacity-60" }`} style={{ paddingInlineStart: `${(depth + 1) * 12 + 22}px` }} title={ @@ -528,7 +530,10 @@ export function LocalFilesystemBrowser({ } if (state.error) { return ( -
+

Failed to load local folder

{state.error}

diff --git a/surfsense_web/components/layout/ui/tabs/DocumentTabContent.tsx b/surfsense_web/components/layout/ui/tabs/DocumentTabContent.tsx index 77668a93d..ac5463873 100644 --- a/surfsense_web/components/layout/ui/tabs/DocumentTabContent.tsx +++ b/surfsense_web/components/layout/ui/tabs/DocumentTabContent.tsx @@ -308,9 +308,7 @@ export function DocumentTabContent({ documentId, searchSpaceId, title }: Documen } }} > - + Download .md diff --git a/surfsense_web/components/new-chat/model-selector.tsx b/surfsense_web/components/new-chat/model-selector.tsx index 3f5a5fa8c..9fe9dd8da 100644 --- a/surfsense_web/components/new-chat/model-selector.tsx +++ b/surfsense_web/components/new-chat/model-selector.tsx @@ -8,9 +8,9 @@ import { ChevronLeft, ChevronRight, ChevronUp, - Pencil, ImageIcon, Layers, + Pencil, Plus, ScanEye, Search, @@ -741,9 +741,7 @@ export function ModelSelector({
{!isMobile && ( @@ -769,9 +767,7 @@ export function ModelSelector({
diff --git a/surfsense_web/components/report-panel/report-panel.tsx b/surfsense_web/components/report-panel/report-panel.tsx index ede63d902..621cf13ce 100644 --- a/surfsense_web/components/report-panel/report-panel.tsx +++ b/surfsense_web/components/report-panel/report-panel.tsx @@ -398,7 +398,8 @@ export function ReportPanelContent({ ); - const editingActions = showReportEditingTier && + const editingActions = + showReportEditingTier && !isReadOnly && (isEditing ? ( <> diff --git a/surfsense_web/components/settings/agent-model-manager.tsx b/surfsense_web/components/settings/agent-model-manager.tsx index 988befdd0..a0b700c2d 100644 --- a/surfsense_web/components/settings/agent-model-manager.tsx +++ b/surfsense_web/components/settings/agent-model-manager.tsx @@ -1,15 +1,7 @@ "use client"; import { useAtomValue } from "jotai"; -import { - AlertCircle, - Dot, - FileText, - Info, - Pencil, - RefreshCw, - Trash2, -} from "lucide-react"; +import { AlertCircle, Dot, FileText, Info, Pencil, RefreshCw, Trash2 } from "lucide-react"; import { useMemo, useState } from "react"; import { membersAtom, myAccessAtom } from "@/atoms/members/members-query.atoms"; import { deleteNewLLMConfigMutationAtom } from "@/atoms/new-llm-config/new-llm-config-mutation.atoms"; diff --git a/surfsense_web/components/settings/roles-manager.tsx b/surfsense_web/components/settings/roles-manager.tsx index e7dadc20f..335cfc8a9 100644 --- a/surfsense_web/components/settings/roles-manager.tsx +++ b/surfsense_web/components/settings/roles-manager.tsx @@ -5,10 +5,8 @@ import { useAtomValue } from "jotai"; import { Bot, ChevronRight, - ScanEye, - Pencil, - FileText, Earth, + FileText, Image, Logs, type LucideIcon, @@ -16,11 +14,13 @@ import { MessageSquare, Mic, MoreHorizontal, - Unplug, + Pencil, + ScanEye, Settings, Shield, SlidersHorizontal, Trash2, + Unplug, Users, Video, } from "lucide-react"; @@ -462,9 +462,19 @@ function RolesContent({ return (
+ {/* biome-ignore lint/a11y/useSemanticElements: row contains nested interactive elements (DropdownMenu); using a )} - {sidebarDocs.length > 0 && ( - - )}
{!hasModelConfigured && (
diff --git a/surfsense_web/components/layout/ui/sidebar/DocumentsSidebar.tsx b/surfsense_web/components/layout/ui/sidebar/DocumentsSidebar.tsx index 00cc2d4ef..3c5a64b0e 100644 --- a/surfsense_web/components/layout/ui/sidebar/DocumentsSidebar.tsx +++ b/surfsense_web/components/layout/ui/sidebar/DocumentsSidebar.tsx @@ -24,7 +24,6 @@ import type React from "react"; import { useCallback, useEffect, useMemo, useRef, useState } from "react"; import { toast } from "sonner"; import { - sidebarMentionEventAtom, sidebarSelectedDocumentsAtom, } from "@/atoms/chat/mentioned-documents.atom"; import { connectorDialogOpenAtom } from "@/atoms/connector-dialog/connector-dialog.atoms"; @@ -416,7 +415,6 @@ function AuthenticatedDocumentsSidebarBase({ const { mutateAsync: deleteDocumentMutation } = useAtomValue(deleteDocumentMutationAtom); const [sidebarDocs, setSidebarDocs] = useAtom(sidebarSelectedDocumentsAtom); - const setSidebarMentionEvent = useSetAtom(sidebarMentionEventAtom); const mentionedDocIds = useMemo(() => new Set(sidebarDocs.map((d) => d.id)), [sidebarDocs]); // Folder state @@ -864,17 +862,6 @@ function AuthenticatedDocumentsSidebarBase({ const key = `${doc.document_type}:${doc.id}`; if (isMentioned) { setSidebarDocs((prev) => prev.filter((d) => `${d.document_type}:${d.id}` !== key)); - setSidebarMentionEvent({ - kind: "remove", - docs: [ - { - id: doc.id, - title: doc.title, - document_type: doc.document_type as DocumentTypeEnum, - }, - ], - nonce: Date.now(), - }); } else { setSidebarDocs((prev) => { if (prev.some((d) => `${d.document_type}:${d.id}` === key)) return prev; @@ -883,20 +870,9 @@ function AuthenticatedDocumentsSidebarBase({ { id: doc.id, title: doc.title, document_type: doc.document_type as DocumentTypeEnum }, ]; }); - setSidebarMentionEvent({ - kind: "add", - docs: [ - { - id: doc.id, - title: doc.title, - document_type: doc.document_type as DocumentTypeEnum, - }, - ], - nonce: Date.now(), - }); } }, - [setSidebarDocs, setSidebarMentionEvent] + [setSidebarDocs] ); const handleToggleFolderSelect = useCallback( @@ -918,14 +894,6 @@ function AuthenticatedDocumentsSidebarBase({ if (subtreeDocs.length === 0) return; if (selectAll) { - const existingKeys = new Set(sidebarDocs.map((d) => `${d.document_type}:${d.id}`)); - const docsToAdd = subtreeDocs - .filter((d) => !existingKeys.has(`${d.document_type}:${d.id}`)) - .map((d) => ({ - id: d.id, - title: d.title, - document_type: d.document_type as DocumentTypeEnum, - })); setSidebarDocs((prev) => { const existingDocKeys = new Set(prev.map((d) => `${d.document_type}:${d.id}`)); const newDocs = subtreeDocs @@ -937,35 +905,14 @@ function AuthenticatedDocumentsSidebarBase({ })); return newDocs.length > 0 ? [...prev, ...newDocs] : prev; }); - if (docsToAdd.length > 0) { - setSidebarMentionEvent({ - kind: "add", - docs: docsToAdd, - nonce: Date.now(), - }); - } } else { const keysToRemove = new Set(subtreeDocs.map((d) => `${d.document_type}:${d.id}`)); - const docsToRemove = sidebarDocs - .filter((d) => keysToRemove.has(`${d.document_type}:${d.id}`)) - .map((d) => ({ - id: d.id, - title: d.title, - document_type: d.document_type as DocumentTypeEnum, - })); setSidebarDocs((prev) => prev.filter((d) => !keysToRemove.has(`${d.document_type}:${d.id}`)) ); - if (docsToRemove.length > 0) { - setSidebarMentionEvent({ - kind: "remove", - docs: docsToRemove, - nonce: Date.now(), - }); - } } }, - [treeDocuments, foldersByParent, sidebarDocs, setSidebarDocs, setSidebarMentionEvent] + [treeDocuments, foldersByParent, setSidebarDocs] ); const searchFilteredDocuments = useMemo(() => { @@ -1626,7 +1573,6 @@ function AnonymousDocumentsSidebar({ const [search, setSearch] = useState(""); const [sidebarDocs, setSidebarDocs] = useAtom(sidebarSelectedDocumentsAtom); - const setSidebarMentionEvent = useSetAtom(sidebarMentionEventAtom); const mentionedDocIds = useMemo(() => new Set(sidebarDocs.map((d) => d.id)), [sidebarDocs]); const handleToggleChatMention = useCallback( @@ -1634,17 +1580,6 @@ function AnonymousDocumentsSidebar({ const key = `${doc.document_type}:${doc.id}`; if (isMentioned) { setSidebarDocs((prev) => prev.filter((d) => `${d.document_type}:${d.id}` !== key)); - setSidebarMentionEvent({ - kind: "remove", - docs: [ - { - id: doc.id, - title: doc.title, - document_type: doc.document_type as DocumentTypeEnum, - }, - ], - nonce: Date.now(), - }); } else { setSidebarDocs((prev) => { if (prev.some((d) => `${d.document_type}:${d.id}` === key)) return prev; @@ -1653,20 +1588,9 @@ function AnonymousDocumentsSidebar({ { id: doc.id, title: doc.title, document_type: doc.document_type as DocumentTypeEnum }, ]; }); - setSidebarMentionEvent({ - kind: "add", - docs: [ - { - id: doc.id, - title: doc.title, - document_type: doc.document_type as DocumentTypeEnum, - }, - ], - nonce: Date.now(), - }); } }, - [setSidebarDocs, setSidebarMentionEvent] + [setSidebarDocs] ); const uploadedDoc = anonMode.isAnonymous ? anonMode.uploadedDoc : null; From 294c719965f9e83867ec8831994bb0ae67caac29 Mon Sep 17 00:00:00 2001 From: Anish Sarkar <104695310+AnishSarkar22@users.noreply.github.com> Date: Tue, 28 Apr 2026 18:36:49 +0530 Subject: [PATCH 06/17] feat(mentions): implement user message rendering with mention chips for referenced documents --- .../components/assistant-ui/user-message.tsx | 94 +++++++++++++++---- 1 file changed, 78 insertions(+), 16 deletions(-) diff --git a/surfsense_web/components/assistant-ui/user-message.tsx b/surfsense_web/components/assistant-ui/user-message.tsx index 86863a501..fb7212119 100644 --- a/surfsense_web/components/assistant-ui/user-message.tsx +++ b/surfsense_web/components/assistant-ui/user-message.tsx @@ -1,11 +1,12 @@ import { ActionBarPrimitive, AuiIf, MessagePrimitive, useAuiState } from "@assistant-ui/react"; import { useAtomValue } from "jotai"; -import { CheckIcon, CopyIcon, FileText, Pencil } from "lucide-react"; +import { CheckIcon, CopyIcon, Pencil } from "lucide-react"; import Image from "next/image"; import { type FC, useState } from "react"; import { currentThreadAtom } from "@/atoms/chat/current-thread.atom"; import { messageDocumentsMapAtom } from "@/atoms/chat/mentioned-documents.atom"; import { TooltipIconButton } from "@/components/assistant-ui/tooltip-icon-button"; +import { getConnectorIcon } from "@/contracts/enums/connectorIcons"; interface AuthorMetadata { displayName: string | null; @@ -48,6 +49,19 @@ const UserAvatar: FC = ({ displayName, avatarUrl }) => { export const UserMessage: FC = () => { const messageId = useAuiState(({ message }) => message?.id); + const messageText = useAuiState(({ message }) => + (message?.content ?? []) + .map((part) => + typeof part === "object" && + part !== null && + "type" in part && + (part as { type?: string }).type === "text" && + "text" in part + ? String((part as { text?: string }).text ?? "") + : "" + ) + .join("") + ); const messageDocumentsMap = useAtomValue(messageDocumentsMapAtom); const mentionedDocs = messageId ? messageDocumentsMap[messageId] : undefined; const metadata = useAuiState(({ message }) => message?.metadata); @@ -63,22 +77,12 @@ export const UserMessage: FC = () => {
- {mentionedDocs && mentionedDocs.length > 0 && ( -
- {mentionedDocs?.map((doc) => ( - - - {doc.title} - - ))} -
- )}
- + {mentionedDocs && mentionedDocs.length > 0 ? ( + + ) : ( + + )}
@@ -95,6 +99,64 @@ export const UserMessage: FC = () => { ); }; +const UserMessageWithMentionChips: FC<{ + text: string; + mentionedDocs: { id: number; title: string; document_type: string }[]; +}> = ({ text, mentionedDocs }) => { + type Segment = + | { type: "text"; value: string; start: number } + | { type: "mention"; doc: { id: number; title: string; document_type: string }; start: number }; + + const tokens = mentionedDocs + .map((doc) => ({ doc, token: `@${doc.title}` })) + .sort((a, b) => b.token.length - a.token.length); + + const segments: Segment[] = []; + let i = 0; + let buffer = ""; + let bufferStart = 0; + while (i < text.length) { + const tokenMatch = tokens.find(({ token }) => text.startsWith(token, i)); + if (tokenMatch) { + if (buffer) { + segments.push({ type: "text", value: buffer, start: bufferStart }); + buffer = ""; + } + segments.push({ type: "mention", doc: tokenMatch.doc, start: i }); + i += tokenMatch.token.length; + bufferStart = i; + continue; + } + if (!buffer) bufferStart = i; + buffer += text[i]; + i += 1; + } + if (buffer) { + segments.push({ type: "text", value: buffer, start: bufferStart }); + } + + return ( + + {segments.map((segment) => + segment.type === "text" ? ( + {segment.value} + ) : ( + + + {getConnectorIcon(segment.doc.document_type ?? "UNKNOWN", "h-3 w-3")} + + {segment.doc.title} + + ) + )} + + ); +}; + const UserActionBar: FC = () => { const isThreadRunning = useAuiState(({ thread }) => thread.isRunning); From 282510f93ce16bf7b76779046960baa44ed7d9fb Mon Sep 17 00:00:00 2001 From: Anish Sarkar <104695310+AnishSarkar22@users.noreply.github.com> Date: Tue, 28 Apr 2026 18:47:57 +0530 Subject: [PATCH 07/17] feat(mentions): add syncEditorState function to manage editor state and mentioned documents --- .../assistant-ui/inline-mention-editor.tsx | 57 ++++++++++--------- 1 file changed, 31 insertions(+), 26 deletions(-) diff --git a/surfsense_web/components/assistant-ui/inline-mention-editor.tsx b/surfsense_web/components/assistant-ui/inline-mention-editor.tsx index 81d6cbd77..e75a840c0 100644 --- a/surfsense_web/components/assistant-ui/inline-mention-editor.tsx +++ b/surfsense_web/components/assistant-ui/inline-mention-editor.tsx @@ -236,6 +236,19 @@ export const InlineMentionEditor = forwardRef) => { + const docs = docsOverride + ? Array.from(docsOverride.values()) + : Array.from(mentionedDocs.values()); + const text = getText(); + const empty = text.length === 0 && docs.length === 0; + setIsEmpty(empty); + onChange?.(text, docs); + }, + [getText, mentionedDocs, onChange] + ); + // Create a chip element for a document const createChipElement = useCallback( (doc: MentionedDocument): HTMLSpanElement => { @@ -275,6 +288,7 @@ export const InlineMentionEditor = forwardRef { const next = new Map(prev); next.delete(docKey); + syncEditorState(next); return next; }); onDocumentRemove?.(doc.id, doc.document_type); @@ -319,7 +333,7 @@ export const InlineMentionEditor = forwardRef new Map(prev).set(docKey, mentionDoc)); + const nextDocs = new Map(mentionedDocs); + nextDocs.set(docKey, mentionDoc); // Find and remove the @query text const selection = window.getSelection(); @@ -436,25 +452,16 @@ export const InlineMentionEditor = forwardRef { - onChange(getText(), getMentionedDocuments()); - }, 0); - } + syncEditorState(nextDocs); }, [ createChipElement, focusAtEnd, - getText, - getMentionedDocuments, isSelectionInsideEditor, - onChange, + mentionedDocs, rememberSelection, restoreRememberedSelection, + syncEditorState, ] ); @@ -462,22 +469,21 @@ export const InlineMentionEditor = forwardRef { if (editorRef.current) { editorRef.current.innerHTML = ""; - setIsEmpty(true); - setMentionedDocs(new Map()); + const emptyDocs = new Map(); + setMentionedDocs(emptyDocs); + syncEditorState(emptyDocs); } - }, []); + }, [syncEditorState]); // Replace editor content with plain text and place cursor at end const setText = useCallback( (text: string) => { if (!editorRef.current) return; editorRef.current.innerText = text; - const empty = text.length === 0; - setIsEmpty(empty); - onChange?.(text, Array.from(mentionedDocs.values())); + syncEditorState(); focusAtEnd(); }, - [focusAtEnd, onChange, mentionedDocs] + [focusAtEnd, syncEditorState] ); const setDocumentChipStatus = useCallback( @@ -538,14 +544,11 @@ export const InlineMentionEditor = forwardRef { const next = new Map(prev); next.delete(chipKey); + syncEditorState(next); return next; }); - - const text = getText(); - const empty = text.length === 0 && mentionedDocs.size <= 1; - setIsEmpty(empty); }, - [getText, mentionedDocs.size] + [syncEditorState] ); // Expose methods via ref @@ -697,6 +700,7 @@ export const InlineMentionEditor = forwardRef { const next = new Map(prev); next.delete(chipKey); + syncEditorState(next); return next; }); // Notify parent that a document was removed @@ -734,6 +738,7 @@ export const InlineMentionEditor = forwardRef { const next = new Map(prev); next.delete(chipKey); + syncEditorState(next); return next; }); // Notify parent that a document was removed @@ -745,7 +750,7 @@ export const InlineMentionEditor = forwardRef Date: Tue, 28 Apr 2026 09:22:19 -0700 Subject: [PATCH 08/17] feat: updated agent harness --- surfsense_backend/.env.example | 39 + .../versions/130_add_agent_action_log.py | 94 ++ .../versions/131_add_document_revisions.py | 119 ++ .../132_add_agent_permission_rules.py | 82 ++ .../app/agents/new_chat/chat_deepagent.py | 459 +++++++- .../app/agents/new_chat/errors.py | 95 ++ .../app/agents/new_chat/feature_flags.py | 188 +++ .../agents/new_chat/middleware/__init__.py | 42 + .../agents/new_chat/middleware/action_log.py | 294 +++++ .../agents/new_chat/middleware/busy_mutex.py | 231 ++++ .../agents/new_chat/middleware/compaction.py | 253 +++++ .../new_chat/middleware/context_editing.py | 349 ++++++ .../new_chat/middleware/dedup_tool_calls.py | 123 +- .../agents/new_chat/middleware/doom_loop.py | 228 ++++ .../new_chat/middleware/knowledge_search.py | 79 +- .../new_chat/middleware/noop_injection.py | 133 +++ .../agents/new_chat/middleware/otel_span.py | 202 ++++ .../agents/new_chat/middleware/permission.py | 344 ++++++ .../agents/new_chat/middleware/retry_after.py | 245 ++++ .../new_chat/middleware/safe_summarization.py | 123 -- .../new_chat/middleware/skills_backends.py | 332 ++++++ .../new_chat/middleware/tool_call_repair.py | 190 ++++ .../app/agents/new_chat/permissions.py | 204 ++++ .../app/agents/new_chat/plugin_loader.py | 157 +++ .../app/agents/new_chat/plugins/__init__.py | 6 + .../new_chat/plugins/year_substituter.py | 87 ++ .../app/agents/new_chat/prompts/__init__.py | 7 + .../agents/new_chat/prompts/base/__init__.py | 1 + .../new_chat/prompts/base/agent_private.md | 7 + .../new_chat/prompts/base/agent_team.md | 9 + .../new_chat/prompts/base/citations_off.md | 16 + .../new_chat/prompts/base/citations_on.md | 90 ++ .../prompts/base/kb_only_policy_private.md | 15 + .../prompts/base/kb_only_policy_team.md | 15 + .../prompts/base/memory_protocol_private.md | 6 + .../prompts/base/memory_protocol_team.md | 6 + .../prompts/base/parameter_resolution.md | 39 + .../prompts/base/tool_routing_private.md | 16 + .../prompts/base/tool_routing_team.md | 16 + .../app/agents/new_chat/prompts/composer.py | 359 ++++++ .../new_chat/prompts/examples/__init__.py | 1 + .../prompts/examples/generate_image.md | 12 + .../prompts/examples/generate_podcast.md | 7 + .../prompts/examples/generate_report.md | 13 + .../prompts/examples/generate_resume.md | 19 + .../examples/generate_video_presentation.md | 7 + .../prompts/examples/scrape_webpage.md | 13 + .../prompts/examples/search_surfsense_docs.md | 9 + .../prompts/examples/update_memory_private.md | 16 + .../prompts/examples/update_memory_team.md | 7 + .../new_chat/prompts/examples/web_search.md | 8 + .../new_chat/prompts/providers/__init__.py | 1 + .../new_chat/prompts/providers/anthropic.md | 5 + .../new_chat/prompts/providers/default.md | 1 + .../new_chat/prompts/providers/google.md | 4 + .../prompts/providers/openai_classic.md | 5 + .../prompts/providers/openai_reasoning.md | 5 + .../new_chat/prompts/routing/__init__.py | 1 + .../agents/new_chat/prompts/routing/jira.md | 1 + .../agents/new_chat/prompts/routing/linear.md | 1 + .../agents/new_chat/prompts/routing/slack.md | 1 + .../agents/new_chat/prompts/tools/__init__.py | 1 + .../new_chat/prompts/tools/_preamble.md | 6 + .../new_chat/prompts/tools/generate_image.md | 11 + .../prompts/tools/generate_podcast.md | 15 + .../new_chat/prompts/tools/generate_report.md | 39 + .../new_chat/prompts/tools/generate_resume.md | 30 + .../tools/generate_video_presentation.md | 9 + .../new_chat/prompts/tools/scrape_webpage.md | 30 + .../prompts/tools/search_surfsense_docs.md | 7 + .../prompts/tools/update_memory_private.md | 31 + .../prompts/tools/update_memory_team.md | 26 + .../new_chat/prompts/tools/web_search.md | 18 + .../app/agents/new_chat/skills/__init__.py | 7 + .../new_chat/skills/builtin/__init__.py | 1 + .../skills/builtin/email-drafting/SKILL.md | 25 + .../skills/builtin/kb-research/SKILL.md | 23 + .../skills/builtin/meeting-prep/SKILL.md | 22 + .../skills/builtin/report-writing/SKILL.md | 23 + .../skills/builtin/slack-summary/SKILL.md | 26 + .../app/agents/new_chat/subagents/__init__.py | 26 + .../app/agents/new_chat/subagents/config.py | 427 +++++++ .../app/agents/new_chat/system_prompt.py | 1003 ++--------------- .../app/agents/new_chat/tools/invalid_tool.py | 52 + .../app/agents/new_chat/tools/registry.py | 50 + surfsense_backend/app/db.py | 196 ++++ .../app/observability/__init__.py | 7 + surfsense_backend/app/observability/otel.py | 319 ++++++ surfsense_backend/app/routes/__init__.py | 10 + .../app/routes/agent_action_log_route.py | 186 +++ .../app/routes/agent_flags_route.py | 71 ++ .../app/routes/agent_permissions_route.py | 280 +++++ .../app/routes/agent_revert_route.py | 122 ++ .../app/services/revert_service.py | 279 +++++ surfsense_backend/app/utils/async_retry.py | 2 +- .../tests/integration/harness/__init__.py | 146 +++ .../harness/test_scripted_harness.py | 53 + .../tests/unit/agents/__init__.py | 1 + .../tests/unit/agents/new_chat/__init__.py | 1 + .../unit/agents/new_chat/prompts/__init__.py | 1 + .../agents/new_chat/prompts/test_composer.py | 201 ++++ .../unit/agents/new_chat/test_action_log.py | 311 +++++ .../unit/agents/new_chat/test_busy_mutex.py | 90 ++ .../unit/agents/new_chat/test_compaction.py | 107 ++ .../agents/new_chat/test_context_editing.py | 107 ++ .../agents/new_chat/test_dedup_tool_calls.py | 132 +++ .../test_default_permissions_layering.py | 128 +++ .../unit/agents/new_chat/test_doom_loop.py | 99 ++ .../agents/new_chat/test_feature_flags.py | 120 ++ .../agents/new_chat/test_noop_injection.py | 119 ++ .../unit/agents/new_chat/test_otel_span.py | 195 ++++ .../new_chat/test_permission_middleware.py | 116 ++ .../unit/agents/new_chat/test_permissions.py | 111 ++ .../agents/new_chat/test_plugin_loader.py | 187 +++ .../unit/agents/new_chat/test_retry_after.py | 107 ++ .../agents/new_chat/test_skills_backends.py | 242 ++++ .../new_chat/test_specialized_subagents.py | 338 ++++++ .../agents/new_chat/test_tool_call_repair.py | 103 ++ .../middleware/test_dedup_hitl_tool_calls.py | 31 +- .../tests/unit/observability/__init__.py | 1 + .../tests/unit/observability/test_otel.py | 84 ++ .../unit/services/test_revert_service.py | 56 + .../components/AgentPermissionsContent.tsx | 451 ++++++++ .../components/AgentStatusContent.tsx | 309 +++++ .../atoms/agent/action-log-sheet.atom.ts | 19 + .../atoms/agent/agent-flags-query.atom.ts | 17 + .../agent-action-log/action-log-button.tsx | 50 + .../agent-action-log/action-log-item.tsx | 215 ++++ .../agent-action-log/action-log-sheet.tsx | 185 +++ .../components/assistant-ui/markdown-text.tsx | 16 +- .../components/assistant-ui/tool-fallback.tsx | 7 + .../layout/providers/LayoutDataProvider.tsx | 4 + .../components/layout/ui/header/Header.tsx | 2 + surfsense_web/components/markdown-viewer.tsx | 6 +- .../settings/user-settings-dialog.tsx | 28 + .../components/tool-ui/doom-loop-approval.tsx | 187 +++ .../lib/apis/agent-actions-api.service.ts | 64 ++ .../lib/apis/agent-flags-api.service.ts | 40 + .../lib/apis/agent-permissions-api.service.ts | 90 ++ 139 files changed, 12583 insertions(+), 1111 deletions(-) create mode 100644 surfsense_backend/alembic/versions/130_add_agent_action_log.py create mode 100644 surfsense_backend/alembic/versions/131_add_document_revisions.py create mode 100644 surfsense_backend/alembic/versions/132_add_agent_permission_rules.py create mode 100644 surfsense_backend/app/agents/new_chat/errors.py create mode 100644 surfsense_backend/app/agents/new_chat/feature_flags.py create mode 100644 surfsense_backend/app/agents/new_chat/middleware/action_log.py create mode 100644 surfsense_backend/app/agents/new_chat/middleware/busy_mutex.py create mode 100644 surfsense_backend/app/agents/new_chat/middleware/compaction.py create mode 100644 surfsense_backend/app/agents/new_chat/middleware/context_editing.py create mode 100644 surfsense_backend/app/agents/new_chat/middleware/doom_loop.py create mode 100644 surfsense_backend/app/agents/new_chat/middleware/noop_injection.py create mode 100644 surfsense_backend/app/agents/new_chat/middleware/otel_span.py create mode 100644 surfsense_backend/app/agents/new_chat/middleware/permission.py create mode 100644 surfsense_backend/app/agents/new_chat/middleware/retry_after.py delete mode 100644 surfsense_backend/app/agents/new_chat/middleware/safe_summarization.py create mode 100644 surfsense_backend/app/agents/new_chat/middleware/skills_backends.py create mode 100644 surfsense_backend/app/agents/new_chat/middleware/tool_call_repair.py create mode 100644 surfsense_backend/app/agents/new_chat/permissions.py create mode 100644 surfsense_backend/app/agents/new_chat/plugin_loader.py create mode 100644 surfsense_backend/app/agents/new_chat/plugins/__init__.py create mode 100644 surfsense_backend/app/agents/new_chat/plugins/year_substituter.py create mode 100644 surfsense_backend/app/agents/new_chat/prompts/__init__.py create mode 100644 surfsense_backend/app/agents/new_chat/prompts/base/__init__.py create mode 100644 surfsense_backend/app/agents/new_chat/prompts/base/agent_private.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/base/agent_team.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/base/citations_off.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/base/citations_on.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/base/kb_only_policy_private.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/base/kb_only_policy_team.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/base/memory_protocol_private.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/base/memory_protocol_team.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/base/parameter_resolution.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/base/tool_routing_private.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/base/tool_routing_team.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/composer.py create mode 100644 surfsense_backend/app/agents/new_chat/prompts/examples/__init__.py create mode 100644 surfsense_backend/app/agents/new_chat/prompts/examples/generate_image.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/examples/generate_podcast.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/examples/generate_report.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/examples/generate_resume.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/examples/generate_video_presentation.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/examples/scrape_webpage.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/examples/search_surfsense_docs.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/examples/update_memory_private.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/examples/update_memory_team.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/examples/web_search.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/providers/__init__.py create mode 100644 surfsense_backend/app/agents/new_chat/prompts/providers/anthropic.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/providers/default.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/providers/google.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/providers/openai_classic.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/providers/openai_reasoning.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/routing/__init__.py create mode 100644 surfsense_backend/app/agents/new_chat/prompts/routing/jira.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/routing/linear.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/routing/slack.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/tools/__init__.py create mode 100644 surfsense_backend/app/agents/new_chat/prompts/tools/_preamble.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/tools/generate_image.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/tools/generate_podcast.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/tools/generate_report.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/tools/generate_resume.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/tools/generate_video_presentation.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/tools/scrape_webpage.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/tools/search_surfsense_docs.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/tools/update_memory_private.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/tools/update_memory_team.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/tools/web_search.md create mode 100644 surfsense_backend/app/agents/new_chat/skills/__init__.py create mode 100644 surfsense_backend/app/agents/new_chat/skills/builtin/__init__.py create mode 100644 surfsense_backend/app/agents/new_chat/skills/builtin/email-drafting/SKILL.md create mode 100644 surfsense_backend/app/agents/new_chat/skills/builtin/kb-research/SKILL.md create mode 100644 surfsense_backend/app/agents/new_chat/skills/builtin/meeting-prep/SKILL.md create mode 100644 surfsense_backend/app/agents/new_chat/skills/builtin/report-writing/SKILL.md create mode 100644 surfsense_backend/app/agents/new_chat/skills/builtin/slack-summary/SKILL.md create mode 100644 surfsense_backend/app/agents/new_chat/subagents/__init__.py create mode 100644 surfsense_backend/app/agents/new_chat/subagents/config.py create mode 100644 surfsense_backend/app/agents/new_chat/tools/invalid_tool.py create mode 100644 surfsense_backend/app/observability/__init__.py create mode 100644 surfsense_backend/app/observability/otel.py create mode 100644 surfsense_backend/app/routes/agent_action_log_route.py create mode 100644 surfsense_backend/app/routes/agent_flags_route.py create mode 100644 surfsense_backend/app/routes/agent_permissions_route.py create mode 100644 surfsense_backend/app/routes/agent_revert_route.py create mode 100644 surfsense_backend/app/services/revert_service.py create mode 100644 surfsense_backend/tests/integration/harness/__init__.py create mode 100644 surfsense_backend/tests/integration/harness/test_scripted_harness.py create mode 100644 surfsense_backend/tests/unit/agents/__init__.py create mode 100644 surfsense_backend/tests/unit/agents/new_chat/__init__.py create mode 100644 surfsense_backend/tests/unit/agents/new_chat/prompts/__init__.py create mode 100644 surfsense_backend/tests/unit/agents/new_chat/prompts/test_composer.py create mode 100644 surfsense_backend/tests/unit/agents/new_chat/test_action_log.py create mode 100644 surfsense_backend/tests/unit/agents/new_chat/test_busy_mutex.py create mode 100644 surfsense_backend/tests/unit/agents/new_chat/test_compaction.py create mode 100644 surfsense_backend/tests/unit/agents/new_chat/test_context_editing.py create mode 100644 surfsense_backend/tests/unit/agents/new_chat/test_dedup_tool_calls.py create mode 100644 surfsense_backend/tests/unit/agents/new_chat/test_default_permissions_layering.py create mode 100644 surfsense_backend/tests/unit/agents/new_chat/test_doom_loop.py create mode 100644 surfsense_backend/tests/unit/agents/new_chat/test_feature_flags.py create mode 100644 surfsense_backend/tests/unit/agents/new_chat/test_noop_injection.py create mode 100644 surfsense_backend/tests/unit/agents/new_chat/test_otel_span.py create mode 100644 surfsense_backend/tests/unit/agents/new_chat/test_permission_middleware.py create mode 100644 surfsense_backend/tests/unit/agents/new_chat/test_permissions.py create mode 100644 surfsense_backend/tests/unit/agents/new_chat/test_plugin_loader.py create mode 100644 surfsense_backend/tests/unit/agents/new_chat/test_retry_after.py create mode 100644 surfsense_backend/tests/unit/agents/new_chat/test_skills_backends.py create mode 100644 surfsense_backend/tests/unit/agents/new_chat/test_specialized_subagents.py create mode 100644 surfsense_backend/tests/unit/agents/new_chat/test_tool_call_repair.py create mode 100644 surfsense_backend/tests/unit/observability/__init__.py create mode 100644 surfsense_backend/tests/unit/observability/test_otel.py create mode 100644 surfsense_backend/tests/unit/services/test_revert_service.py create mode 100644 surfsense_web/app/dashboard/[search_space_id]/user-settings/components/AgentPermissionsContent.tsx create mode 100644 surfsense_web/app/dashboard/[search_space_id]/user-settings/components/AgentStatusContent.tsx create mode 100644 surfsense_web/atoms/agent/action-log-sheet.atom.ts create mode 100644 surfsense_web/atoms/agent/agent-flags-query.atom.ts create mode 100644 surfsense_web/components/agent-action-log/action-log-button.tsx create mode 100644 surfsense_web/components/agent-action-log/action-log-item.tsx create mode 100644 surfsense_web/components/agent-action-log/action-log-sheet.tsx create mode 100644 surfsense_web/components/tool-ui/doom-loop-approval.tsx create mode 100644 surfsense_web/lib/apis/agent-actions-api.service.ts create mode 100644 surfsense_web/lib/apis/agent-flags-api.service.ts create mode 100644 surfsense_web/lib/apis/agent-permissions-api.service.ts diff --git a/surfsense_backend/.env.example b/surfsense_backend/.env.example index 86bac0aaf..e133a2bc5 100644 --- a/surfsense_backend/.env.example +++ b/surfsense_backend/.env.example @@ -247,3 +247,42 @@ LANGSMITH_TRACING=true LANGSMITH_ENDPOINT=https://api.smith.langchain.com LANGSMITH_API_KEY=lsv2_pt_..... LANGSMITH_PROJECT=surfsense + + +# ============================================================================= +# OPTIONAL: New-chat agent feature flags (OpenCode-port) +# ============================================================================= +# Master kill-switch — when true, every flag below is forced OFF. +# SURFSENSE_DISABLE_NEW_AGENT_STACK=false + +# Tier 1 — Agent quality +# SURFSENSE_ENABLE_CONTEXT_EDITING=false +# SURFSENSE_ENABLE_COMPACTION_V2=false +# SURFSENSE_ENABLE_RETRY_AFTER=false +# SURFSENSE_ENABLE_MODEL_FALLBACK=false +# SURFSENSE_ENABLE_MODEL_CALL_LIMIT=false +# SURFSENSE_ENABLE_TOOL_CALL_LIMIT=false +# SURFSENSE_ENABLE_TOOL_CALL_REPAIR=false +# SURFSENSE_ENABLE_DOOM_LOOP=false # leave OFF until UI handles permission='doom_loop' + +# Tier 2 — Safety +# SURFSENSE_ENABLE_PERMISSION=false +# SURFSENSE_ENABLE_BUSY_MUTEX=false +# SURFSENSE_ENABLE_LLM_TOOL_SELECTOR=false # adds a per-turn LLM call + +# Tier 3b — Observability (also requires OTEL_EXPORTER_OTLP_ENDPOINT) +# SURFSENSE_ENABLE_OTEL=false + +# Tier 4 — Skills + subagents +# SURFSENSE_ENABLE_SKILLS=false +# SURFSENSE_ENABLE_SPECIALIZED_SUBAGENTS=false +# SURFSENSE_ENABLE_KB_PLANNER_RUNNABLE=false + +# Tier 5 — Snapshot / revert +# SURFSENSE_ENABLE_ACTION_LOG=false +# SURFSENSE_ENABLE_REVERT_ROUTE=false # Backend-only; flip when UI ships + +# Tier 6 — Plugins +# SURFSENSE_ENABLE_PLUGIN_LOADER=false +# Comma-separated allowlist of plugin entry-point names +# SURFSENSE_ALLOWED_PLUGINS=year_substituter diff --git a/surfsense_backend/alembic/versions/130_add_agent_action_log.py b/surfsense_backend/alembic/versions/130_add_agent_action_log.py new file mode 100644 index 000000000..5793988cb --- /dev/null +++ b/surfsense_backend/alembic/versions/130_add_agent_action_log.py @@ -0,0 +1,94 @@ +"""130_add_agent_action_log + +Revision ID: 130 +Revises: 129 +Create Date: 2026-04-28 + +Tier 5.2 in the OpenCode-port plan. Adds the append-only ``agent_action_log`` +table that :class:`ActionLogMiddleware` writes to after every tool call. +""" + +from __future__ import annotations + +from collections.abc import Sequence + +import sqlalchemy as sa +from sqlalchemy.dialects import postgresql + +from alembic import op + +revision: str = "130" +down_revision: str | None = "129" +branch_labels: str | Sequence[str] | None = None +depends_on: str | Sequence[str] | None = None + + +def upgrade() -> None: + op.create_table( + "agent_action_log", + sa.Column("id", sa.Integer(), primary_key=True, index=True), + sa.Column( + "thread_id", + sa.Integer(), + sa.ForeignKey("new_chat_threads.id", ondelete="CASCADE"), + nullable=False, + index=True, + ), + sa.Column( + "user_id", + postgresql.UUID(as_uuid=True), + sa.ForeignKey("user.id", ondelete="SET NULL"), + nullable=True, + index=True, + ), + sa.Column( + "search_space_id", + sa.Integer(), + sa.ForeignKey("searchspaces.id", ondelete="CASCADE"), + nullable=False, + index=True, + ), + sa.Column("turn_id", sa.String(length=64), nullable=True, index=True), + sa.Column("message_id", sa.String(length=128), nullable=True, index=True), + sa.Column("tool_name", sa.String(length=255), nullable=False, index=True), + sa.Column("args", postgresql.JSONB(astext_type=sa.Text()), nullable=True), + sa.Column("result_id", sa.String(length=255), nullable=True), + sa.Column( + "reversible", + sa.Boolean(), + nullable=False, + server_default=sa.text("false"), + ), + sa.Column( + "reverse_descriptor", + postgresql.JSONB(astext_type=sa.Text()), + nullable=True, + ), + sa.Column("error", postgresql.JSONB(astext_type=sa.Text()), nullable=True), + sa.Column( + "reverse_of", + sa.Integer(), + sa.ForeignKey("agent_action_log.id", ondelete="SET NULL"), + nullable=True, + index=True, + ), + sa.Column( + "created_at", + sa.TIMESTAMP(timezone=True), + nullable=False, + server_default=sa.text("(now() AT TIME ZONE 'utc')"), + index=True, + ), + ) + op.create_index( + "ix_agent_action_log_thread_created", + "agent_action_log", + ["thread_id", "created_at"], + ) + + +def downgrade() -> None: + op.drop_index( + "ix_agent_action_log_thread_created", table_name="agent_action_log" + ) + op.drop_table("agent_action_log") diff --git a/surfsense_backend/alembic/versions/131_add_document_revisions.py b/surfsense_backend/alembic/versions/131_add_document_revisions.py new file mode 100644 index 000000000..46c6991b6 --- /dev/null +++ b/surfsense_backend/alembic/versions/131_add_document_revisions.py @@ -0,0 +1,119 @@ +"""131_add_document_revisions + +Revision ID: 131 +Revises: 130 +Create Date: 2026-04-28 + +Tier 5.1 in the OpenCode-port plan. Adds two snapshot tables: + +* ``document_revisions``: pre-mutation snapshot of NOTE/FILE/EXTENSION docs. +* ``folder_revisions``: pre-mutation snapshot of folder mkdir/move/delete. + +Both are written by :class:`KnowledgeBasePersistenceMiddleware` ahead of +state-changing tool calls and consumed by ``revert_service.revert_action``. +""" + +from __future__ import annotations + +from collections.abc import Sequence + +import sqlalchemy as sa +from sqlalchemy.dialects import postgresql + +from alembic import op + +revision: str = "131" +down_revision: str | None = "130" +branch_labels: str | Sequence[str] | None = None +depends_on: str | Sequence[str] | None = None + + +def upgrade() -> None: + op.create_table( + "document_revisions", + sa.Column("id", sa.Integer(), primary_key=True, index=True), + sa.Column( + "document_id", + sa.Integer(), + sa.ForeignKey("documents.id", ondelete="CASCADE"), + nullable=False, + index=True, + ), + sa.Column( + "search_space_id", + sa.Integer(), + sa.ForeignKey("searchspaces.id", ondelete="CASCADE"), + nullable=False, + index=True, + ), + sa.Column("content_before", sa.Text(), nullable=True), + sa.Column("title_before", sa.String(), nullable=True), + sa.Column("folder_id_before", sa.Integer(), nullable=True), + sa.Column( + "chunks_before", postgresql.JSONB(astext_type=sa.Text()), nullable=True + ), + sa.Column( + "metadata_before", postgresql.JSONB(astext_type=sa.Text()), nullable=True + ), + sa.Column( + "created_by_turn_id", sa.String(length=64), nullable=True, index=True + ), + sa.Column( + "agent_action_id", + sa.Integer(), + sa.ForeignKey("agent_action_log.id", ondelete="SET NULL"), + nullable=True, + index=True, + ), + sa.Column( + "created_at", + sa.TIMESTAMP(timezone=True), + nullable=False, + server_default=sa.text("(now() AT TIME ZONE 'utc')"), + index=True, + ), + ) + + op.create_table( + "folder_revisions", + sa.Column("id", sa.Integer(), primary_key=True, index=True), + sa.Column( + "folder_id", + sa.Integer(), + sa.ForeignKey("folders.id", ondelete="CASCADE"), + nullable=False, + index=True, + ), + sa.Column( + "search_space_id", + sa.Integer(), + sa.ForeignKey("searchspaces.id", ondelete="CASCADE"), + nullable=False, + index=True, + ), + sa.Column("name_before", sa.String(length=255), nullable=True), + sa.Column("parent_id_before", sa.Integer(), nullable=True), + sa.Column("position_before", sa.String(length=50), nullable=True), + sa.Column( + "created_by_turn_id", sa.String(length=64), nullable=True, index=True + ), + sa.Column( + "agent_action_id", + sa.Integer(), + sa.ForeignKey("agent_action_log.id", ondelete="SET NULL"), + nullable=True, + index=True, + ), + sa.Column( + "created_at", + sa.TIMESTAMP(timezone=True), + nullable=False, + server_default=sa.text("(now() AT TIME ZONE 'utc')"), + index=True, + ), + ) + + +def downgrade() -> None: + op.drop_table("folder_revisions") + op.drop_table("document_revisions") diff --git a/surfsense_backend/alembic/versions/132_add_agent_permission_rules.py b/surfsense_backend/alembic/versions/132_add_agent_permission_rules.py new file mode 100644 index 000000000..0e81eacb5 --- /dev/null +++ b/surfsense_backend/alembic/versions/132_add_agent_permission_rules.py @@ -0,0 +1,82 @@ +"""132_add_agent_permission_rules + +Revision ID: 132 +Revises: 131 +Create Date: 2026-04-28 + +Tier 2.1 in the OpenCode-port plan. Adds the persistent ``agent_permission_rules`` +table consumed by :class:`PermissionMiddleware` at agent build time. Rules +can be scoped at search-space (``user_id`` / ``thread_id`` NULL), +user-wide (``user_id`` set, ``thread_id`` NULL), or per-thread +(``thread_id`` set). +""" + +from __future__ import annotations + +from collections.abc import Sequence + +import sqlalchemy as sa +from sqlalchemy.dialects import postgresql + +from alembic import op + +revision: str = "132" +down_revision: str | None = "131" +branch_labels: str | Sequence[str] | None = None +depends_on: str | Sequence[str] | None = None + + +def upgrade() -> None: + op.create_table( + "agent_permission_rules", + sa.Column("id", sa.Integer(), primary_key=True, index=True), + sa.Column( + "search_space_id", + sa.Integer(), + sa.ForeignKey("searchspaces.id", ondelete="CASCADE"), + nullable=False, + index=True, + ), + sa.Column( + "user_id", + postgresql.UUID(as_uuid=True), + sa.ForeignKey("user.id", ondelete="CASCADE"), + nullable=True, + index=True, + ), + sa.Column( + "thread_id", + sa.Integer(), + sa.ForeignKey("new_chat_threads.id", ondelete="CASCADE"), + nullable=True, + index=True, + ), + sa.Column("permission", sa.String(length=255), nullable=False), + sa.Column( + "pattern", + sa.String(length=255), + nullable=False, + server_default="*", + ), + sa.Column("action", sa.String(length=16), nullable=False), + sa.Column( + "created_at", + sa.TIMESTAMP(timezone=True), + nullable=False, + server_default=sa.text("(now() AT TIME ZONE 'utc')"), + index=True, + ), + sa.UniqueConstraint( + "search_space_id", + "user_id", + "thread_id", + "permission", + "pattern", + "action", + name="uq_agent_permission_rules_scope", + ), + ) + + +def downgrade() -> None: + op.drop_table("agent_permission_rules") diff --git a/surfsense_backend/app/agents/new_chat/chat_deepagent.py b/surfsense_backend/app/agents/new_chat/chat_deepagent.py index 61de1fffa..672570696 100644 --- a/surfsense_backend/app/agents/new_chat/chat_deepagent.py +++ b/surfsense_backend/app/agents/new_chat/chat_deepagent.py @@ -23,9 +23,16 @@ from deepagents import SubAgent, SubAgentMiddleware, __version__ as deepagents_v from deepagents.backends import StateBackend from deepagents.graph import BASE_AGENT_PROMPT from deepagents.middleware.patch_tool_calls import PatchToolCallsMiddleware +from deepagents.middleware.skills import SkillsMiddleware from deepagents.middleware.subagents import GENERAL_PURPOSE_SUBAGENT from langchain.agents import create_agent -from langchain.agents.middleware import TodoListMiddleware +from langchain.agents.middleware import ( + LLMToolSelectorMiddleware, + ModelCallLimitMiddleware, + ModelFallbackMiddleware, + TodoListMiddleware, + ToolCallLimitMiddleware, +) from langchain_anthropic.middleware import AnthropicPromptCachingMiddleware from langchain_core.language_models import BaseChatModel from langchain_core.tools import BaseTool @@ -33,27 +40,51 @@ from langgraph.types import Checkpointer from sqlalchemy.ext.asyncio import AsyncSession from app.agents.new_chat.context import SurfSenseContextSchema +from app.agents.new_chat.feature_flags import AgentFeatureFlags, get_flags from app.agents.new_chat.filesystem_backends import build_backend_resolver from app.agents.new_chat.filesystem_selection import FilesystemMode, FilesystemSelection from app.agents.new_chat.llm_config import AgentConfig from app.agents.new_chat.middleware import ( + ActionLogMiddleware, AnonymousDocumentMiddleware, + BusyMutexMiddleware, + ClearToolUsesEdit, DedupHITLToolCallsMiddleware, + DoomLoopMiddleware, FileIntentMiddleware, KnowledgeBasePersistenceMiddleware, KnowledgePriorityMiddleware, KnowledgeTreeMiddleware, MemoryInjectionMiddleware, + NoopInjectionMiddleware, + OtelSpanMiddleware, + PermissionMiddleware, + RetryAfterMiddleware, + SpillingContextEditingMiddleware, + SpillToBackendEdit, SurfSenseFilesystemMiddleware, + ToolCallNameRepairMiddleware, + build_skills_backend_factory, + create_surfsense_compaction_middleware, + default_skills_sources, ) -from app.agents.new_chat.middleware.safe_summarization import ( - create_safe_summarization_middleware, +from app.agents.new_chat.permissions import Rule, Ruleset +from app.agents.new_chat.plugin_loader import ( + PluginContext, + load_allowed_plugin_names_from_env, + load_plugin_middlewares, ) +from app.agents.new_chat.subagents import build_specialized_subagents from app.agents.new_chat.system_prompt import ( build_configurable_system_prompt, build_surfsense_system_prompt, ) +from app.agents.new_chat.tools.invalid_tool import ( + INVALID_TOOL_NAME, + invalid_tool, +) from app.agents.new_chat.tools.registry import ( + BUILTIN_TOOLS, build_tools_async, get_connector_gated_tools, ) @@ -321,6 +352,17 @@ async def create_surfsense_deep_agent( disabled_tools=modified_disabled_tools, additional_tools=list(additional_tools) if additional_tools else None, ) + + # Tier 1.6: register `invalid` tool. It is dispatched only when + # ToolCallNameRepairMiddleware rewrites a malformed call. We + # intentionally append it AFTER ``build_tools_async`` so it never + # appears in the system-prompt tool list (which is built from the + # registry, not the bound tool list). + _flags: AgentFeatureFlags = get_flags() + if _flags.enable_tool_call_repair and INVALID_TOOL_NAME not in { + t.name for t in tools + }: + tools = [*list(tools), invalid_tool] _perf_log.info( "[create_agent] build_tools_async in %.3fs (%d tools)", time.perf_counter() - _t0, @@ -397,6 +439,8 @@ async def create_surfsense_deep_agent( available_connectors=available_connectors, available_document_types=available_document_types, mentioned_document_ids=mentioned_document_ids, + max_input_tokens=_max_input_tokens, + flags=_flags, checkpointer=checkpointer, ) _perf_log.info( @@ -411,6 +455,71 @@ async def create_surfsense_deep_agent( return agent +# Tier 1.1: tools whose output is too costly / lossy to discard. Keep +# this conservative — anything listed here is *never* pruned by +# ContextEditingMiddleware. The list is filtered against actually-bound +# tool names so disabled connectors don't show up here. +_PRUNE_PROTECTED_TOOL_NAMES: frozenset[str] = frozenset( + { + "generate_report", + "generate_resume", + "generate_podcast", + "generate_video_presentation", + "generate_image", + # Read-heavy connector reads — recomputing them is expensive + "read_email", + "search_emails", + # The fallback for malformed tool calls — keep its replies visible + "invalid", + } +) + + +def _safe_exclude_tools(tools: Sequence[BaseTool]) -> tuple[str, ...]: + """Return ``exclude_tools`` derived from the actually-bound tool list. + + Filters :data:`_PRUNE_PROTECTED_TOOL_NAMES` against the bound tools + so we never list tools that don't exist (would be a silent no-op). + """ + enabled = {t.name for t in tools} + return tuple(name for name in _PRUNE_PROTECTED_TOOL_NAMES if name in enabled) + + +# Tier 2.1 / cleanup: opencode `Permission.disabled` parity. Replaces the +# legacy binary ``_CONNECTOR_TYPE_TO_SEARCHABLE``-based gating with a +# declarative pass over :data:`BUILTIN_TOOLS`. Each tool that declares a +# ``required_connector`` not present in ``available_connectors`` gets a +# deny rule so any execution attempt short-circuits with permission_denied. +def _synthesize_connector_deny_rules( + *, + available_connectors: list[str] | None, + enabled_tool_names: set[str], +) -> list[Rule]: + """Build deny rules for tools whose required connector is not enabled. + + Source of truth is ``ToolDefinition.required_connector`` in + :data:`BUILTIN_TOOLS`. A tool only gets a deny rule when: + + 1. It is currently bound (``enabled_tool_names``). + 2. It declares a ``required_connector``. + 3. That connector is *not* in ``available_connectors``. + + This expresses the OpenCode ``Permission.disabled`` semantics + declaratively, replacing the substring-heuristic binary gating + that used to consult the hardcoded ``_CONNECTOR_TYPE_TO_SEARCHABLE`` + map. + """ + available = set(available_connectors or []) + deny: list[Rule] = [] + for tool_def in BUILTIN_TOOLS: + if tool_def.name not in enabled_tool_names: + continue + rc = tool_def.required_connector + if rc and rc not in available: + deny.append(Rule(permission=tool_def.name, pattern="*", action="deny")) + return deny + + def _build_compiled_agent_blocking( *, llm: BaseChatModel, @@ -426,6 +535,8 @@ def _build_compiled_agent_blocking( available_connectors: list[str] | None, available_document_types: list[str] | None, mentioned_document_ids: list[int] | None, + max_input_tokens: int | None, + flags: AgentFeatureFlags, checkpointer: Checkpointer, ): """Build the middleware stack and compile the agent graph synchronously. @@ -458,7 +569,7 @@ def _build_compiled_agent_blocking( created_by_id=user_id, thread_id=thread_id, ), - create_safe_summarization_middleware(llm, StateBackend), + create_surfsense_compaction_middleware(llm, StateBackend), PatchToolCallsMiddleware(), AnthropicPromptCachingMiddleware(unsupported_model_behavior="ignore"), ] @@ -470,13 +581,319 @@ def _build_compiled_agent_blocking( "middleware": gp_middleware, } + # Tier 4.3: specialized user-facing subagents (explore, report_writer, + # connector_negotiator). Registered through SubAgentMiddleware alongside + # the general-purpose spec so the parent's `task` tool can address them + # by name. Off by default until the flag flips so existing deployments + # don't see new agent types in the task tool description. + specialized_subagents: list[SubAgent] = [] + if ( + flags.enable_specialized_subagents + and not flags.disable_new_agent_stack + ): + try: + # Specialized subagents share the parent's filesystem + + # todo view so their system prompts (which promise + # ``read_file``, ``ls``, ``grep``, ``glob``, ``write_todos``) + # actually match runtime behavior. Build *fresh* instances + # rather than aliasing the parent's GP middleware to avoid + # subtle state coupling across compiled graphs. + subagent_extra_middleware: list = [ + TodoListMiddleware(), + SurfSenseFilesystemMiddleware( + backend=backend_resolver, + filesystem_mode=filesystem_mode, + search_space_id=search_space_id, + created_by_id=user_id, + thread_id=thread_id, + ), + ] + specialized_subagents = build_specialized_subagents( + tools=tools, + model=llm, + extra_middleware=subagent_extra_middleware, + ) + except Exception as exc: # pragma: no cover - defensive + logging.warning( + "Specialized subagent build failed; running without them: %s", + exc, + ) + specialized_subagents = [] + + subagent_specs: list[SubAgent] = [general_purpose_spec, *specialized_subagents] + # Main agent middleware # Order: AnonDoc -> Tree -> Priority -> FileIntent -> Filesystem -> Persistence -> ... # before_agent hooks run in declared order; later injections sit closer to # the latest human turn. Tree (large + cacheable) is injected earliest so # provider-side prefix caching has more material to hit; FileIntent (most # actionable per-turn contract) is injected closest to the user message. + # + # ``wrap_model_call`` ordering: the FIRST middleware in the list is the + # OUTERMOST wrapper. To ensure prune executes before summarization, + # place ``SpillingContextEditingMiddleware`` before + # ``SurfSenseCompactionMiddleware`` (Tier 1.1 + 1.3). + # Compaction is the canonical token-budget defense after the + # cleanup tier removed ``SafeSummarizationMiddleware``. The Bedrock + # buffer-empty defense is folded into ``SurfSenseCompactionMiddleware``. + summarization_mw = create_surfsense_compaction_middleware(llm, StateBackend) + _ = flags.enable_compaction_v2 # historical flag; retained for telemetry parity + + # Tier 1.1: ContextEditing prune. Trigger at 55% of model_max_input, + # earlier than summarization (~85%). When disabled, no edit runs. + context_edit_mw = None + if ( + flags.enable_context_editing + and not flags.disable_new_agent_stack + and max_input_tokens + ): + spill_edit = SpillToBackendEdit( + trigger=int(max_input_tokens * 0.55), + clear_at_least=int(max_input_tokens * 0.15), + keep=5, + exclude_tools=_safe_exclude_tools(tools), + clear_tool_inputs=True, + ) + clear_edit = ClearToolUsesEdit( + trigger=int(max_input_tokens * 0.55), + clear_at_least=int(max_input_tokens * 0.15), + keep=5, + exclude_tools=_safe_exclude_tools(tools), + clear_tool_inputs=True, + placeholder="[cleared - older tool output trimmed for context]", + ) + context_edit_mw = SpillingContextEditingMiddleware( + edits=[spill_edit, clear_edit], + backend_resolver=backend_resolver, + ) + + # Tier 1.4 / 1.8 / 1.9 / 1.10: built-in retry/fallback/limits. + retry_mw = ( + RetryAfterMiddleware(max_retries=3) + if flags.enable_retry_after and not flags.disable_new_agent_stack + else None + ) + # Fallback chain — primary is the agent's own model; we add cheap + # alternatives. Off by default; only the first call site that + # configures the chain via env should enable it. + fallback_mw: ModelFallbackMiddleware | None = None + if flags.enable_model_fallback and not flags.disable_new_agent_stack: + try: + fallback_mw = ModelFallbackMiddleware( + "openai:gpt-4o-mini", + "anthropic:claude-3-5-haiku-20241022", + ) + except Exception: + logging.warning("ModelFallbackMiddleware init failed; skipping.") + fallback_mw = None + model_call_limit_mw = ( + ModelCallLimitMiddleware( + thread_limit=120, + run_limit=80, + exit_behavior="end", + ) + if flags.enable_model_call_limit and not flags.disable_new_agent_stack + else None + ) + tool_call_limit_mw = ( + ToolCallLimitMiddleware(thread_limit=300, run_limit=80, exit_behavior="continue") + if flags.enable_tool_call_limit and not flags.disable_new_agent_stack + else None + ) + + # Tier 1.5: provider-compat _noop injection. + noop_mw = ( + NoopInjectionMiddleware() + if flags.enable_compaction_v2 and not flags.disable_new_agent_stack + else None + ) + + # Tier 1.7: tool-call name repair (lowercase + invalid fallback). + # + # ``registered_tool_names`` MUST cover every tool the model can legitimately + # call. That includes the bound ``tools`` list AND every tool provided by + # middleware in the stack — ``FilesystemMiddleware`` (read_file, ls, grep, + # glob, edit_file, write_file, execute), ``TodoListMiddleware`` + # (write_todos), ``SubAgentMiddleware`` (task), ``SkillsMiddleware`` (skill + # loaders), etc. If we only inspect ``tools`` here, every call to + # ``read_file`` / ``ls`` / ``grep`` from the model will be rewritten to + # ``invalid`` because the repair middleware doesn't recognize them. The + # built-in deepagents middleware aren't in scope yet at this point of the + # function but they're added unconditionally below, so we hard-code their + # canonical names alongside the dynamic ``tools`` set. + repair_mw = None + if flags.enable_tool_call_repair and not flags.disable_new_agent_stack: + registered_names: set[str] = {t.name for t in tools} + # Tools owned by the standard deepagents middleware stack. + registered_names |= { + "write_todos", + "ls", + "read_file", + "write_file", + "edit_file", + "glob", + "grep", + "execute", + "task", + } + repair_mw = ToolCallNameRepairMiddleware( + registered_tool_names=registered_names, + fuzzy_match_threshold=None, # opencode parity: no fuzzy step + ) + + # Tier 1.11: doom-loop detector. Off by default until UI handles. + doom_loop_mw = ( + DoomLoopMiddleware(threshold=3) + if flags.enable_doom_loop and not flags.disable_new_agent_stack + else None + ) + + # Tier 2.1: PermissionMiddleware. Layers, earliest -> latest (last + # match wins per opencode): + # + # 1. ``surfsense_defaults`` — single ``allow */*`` rule. SurfSense + # already runs per-tool HITL (see ``tools/hitl.py``) for mutating + # connector tools, so we only want PermissionMiddleware to *deny* + # things the user has gated off; the default fallback in + # ``permissions.evaluate`` is ``ask``, which would double-prompt + # on every safe read-only call (``ls``, ``read_file``, ``grep``, + # ``glob``, ``web_search`` …) and, on resume, replay the previous + # reject decision into innocent calls. + # 2. ``connector_synthesized`` — deny rules for tools whose required + # connector is not connected to this space. Overrides #1. + # 3. (future) user-defined rules from ``agent_permission_rules`` table + # via the Agent Permissions UI. Loaded last so they override both. + permission_mw: PermissionMiddleware | None = None + if flags.enable_permission and not flags.disable_new_agent_stack: + synthesized = _synthesize_connector_deny_rules( + available_connectors=available_connectors, + enabled_tool_names={t.name for t in tools}, + ) + permission_mw = PermissionMiddleware( + rulesets=[ + Ruleset( + rules=[Rule(permission="*", pattern="*", action="allow")], + origin="surfsense_defaults", + ), + Ruleset(rules=synthesized, origin="connector_synthesized"), + ], + ) + + # Tier 5.2: ActionLogMiddleware. Off by default until the + # ``agent_action_log`` table is migrated. When enabled, persists one + # row per tool call with optional reverse_descriptor for + # /api/threads/{thread_id}/revert/{action_id}. Sits inside permission + # so denied calls aren't logged as completions. + action_log_mw: ActionLogMiddleware | None = None + if ( + flags.enable_action_log + and not flags.disable_new_agent_stack + and thread_id is not None + ): + try: + tool_defs_by_name = {td.name: td for td in BUILTIN_TOOLS} + action_log_mw = ActionLogMiddleware( + thread_id=thread_id, + search_space_id=search_space_id, + user_id=user_id, + tool_definitions=tool_defs_by_name, + ) + except Exception: # pragma: no cover - defensive + logging.warning( + "ActionLogMiddleware init failed; running without it.", + exc_info=True, + ) + action_log_mw = None + + # Tier 2.2: per-thread busy mutex. + busy_mutex_mw: BusyMutexMiddleware | None = ( + BusyMutexMiddleware() + if flags.enable_busy_mutex and not flags.disable_new_agent_stack + else None + ) + + # Tier 3b: OpenTelemetry spans (model.call + tool.call). Lives just + # inside BusyMutex so it spans every retry/fallback attempt of the + # current turn but never wraps a queued/blocked turn. + otel_mw: OtelSpanMiddleware | None = ( + OtelSpanMiddleware() + if flags.enable_otel and not flags.disable_new_agent_stack + else None + ) + + # Tier 6: plugin entry-point loader. Off by default; opt-in via the + # ``SURFSENSE_ENABLE_PLUGIN_LOADER`` flag. The allowlist is read from + # the ``SURFSENSE_ALLOWED_PLUGINS`` env var (comma-separated). A future + # PR can wire it through ``global_llm_config.yaml``. + plugin_middlewares: list[Any] = [] + if flags.enable_plugin_loader and not flags.disable_new_agent_stack: + try: + allowed_names = load_allowed_plugin_names_from_env() + if allowed_names: + plugin_middlewares = load_plugin_middlewares( + PluginContext.build( + search_space_id=search_space_id, + user_id=user_id, + thread_visibility=visibility, + llm=llm, + ), + allowed_plugin_names=allowed_names, + ) + except Exception: # pragma: no cover - defensive + logging.warning( + "Plugin loader failed; continuing without plugins.", + exc_info=True, + ) + plugin_middlewares = [] + + # Tier 4.1: SkillsMiddleware. Loads built-in + space-authored skills + # via a CompositeBackend. Sources are layered: built-in first, space + # last, so a search-space-authored skill of the same name overrides + # the bundled one. + skills_mw: SkillsMiddleware | None = None + if flags.enable_skills and not flags.disable_new_agent_stack: + try: + skills_factory = build_skills_backend_factory( + search_space_id=search_space_id + if filesystem_mode == FilesystemMode.CLOUD + else None, + ) + skills_mw = SkillsMiddleware( + backend=skills_factory, + sources=default_skills_sources(), + ) + except Exception as exc: # pragma: no cover - defensive + logging.warning("SkillsMiddleware init failed; skipping: %s", exc) + skills_mw = None + + # Tier 2.5: LLM-driven tool selection for >30 tools. + selector_mw: LLMToolSelectorMiddleware | None = None + if ( + flags.enable_llm_tool_selector + and not flags.disable_new_agent_stack + and len(tools) > 30 + ): + try: + selector_mw = LLMToolSelectorMiddleware( + model="openai:gpt-4o-mini", + max_tools=12, + always_include=[ + name + for name in ("update_memory", "get_connected_accounts", "scrape_webpage") + if name in {t.name for t in tools} + ], + ) + except Exception: + logging.warning("LLMToolSelectorMiddleware init failed; skipping.") + selector_mw = None + deepagent_middleware = [ + # BusyMutex is OUTERMOST: it must wrap the entire stream so no + # other turn can sneak in while this one is mid-flight. + busy_mutex_mw, + # OTel spans sit just inside BusyMutex so each retry attempt + # gets its own model.call / tool.call span. + otel_mw, TodoListMiddleware(), _memory_middleware, AnonymousDocumentMiddleware( @@ -514,10 +931,40 @@ def _build_compiled_agent_blocking( ) if filesystem_mode == FilesystemMode.CLOUD else None, - SubAgentMiddleware(backend=StateBackend, subagents=[general_purpose_spec]), - create_safe_summarization_middleware(llm, StateBackend), + # Tier 4.1: skill loader. Placed before SubAgentMiddleware so + # subagents inherit the same skill metadata (subagent specs reference + # the same source paths via `default_skills_sources()`). + skills_mw, + SubAgentMiddleware(backend=StateBackend, subagents=subagent_specs), + # Tier 2.5: tool selection (only when >30 tools and flag on). + selector_mw, + # Defensive caps, then prune, then summarize. + model_call_limit_mw, + tool_call_limit_mw, + context_edit_mw, + summarization_mw, + # Provider compatibility + retry chain — placed after prune/compact + # so retries happen on the already-trimmed payload. + noop_mw, + retry_mw, + fallback_mw, + # Tool-call repair must run after model emits but before + # permission / dedup / doom-loop interpret the calls. + repair_mw, + # Tier 2.1: deny/ask BEFORE the calls are forwarded to tool nodes. + permission_mw, + doom_loop_mw, + # Tier 5.2: action log sits inside permission so denied calls + # don't appear as completions, and outside dedup so each unique + # tool invocation gets its own row. + action_log_mw, PatchToolCallsMiddleware(), DedupHITLToolCallsMiddleware(agent_tools=list(tools)), + # Tier 6: plugin slot — sits just before AnthropicCache so plugin-side + # transforms see the final tool result and run before any caching + # heuristics. Multiple plugins in declared order; loader filtered by + # the admin allowlist already. + *plugin_middlewares, AnthropicPromptCachingMiddleware(unsupported_model_behavior="ignore"), ] deepagent_middleware = [m for m in deepagent_middleware if m is not None] diff --git a/surfsense_backend/app/agents/new_chat/errors.py b/surfsense_backend/app/agents/new_chat/errors.py new file mode 100644 index 000000000..b7bac4536 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/errors.py @@ -0,0 +1,95 @@ +""" +Typed error taxonomy for the SurfSense agent stack. + +Used by: +- :class:`RetryAfterMiddleware` (Tier 1.4) — its ``retry_on`` callable + consults the error code to decide whether a retry is appropriate. +- :class:`PermissionMiddleware` (Tier 2.1) — emits + ``code="permission_denied"`` errors when a deny rule trips. +- All tools — return :class:`StreamingError` payloads in + ``ToolMessage.additional_kwargs["error"]`` so the model and the + retry/permission layers share a contract. +""" + +from __future__ import annotations + +from typing import Literal + +from pydantic import BaseModel, Field + +ErrorCode = Literal[ + "rate_limit", + "auth", + "tool_validation", + "tool_runtime", + "context_overflow", + "provider", + "permission_denied", + "doom_loop", + "busy", + "cancelled", +] + + +class StreamingError(BaseModel): + """Structured error payload attached to ``ToolMessage.additional_kwargs["error"]``. + + Tools and middleware emit this so retry, permission, and routing + layers can decide what to do without parsing free-form strings. + """ + + code: ErrorCode + retryable: bool = False + suggestion: str | None = None + correlation_id: str | None = None + detail: str | None = Field( + default=None, + description="Free-form additional context. Not surfaced to the model.", + ) + + class Config: + frozen = True + + +class RejectedError(Exception): + """Raised when the user rejects a permission ask without feedback. + + Caught by :class:`PermissionMiddleware`; the agent stops the current + tool fan-out and surfaces a user-facing rejection. + """ + + def __init__(self, *, tool: str | None = None, pattern: str | None = None) -> None: + super().__init__(f"Permission rejected for tool {tool!r}, pattern {pattern!r}") + self.tool = tool + self.pattern = pattern + + +class CorrectedError(Exception): + """Raised when the user rejects a permission ask *with* feedback. + + The :class:`PermissionMiddleware` translates the feedback into a + synthetic ``ToolMessage`` so the model sees the user's correction + and can retry the request differently. + """ + + def __init__(self, feedback: str, *, tool: str | None = None) -> None: + super().__init__(feedback) + self.feedback = feedback + self.tool = tool + + +class BusyError(Exception): + """Raised when a second prompt arrives while the same thread is mid-stream.""" + + def __init__(self, request_id: str | None = None) -> None: + super().__init__("Thread is busy with another request") + self.request_id = request_id + + +__all__ = [ + "BusyError", + "CorrectedError", + "ErrorCode", + "RejectedError", + "StreamingError", +] diff --git a/surfsense_backend/app/agents/new_chat/feature_flags.py b/surfsense_backend/app/agents/new_chat/feature_flags.py new file mode 100644 index 000000000..ce0a3b3fa --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/feature_flags.py @@ -0,0 +1,188 @@ +""" +Feature flags for the SurfSense new_chat agent stack. + +These flags control rollout of OpenCode-pattern middleware ported into +SurfSense. They follow a "default-OFF for risky things, default-ON for +safe upgrades, master kill-switch for everything new" model. + +All new middleware checks its flag at agent build time. If the master +kill-switch ``SURFSENSE_DISABLE_NEW_AGENT_STACK`` is set, every new +middleware is disabled regardless of its individual flag. This gives +operators a single switch to revert to pre-port behavior. + +Examples +-------- + +Local development (recommended for trying everything except doom-loop / selector): + + SURFSENSE_ENABLE_CONTEXT_EDITING=true + SURFSENSE_ENABLE_COMPACTION_V2=true + SURFSENSE_ENABLE_RETRY_AFTER=true + SURFSENSE_ENABLE_TOOL_CALL_REPAIR=true + SURFSENSE_ENABLE_PERMISSION=false # default off, opt-in per deploy + SURFSENSE_ENABLE_DOOM_LOOP=false # default off until UI ships + SURFSENSE_ENABLE_LLM_TOOL_SELECTOR=false + +Master kill-switch (overrides everything else): + + SURFSENSE_DISABLE_NEW_AGENT_STACK=true +""" + +from __future__ import annotations + +import logging +import os +from dataclasses import dataclass + +logger = logging.getLogger(__name__) + + +def _env_bool(name: str, default: bool) -> bool: + """Parse a boolean env var. Accepts ``1``/``true``/``yes``/``on`` (case-insensitive).""" + raw = os.environ.get(name) + if raw is None: + return default + return raw.strip().lower() in ("1", "true", "yes", "on") + + +@dataclass(frozen=True) +class AgentFeatureFlags: + """Resolved feature-flag state for one agent build. + + Constructed via :meth:`from_env`. The dataclass is frozen so it can be + safely shared across coroutines. + """ + + # Master kill-switch — when true, every flag below resolves to False + # regardless of its env value. Used for rapid rollback. + disable_new_agent_stack: bool = False + + # Tier 1 — Agent quality + enable_context_editing: bool = False + enable_compaction_v2: bool = False + enable_retry_after: bool = False + enable_model_fallback: bool = False + enable_model_call_limit: bool = False + enable_tool_call_limit: bool = False + enable_tool_call_repair: bool = False + enable_doom_loop: bool = False # Default OFF until UI handles permission='doom_loop' + + # Tier 2 — Safety + enable_permission: bool = False # Default OFF for first deploy + enable_busy_mutex: bool = False + enable_llm_tool_selector: bool = False # Default OFF — adds per-turn LLM cost + + # Tier 4 — Skills + subagents + enable_skills: bool = False + enable_specialized_subagents: bool = False + enable_kb_planner_runnable: bool = False + + # Tier 5 — Snapshot / revert + enable_action_log: bool = False + enable_revert_route: bool = False # Backend ships before UI; route returns 503 until this flips + + # Tier 6 — Plugins + enable_plugin_loader: bool = False + + # Tier 3b — OTel (orthogonal: also requires OTEL_EXPORTER_OTLP_ENDPOINT) + enable_otel: bool = False + + @classmethod + def from_env(cls) -> AgentFeatureFlags: + """Read flags from environment. + + Master kill-switch is evaluated first; when set, all other flags + force to False. + """ + master_off = _env_bool("SURFSENSE_DISABLE_NEW_AGENT_STACK", False) + if master_off: + logger.info( + "SURFSENSE_DISABLE_NEW_AGENT_STACK is set: every new agent " + "middleware is forced OFF for this build." + ) + return cls(disable_new_agent_stack=True) + + return cls( + disable_new_agent_stack=False, + # Tier 1 + enable_context_editing=_env_bool("SURFSENSE_ENABLE_CONTEXT_EDITING", False), + enable_compaction_v2=_env_bool("SURFSENSE_ENABLE_COMPACTION_V2", False), + enable_retry_after=_env_bool("SURFSENSE_ENABLE_RETRY_AFTER", False), + enable_model_fallback=_env_bool("SURFSENSE_ENABLE_MODEL_FALLBACK", False), + enable_model_call_limit=_env_bool("SURFSENSE_ENABLE_MODEL_CALL_LIMIT", False), + enable_tool_call_limit=_env_bool("SURFSENSE_ENABLE_TOOL_CALL_LIMIT", False), + enable_tool_call_repair=_env_bool("SURFSENSE_ENABLE_TOOL_CALL_REPAIR", False), + enable_doom_loop=_env_bool("SURFSENSE_ENABLE_DOOM_LOOP", False), + # Tier 2 + enable_permission=_env_bool("SURFSENSE_ENABLE_PERMISSION", False), + enable_busy_mutex=_env_bool("SURFSENSE_ENABLE_BUSY_MUTEX", False), + enable_llm_tool_selector=_env_bool("SURFSENSE_ENABLE_LLM_TOOL_SELECTOR", False), + # Tier 4 + enable_skills=_env_bool("SURFSENSE_ENABLE_SKILLS", False), + enable_specialized_subagents=_env_bool( + "SURFSENSE_ENABLE_SPECIALIZED_SUBAGENTS", False + ), + enable_kb_planner_runnable=_env_bool( + "SURFSENSE_ENABLE_KB_PLANNER_RUNNABLE", False + ), + # Tier 5 + enable_action_log=_env_bool("SURFSENSE_ENABLE_ACTION_LOG", False), + enable_revert_route=_env_bool("SURFSENSE_ENABLE_REVERT_ROUTE", False), + # Tier 6 + enable_plugin_loader=_env_bool("SURFSENSE_ENABLE_PLUGIN_LOADER", False), + # Tier 3b + enable_otel=_env_bool("SURFSENSE_ENABLE_OTEL", False), + ) + + def any_new_middleware_enabled(self) -> bool: + """Return True if any new middleware flag is on.""" + if self.disable_new_agent_stack: + return False + return any( + ( + self.enable_context_editing, + self.enable_compaction_v2, + self.enable_retry_after, + self.enable_model_fallback, + self.enable_model_call_limit, + self.enable_tool_call_limit, + self.enable_tool_call_repair, + self.enable_doom_loop, + self.enable_permission, + self.enable_busy_mutex, + self.enable_llm_tool_selector, + self.enable_skills, + self.enable_specialized_subagents, + self.enable_kb_planner_runnable, + self.enable_action_log, + self.enable_revert_route, + self.enable_plugin_loader, + ) + ) + + +# Module-level cache. Read once at import time so the values are consistent +# across the process lifetime. Use ``reload_for_tests`` to reset in tests. +_FLAGS: AgentFeatureFlags | None = None + + +def get_flags() -> AgentFeatureFlags: + """Return the resolved feature-flag state, caching on first call.""" + global _FLAGS + if _FLAGS is None: + _FLAGS = AgentFeatureFlags.from_env() + return _FLAGS + + +def reload_for_tests() -> AgentFeatureFlags: + """Force a fresh read from env. Tests should call this after monkeypatching env.""" + global _FLAGS + _FLAGS = AgentFeatureFlags.from_env() + return _FLAGS + + +__all__ = [ + "AgentFeatureFlags", + "get_flags", + "reload_for_tests", +] diff --git a/surfsense_backend/app/agents/new_chat/middleware/__init__.py b/surfsense_backend/app/agents/new_chat/middleware/__init__.py index e885d9e6b..094c102f8 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/__init__.py +++ b/surfsense_backend/app/agents/new_chat/middleware/__init__.py @@ -1,11 +1,23 @@ """Middleware components for the SurfSense new chat agent.""" +from app.agents.new_chat.middleware.action_log import ActionLogMiddleware from app.agents.new_chat.middleware.anonymous_document import ( AnonymousDocumentMiddleware, ) +from app.agents.new_chat.middleware.busy_mutex import BusyMutexMiddleware +from app.agents.new_chat.middleware.compaction import ( + SurfSenseCompactionMiddleware, + create_surfsense_compaction_middleware, +) +from app.agents.new_chat.middleware.context_editing import ( + ClearToolUsesEdit, + SpillingContextEditingMiddleware, + SpillToBackendEdit, +) from app.agents.new_chat.middleware.dedup_tool_calls import ( DedupHITLToolCallsMiddleware, ) +from app.agents.new_chat.middleware.doom_loop import DoomLoopMiddleware from app.agents.new_chat.middleware.file_intent import ( FileIntentMiddleware, ) @@ -26,16 +38,46 @@ from app.agents.new_chat.middleware.knowledge_tree import ( from app.agents.new_chat.middleware.memory_injection import ( MemoryInjectionMiddleware, ) +from app.agents.new_chat.middleware.noop_injection import NoopInjectionMiddleware +from app.agents.new_chat.middleware.otel_span import OtelSpanMiddleware +from app.agents.new_chat.middleware.permission import PermissionMiddleware +from app.agents.new_chat.middleware.retry_after import RetryAfterMiddleware +from app.agents.new_chat.middleware.skills_backends import ( + BuiltinSkillsBackend, + SearchSpaceSkillsBackend, + build_skills_backend_factory, + default_skills_sources, +) +from app.agents.new_chat.middleware.tool_call_repair import ( + ToolCallNameRepairMiddleware, +) __all__ = [ + "ActionLogMiddleware", "AnonymousDocumentMiddleware", + "BuiltinSkillsBackend", + "BusyMutexMiddleware", + "ClearToolUsesEdit", "DedupHITLToolCallsMiddleware", + "DoomLoopMiddleware", "FileIntentMiddleware", "KnowledgeBasePersistenceMiddleware", "KnowledgeBaseSearchMiddleware", "KnowledgePriorityMiddleware", "KnowledgeTreeMiddleware", "MemoryInjectionMiddleware", + "NoopInjectionMiddleware", + "OtelSpanMiddleware", + "PermissionMiddleware", + "RetryAfterMiddleware", + "SearchSpaceSkillsBackend", + "SpillToBackendEdit", + "SpillingContextEditingMiddleware", + "SurfSenseCompactionMiddleware", "SurfSenseFilesystemMiddleware", + "ToolCallNameRepairMiddleware", + "build_skills_backend_factory", "commit_staged_filesystem_state", + "create_surfsense_compaction_middleware", + "default_skills_sources", ] diff --git a/surfsense_backend/app/agents/new_chat/middleware/action_log.py b/surfsense_backend/app/agents/new_chat/middleware/action_log.py new file mode 100644 index 000000000..cf0b57fd4 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/middleware/action_log.py @@ -0,0 +1,294 @@ +"""Append-only action-log middleware for the SurfSense agent. + +Wraps every tool call via :meth:`AgentMiddleware.awrap_tool_call` and writes +a row to :class:`~app.db.AgentActionLog` after the tool returns. Tools opt +into reversibility by declaring a ``reverse`` callable on their +:class:`~app.agents.new_chat.tools.registry.ToolDefinition`; the rendered +descriptor is persisted in ``reverse_descriptor`` for use by +``/api/threads/{thread_id}/revert/{action_id}``. + +Design points: + +* **Defensive.** Logging never blocks the agent. We catch every exception + on the DB write path and emit a warning; the tool's ``ToolMessage`` + result is always returned untouched. +* **Lightweight payload.** Only the tool ``name`` + ``args`` (capped) + + ``result_id`` + ``reverse_descriptor`` are stored. Tool output text + remains in the LangGraph checkpoint / spilled tool-output files. +* **Best-effort reversibility.** We invoke ``reverse(args, result_obj)`` + with the parsed JSON result when the tool's content is a JSON object; + otherwise the raw text is passed. Exceptions in the reverse callable + are swallowed and logged — a failed descriptor render simply means the + action is NOT marked reversible. +""" + +from __future__ import annotations + +import json +import logging +from collections.abc import Awaitable, Callable +from typing import TYPE_CHECKING, Any + +from langchain.agents.middleware import AgentMiddleware +from langchain_core.messages import ToolMessage + +from app.agents.new_chat.feature_flags import get_flags +from app.agents.new_chat.tools.registry import ToolDefinition + +if TYPE_CHECKING: # pragma: no cover - type-only + from langchain.agents.middleware.types import ToolCallRequest + from langgraph.types import Command + + +logger = logging.getLogger(__name__) + + +# Cap for the persisted ``args`` JSON to avoid bloating the action log with +# accidentally-huge inputs. Values are truncated and a flag is set in the +# stored payload so consumers can detect truncation. +_MAX_ARGS_PERSIST_BYTES = 32 * 1024 # 32KB + + +class ActionLogMiddleware(AgentMiddleware): + """Persist a row in :class:`AgentActionLog` after every tool call. + + Should be placed near the OUTERMOST end of the tool-call wrapping stack + so that it sees the *final* :class:`ToolMessage` after all retries, + permission checks, and dedup logic have run. In practice that means + placing it just inside :class:`PermissionMiddleware` and outside + :class:`DedupHITLToolCallsMiddleware`. + + The middleware is fully a no-op when: + + * the master kill-switch ``SURFSENSE_DISABLE_NEW_AGENT_STACK`` is set + (checked via :func:`get_flags`), + * the per-feature flag ``enable_action_log`` is off, or + * persistence raises (defensive: tool-call dispatch always succeeds). + + Args: + thread_id: The current chat thread's primary-key id. Required to + persist a row; if ``None`` the middleware silently no-ops. + search_space_id: Search-space id for cascade-on-delete safety. + user_id: UUID string of the user driving this turn (nullable in + anonymous mode). + tool_definitions: Optional mapping of tool name -> :class:`ToolDefinition` + so the middleware can look up the tool's ``reverse`` callable. + When omitted, no actions are marked reversible. + """ + + tools = () + + def __init__( + self, + *, + thread_id: int | None, + search_space_id: int, + user_id: str | None, + tool_definitions: dict[str, ToolDefinition] | None = None, + ) -> None: + super().__init__() + self._thread_id = thread_id + self._search_space_id = search_space_id + self._user_id = user_id + self._tool_definitions = dict(tool_definitions or {}) + + def _enabled(self) -> bool: + flags = get_flags() + if flags.disable_new_agent_stack: + return False + return bool(flags.enable_action_log) and self._thread_id is not None + + async def awrap_tool_call( + self, + request: ToolCallRequest, + handler: Callable[ + [ToolCallRequest], Awaitable[ToolMessage | Command[Any]] + ], + ) -> ToolMessage | Command[Any]: + if not self._enabled(): + return await handler(request) + + result: ToolMessage | Command[Any] + error_payload: dict[str, Any] | None = None + try: + result = await handler(request) + except Exception as exc: + # Persist the failure too so revert/audit can see it, then + # re-raise so downstream middleware (RetryAfter, etc.) handles it. + error_payload = {"type": type(exc).__name__, "message": str(exc)} + await self._record( + request=request, + result=None, + error_payload=error_payload, + ) + raise + + await self._record(request=request, result=result, error_payload=None) + return result + + async def _record( + self, + *, + request: ToolCallRequest, + result: ToolMessage | Command[Any] | None, + error_payload: dict[str, Any] | None, + ) -> None: + """Persist one ``agent_action_log`` row. Defensive: never raises.""" + try: + from app.db import AgentActionLog, shielded_async_session + + tool_name = _resolve_tool_name(request) + args_payload = _resolve_args_payload(request) + result_id = _resolve_result_id(result) + reverse_descriptor, reversible = self._render_reverse( + tool_name=tool_name, + args=_resolve_args_dict(request), + result=result, + ) + + row = AgentActionLog( + thread_id=self._thread_id, + user_id=self._user_id, + search_space_id=self._search_space_id, + turn_id=_resolve_turn_id(request), + message_id=_resolve_message_id(request), + tool_name=tool_name, + args=args_payload, + result_id=result_id, + reversible=reversible, + reverse_descriptor=reverse_descriptor, + error=error_payload, + ) + async with shielded_async_session() as session: + session.add(row) + await session.commit() + except Exception: + logger.warning( + "ActionLogMiddleware failed to persist action log row", + exc_info=True, + ) + + def _render_reverse( + self, + *, + tool_name: str, + args: dict[str, Any] | None, + result: ToolMessage | Command[Any] | None, + ) -> tuple[dict[str, Any] | None, bool]: + """Run the tool's ``reverse`` callable and return its descriptor. + + Returns a tuple of ``(descriptor_or_None, reversible_bool)``. When + the tool has no ``reverse`` callable, or when the callable raises, + the action is marked non-reversible. + """ + if not result or not isinstance(result, ToolMessage): + return None, False + if args is None: + return None, False + tool_def = self._tool_definitions.get(tool_name) + if tool_def is None or tool_def.reverse is None: + return None, False + try: + parsed_result = _parse_tool_result_content(result) + descriptor = tool_def.reverse(args, parsed_result) + except Exception: + logger.warning( + "Reverse descriptor render failed for tool %s", + tool_name, + exc_info=True, + ) + return None, False + if not isinstance(descriptor, dict): + return None, False + return descriptor, True + + +# --------------------------------------------------------------------------- +# Resolution helpers — defensive against tool_call request shape variation. +# --------------------------------------------------------------------------- + + +def _resolve_tool_name(request: Any) -> str: + try: + tool = getattr(request, "tool", None) + if tool is not None: + name = getattr(tool, "name", None) + if isinstance(name, str) and name: + return name + call = getattr(request, "tool_call", None) or {} + if isinstance(call, dict): + name = call.get("name") + if isinstance(name, str) and name: + return name + except Exception: # pragma: no cover - defensive + pass + return "unknown" + + +def _resolve_args_dict(request: Any) -> dict[str, Any] | None: + try: + call = getattr(request, "tool_call", None) + if not isinstance(call, dict): + return None + args = call.get("args") + if isinstance(args, dict): + return args + return None + except Exception: # pragma: no cover - defensive + return None + + +def _resolve_args_payload(request: Any) -> dict[str, Any] | None: + """Return a JSON-serializable args dict, truncated if too big.""" + args = _resolve_args_dict(request) + if args is None: + return None + try: + encoded = json.dumps(args, default=str) + except Exception: + return {"_repr": repr(args)[:_MAX_ARGS_PERSIST_BYTES]} + if len(encoded) <= _MAX_ARGS_PERSIST_BYTES: + return args + return { + "_truncated": True, + "_size": len(encoded), + "_preview": encoded[:_MAX_ARGS_PERSIST_BYTES], + } + + +def _resolve_turn_id(request: Any) -> str | None: + try: + call = getattr(request, "tool_call", None) or {} + if isinstance(call, dict): + tid = call.get("id") + if isinstance(tid, str): + return tid + except Exception: # pragma: no cover + pass + return None + + +def _resolve_message_id(request: Any) -> str | None: + """Tool-call IDs serve as best-available message correlator at this layer.""" + return _resolve_turn_id(request) + + +def _resolve_result_id(result: Any) -> str | None: + if isinstance(result, ToolMessage): + msg_id = getattr(result, "id", None) + if isinstance(msg_id, str): + return msg_id + return None + + +def _parse_tool_result_content(result: ToolMessage) -> Any: + content = result.content + if isinstance(content, str): + try: + return json.loads(content) + except (json.JSONDecodeError, ValueError): + return content + return content + + +__all__ = ["ActionLogMiddleware"] diff --git a/surfsense_backend/app/agents/new_chat/middleware/busy_mutex.py b/surfsense_backend/app/agents/new_chat/middleware/busy_mutex.py new file mode 100644 index 000000000..1d95638d0 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/middleware/busy_mutex.py @@ -0,0 +1,231 @@ +""" +BusyMutexMiddleware — per-thread asyncio lock + cancel token. + +Tier 2.2 in the OpenCode-port plan. Mirrors opencode's +``Stream.scoped(AbortController)`` pattern (single-process, in-memory +lock + cooperative cancellation). For multi-worker deployments a +distributed lock backend (Redis or PostgreSQL advisory locks) is a +phase-2 follow-up. + +What this provides: +- A ``WeakValueDictionary[str, asyncio.Lock]`` keyed by ``thread_id``; + acquiring the lock during ``before_agent`` blocks any concurrent + prompt on the same thread until release. +- A per-thread ``asyncio.Event`` (``cancel_event``) that long-running + tools can poll to abort cooperatively. The event is reset between + turns. Tools should check ``runtime.context.cancel_event.is_set()`` + in tight inner loops. +- A typed :class:`~app.agents.new_chat.errors.BusyError` raised when a + second turn arrives while the lock is held. + +Note: SurfSense's ``stream_new_chat`` is the call site that should +acquire/release. Wiring this as middleware means the contract is +explicit and the lock manager is shared with subagents that compile +their own ``create_agent`` runnables. +""" + +from __future__ import annotations + +import asyncio +import logging +import weakref +from typing import Any + +from langchain.agents.middleware.types import ( + AgentMiddleware, + AgentState, + ContextT, + ResponseT, +) +from langgraph.config import get_config +from langgraph.runtime import Runtime + +from app.agents.new_chat.errors import BusyError + +logger = logging.getLogger(__name__) + + +class _ThreadLockManager: + """Process-local registry of per-thread asyncio locks + cancel events.""" + + def __init__(self) -> None: + self._locks: weakref.WeakValueDictionary[str, asyncio.Lock] = ( + weakref.WeakValueDictionary() + ) + self._cancel_events: dict[str, asyncio.Event] = {} + + def lock_for(self, thread_id: str) -> asyncio.Lock: + lock = self._locks.get(thread_id) + if lock is None: + lock = asyncio.Lock() + self._locks[thread_id] = lock + return lock + + def cancel_event(self, thread_id: str) -> asyncio.Event: + event = self._cancel_events.get(thread_id) + if event is None: + event = asyncio.Event() + self._cancel_events[thread_id] = event + return event + + def request_cancel(self, thread_id: str) -> bool: + event = self._cancel_events.get(thread_id) + if event is None: + return False + event.set() + return True + + def reset(self, thread_id: str) -> None: + event = self._cancel_events.get(thread_id) + if event is not None: + event.clear() + + +# Module-level singleton — process-local but reused across all agent +# instances built in this process. Subagents created in nested +# ``create_agent`` calls also get this so locks are coherent. +manager = _ThreadLockManager() + + +def get_cancel_event(thread_id: str) -> asyncio.Event: + """Public accessor used by long-running tools to poll cancellation.""" + return manager.cancel_event(thread_id) + + +def request_cancel(thread_id: str) -> bool: + """Trip the cancel event for ``thread_id``. Returns True if found.""" + return manager.request_cancel(thread_id) + + +def reset_cancel(thread_id: str) -> None: + """Reset the cancel event for ``thread_id`` (called between turns).""" + manager.reset(thread_id) + + +class BusyMutexMiddleware(AgentMiddleware[AgentState[ResponseT], ContextT, ResponseT]): + """Block concurrent prompts on the same thread. + + Acquires the thread's lock in ``abefore_agent`` and releases in + ``aafter_agent``. If the lock is held, raises :class:`BusyError` + so the caller can emit a ``surfsense.busy`` SSE event with the + in-flight request id. + + Args: + require_thread_id: When True, raise :class:`BusyError` if no + ``thread_id`` can be resolved from the active + ``RunnableConfig``. Default is False — we treat a missing + thread_id as "this turn has nothing to lock against" and + no-op the mutex. Set True only when you trust the call + site to always provide ``configurable.thread_id`` (e.g. + in production where ``stream_new_chat`` always does). + """ + + def __init__(self, *, require_thread_id: bool = False) -> None: + super().__init__() + self._require_thread_id = require_thread_id + self.tools = [] + # Per-call locks owned by this middleware. We track them as + # an instance attribute so ``aafter_agent`` knows which lock + # to release. + self._held_locks: dict[str, asyncio.Lock] = {} + + @staticmethod + def _thread_id(runtime: Runtime[ContextT]) -> str | None: + """Extract ``thread_id`` from the active LangGraph ``RunnableConfig``. + + ``langgraph.runtime.Runtime`` deliberately does NOT expose ``config``. + The runnable config (where ``configurable.thread_id`` lives) must be + fetched via :func:`langgraph.config.get_config` from inside a node / + middleware. We fall back to ``getattr(runtime, "config", None)`` for + unit tests / legacy runtimes that synthesize a config-bearing stub. + """ + + def _from_dict(cfg: Any) -> str | None: + if not isinstance(cfg, dict): + return None + tid = (cfg.get("configurable") or {}).get("thread_id") + return str(tid) if tid is not None else None + + # Preferred path: real LangGraph runtime context. + try: + tid = _from_dict(get_config()) + except Exception: + tid = None + if tid is not None: + return tid + + # Fallback for tests and any runtime that surfaces a config dict + # directly on the runtime instance. + return _from_dict(getattr(runtime, "config", None)) + + async def abefore_agent( # type: ignore[override] + self, + state: AgentState[Any], + runtime: Runtime[ContextT], + ) -> dict[str, Any] | None: + del state + thread_id = self._thread_id(runtime) + if thread_id is None: + if self._require_thread_id: + raise BusyError("no thread_id configured") + logger.debug( + "BusyMutexMiddleware: no thread_id resolved from RunnableConfig; " + "skipping per-thread lock for this turn." + ) + return None + + lock = manager.lock_for(thread_id) + if lock.locked(): + raise BusyError(request_id=thread_id) + await lock.acquire() + self._held_locks[thread_id] = lock + # Reset the cancel event so this turn starts fresh + reset_cancel(thread_id) + return None + + async def aafter_agent( # type: ignore[override] + self, + state: AgentState[Any], + runtime: Runtime[ContextT], + ) -> dict[str, Any] | None: + del state + thread_id = self._thread_id(runtime) + if thread_id is None: + return None + lock = self._held_locks.pop(thread_id, None) + if lock is not None and lock.locked(): + lock.release() + # Always clear cancel event between turns so a stale signal + # doesn't leak into the next request. + reset_cancel(thread_id) + return None + + # Provide sync no-ops because the middleware base class allows them + def before_agent( # type: ignore[override] + self, state: AgentState[Any], runtime: Runtime[ContextT] + ) -> dict[str, Any] | None: + # Sync path: no asyncio.Lock to acquire. Best we can do is reject + # if anyone else is in flight. + thread_id = self._thread_id(runtime) + if thread_id is None: + if self._require_thread_id: + raise BusyError("no thread_id configured") + return None + lock = manager.lock_for(thread_id) + if lock.locked(): + raise BusyError(request_id=thread_id) + return None + + def after_agent( # type: ignore[override] + self, state: AgentState[Any], runtime: Runtime[ContextT] + ) -> dict[str, Any] | None: + return None + + +__all__ = [ + "BusyMutexMiddleware", + "get_cancel_event", + "manager", + "request_cancel", + "reset_cancel", +] diff --git a/surfsense_backend/app/agents/new_chat/middleware/compaction.py b/surfsense_backend/app/agents/new_chat/middleware/compaction.py new file mode 100644 index 000000000..8b02089c9 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/middleware/compaction.py @@ -0,0 +1,253 @@ +""" +SurfSense compaction middleware. + +Subclasses :class:`deepagents.middleware.summarization.SummarizationMiddleware` +to add SurfSense-specific behavior: + +1. **Structured summary template** (OpenCode-style ``## Goal / Constraints / + Progress / Key Decisions / Next Steps / Critical Context / Relevant Files``). +2. **Protect SurfSense-specific SystemMessages** so injected hints + (````, ````, ````, + ````, ````, ````, ````) + are *not* summarized away and are kept verbatim in the post-summary + message list. +3. **Sanitize ``content=None``** when feeding messages into ``get_buffer_string`` + (Azure OpenAI / LiteLLM defense — when a provider streams an AIMessage + containing only tool_calls and no text, ``content`` can be ``None`` and + ``get_buffer_string`` crashes iterating over ``None``). This used to live in + ``safe_summarization.py``; folded in here. + +This replaces ``app.agents.new_chat.middleware.safe_summarization``. + +Tier 1.3 in the OpenCode-port plan. +""" + +from __future__ import annotations + +import logging +from typing import TYPE_CHECKING, Any + +from deepagents.middleware.summarization import ( + SummarizationMiddleware, + compute_summarization_defaults, +) +from langchain_core.messages import SystemMessage + +from app.observability import otel as ot + +if TYPE_CHECKING: + from deepagents.backends.protocol import BACKEND_TYPES + from langchain_core.language_models import BaseChatModel + from langchain_core.messages import AnyMessage + +logger = logging.getLogger(__name__) + +# OpenCode-faithful structured summary template. Mirrors +# ``opencode/packages/opencode/src/session/compaction.ts:40-75``. Kept as a +# module-level constant so unit tests can assert on its sections. +SURFSENSE_SUMMARY_PROMPT = """ +SurfSense Conversation Compaction Assistant + + + +Extract the most important context from the conversation history below into a structured summary that will replace the older messages. + + + +You are running because the conversation has grown beyond the model's input window. The conversation history below will be summarized and replaced with your output. Use the structured template that follows; keep each section concise but comprehensive enough that the agent can resume work without losing context. Each section is a checklist — populate it with relevant content or write "None" if there is nothing to report. + +## Goal +What is the user's primary goal or request? State it in one or two sentences. + +## Constraints +What boundaries must the agent respect (citations rules, visibility scope, allowed tools, user-imposed style, deadlines, deny-listed topics)? + +## Progress +What has the agent already accomplished? List each completed step succinctly. Do not reproduce tool output; just record the conclusion. + +## Key Decisions +What choices were made and why? Include rejected alternatives and the reasoning behind selecting the current path. + +## Next Steps +What specific tasks remain to achieve the goal? Order them by dependency. + +## Critical Context +What facts, IDs, document titles, query keywords, error messages, or partial answers must persist into the next turn? Include verbatim quotes only when the exact wording matters (e.g. a precise filter clause or a literal name). + +## Relevant Files +What documents or paths in the SurfSense knowledge base are in play? Use ``/documents/...`` paths exactly as they appeared in the workspace tree. + + + +Messages to summarize: +{messages} + + +Respond ONLY with the structured summary. Do not include any text before or after. +""" + +# SystemMessage prefixes that must NOT be summarized away. They are +# re-injected on every turn by the corresponding middleware, but the +# compaction step happens *before* re-injection in some paths, so we +# must preserve them verbatim across the cutoff. +PROTECTED_SYSTEM_PREFIXES: tuple[str, ...] = ( + "", # KnowledgePriorityMiddleware + "", # KnowledgeTreeMiddleware + "", # FileIntentMiddleware + "", # MemoryInjectionMiddleware + "", # MemoryInjectionMiddleware + "", # MemoryInjectionMiddleware + "", # MemoryInjectionMiddleware +) + + +def _is_protected_system_message(msg: AnyMessage) -> bool: + """Return True if ``msg`` is a SystemMessage we must not summarize.""" + if not isinstance(msg, SystemMessage): + return False + content = msg.content + if not isinstance(content, str): + return False + stripped = content.lstrip() + return any(stripped.startswith(prefix) for prefix in PROTECTED_SYSTEM_PREFIXES) + + +def _sanitize_message_content(msg: AnyMessage) -> AnyMessage: + """Return ``msg`` with ``content=None`` coerced to ``""``. + + Folds in the historical defense from ``safe_summarization.py`` — + ``get_buffer_string`` reads ``m.text`` which iterates ``self.content``, + so a ``None`` content (Azure OpenAI / LiteLLM streaming a tool-only + AIMessage) explodes. We return a copy with empty string content so + downstream consumers see an empty body without mutating the original. + """ + if getattr(msg, "content", "not-missing") is not None: + return msg + try: + return msg.model_copy(update={"content": ""}) + except AttributeError: + import copy + + new_msg = copy.copy(msg) + try: + new_msg.content = "" + except Exception: + logger.debug( + "Could not sanitize content=None on message of type %s", + type(msg).__name__, + ) + return msg + return new_msg + + +class SurfSenseCompactionMiddleware(SummarizationMiddleware): + """SummarizationMiddleware tuned for SurfSense. + + Notes + ----- + - Overrides :meth:`_partition_messages` so protected SystemMessages + survive into the ``preserved_messages`` half regardless of cutoff. + - Overrides :meth:`_filter_summary_messages` so the buffer-string path + never iterates ``None`` content. + - Inherits everything else (auto-trigger, backend offload, + ``_summarization_event`` plumbing, ``ContextOverflowError`` fallback). + """ + + def _partition_messages( # type: ignore[override] + self, + conversation_messages: list[AnyMessage], + cutoff_index: int, + ) -> tuple[list[AnyMessage], list[AnyMessage]]: + """Split messages but always preserve SurfSense protected SystemMessages. + + Mirrors OpenCode's ``PRUNE_PROTECTED_TOOLS`` philosophy + (``opencode/packages/opencode/src/session/compaction.ts``): some + message types are always kept verbatim because they are part of the + agent's working contract, not transient output. + + Also opens a ``compaction.run`` OTel span (no-op when OTel is off) + so dashboards can count compaction events and message-volume + without having to instrument upstream callers. + """ + # Opening a span here is appropriate because partitioning is the + # first call SummarizationMiddleware makes when it has decided to + # summarize; we record the volume and then close as a normal span. + with ot.compaction_span( + reason="auto", + messages_in=len(conversation_messages), + extra={"compaction.cutoff_index": int(cutoff_index)}, + ): + messages_to_summarize, preserved_messages = ( + super()._partition_messages(conversation_messages, cutoff_index) + ) + + protected: list[AnyMessage] = [] + kept_for_summary: list[AnyMessage] = [] + for msg in messages_to_summarize: + if _is_protected_system_message(msg): + protected.append(msg) + else: + kept_for_summary.append(msg) + + # Place protected blocks at the *front* of preserved_messages so + # they keep their original ordering relative to the summary + # HumanMessage that precedes the rest of the preserved tail. + return kept_for_summary, [*protected, *preserved_messages] + + def _filter_summary_messages( # type: ignore[override] + self, messages: list[AnyMessage] + ) -> list[AnyMessage]: + """Filter previous summaries AND sanitize ``content=None``. + + Folds the ``safe_summarization.py`` defense in: when the buffer + builder iterates ``m.text`` over ``None`` it explodes; sanitizing + here covers both the sync and async offload paths. + """ + filtered = super()._filter_summary_messages(messages) + return [_sanitize_message_content(m) for m in filtered] + + +def create_surfsense_compaction_middleware( + model: BaseChatModel, + backend: BACKEND_TYPES, + *, + summary_prompt: str | None = None, + history_path_prefix: str = "/conversation_history", + **overrides: Any, +) -> SurfSenseCompactionMiddleware: + """Build a :class:`SurfSenseCompactionMiddleware` with sensible defaults. + + Pulls profile-aware ``trigger`` / ``keep`` / ``truncate_args_settings`` + via :func:`deepagents.middleware.summarization.compute_summarization_defaults` + so callers get the same behavior as ``create_summarization_middleware`` + plus our overrides. + + Args: + model: Chat model to call for summary generation. + backend: Backend instance or factory for offloading conversation history. + summary_prompt: Optional override; defaults to :data:`SURFSENSE_SUMMARY_PROMPT`. + history_path_prefix: Path prefix for offloaded conversation history. + **overrides: Forwarded to :class:`SurfSenseCompactionMiddleware`. + """ + defaults = compute_summarization_defaults(model) + return SurfSenseCompactionMiddleware( + model=model, + backend=backend, + trigger=overrides.pop("trigger", defaults["trigger"]), + keep=overrides.pop("keep", defaults["keep"]), + trim_tokens_to_summarize=overrides.pop("trim_tokens_to_summarize", None), + truncate_args_settings=overrides.pop( + "truncate_args_settings", defaults["truncate_args_settings"] + ), + summary_prompt=summary_prompt or SURFSENSE_SUMMARY_PROMPT, + history_path_prefix=history_path_prefix, + **overrides, + ) + + +__all__ = [ + "PROTECTED_SYSTEM_PREFIXES", + "SURFSENSE_SUMMARY_PROMPT", + "SurfSenseCompactionMiddleware", + "create_surfsense_compaction_middleware", +] diff --git a/surfsense_backend/app/agents/new_chat/middleware/context_editing.py b/surfsense_backend/app/agents/new_chat/middleware/context_editing.py new file mode 100644 index 000000000..93ceab8ee --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/middleware/context_editing.py @@ -0,0 +1,349 @@ +""" +SpillToBackendEdit + SpillingContextEditingMiddleware. + +Mirrors OpenCode's spill-to-disk behavior in +``opencode/packages/opencode/src/tool/truncate.ts``. Before +``ClearToolUsesEdit`` rewrites old ``ToolMessage.content`` to a placeholder, +we capture the full original content and write it to the runtime backend +under ``/tool_outputs/{thread_id}/{message_id}.txt``. The placeholder is +upgraded to ``"[cleared — full output at /tool_outputs/.../{id}.txt; ask the +explore subagent to read it]"`` so the agent can recover it on demand. + +Tier 1.2 in the OpenCode-port plan. + +Why this is a middleware subclass instead of a plain ``ContextEdit``: +``ContextEdit.apply`` is sync, but writing to the backend is async. We +capture the spill payloads inside ``apply`` and flush them via +``await backend.aupload_files(...)`` from ``awrap_model_call`` *before* +delegating to the handler, so the explore subagent can always read what +the placeholder advertises. +""" + +from __future__ import annotations + +import logging +import threading +from collections.abc import Awaitable, Callable, Sequence +from copy import deepcopy +from dataclasses import dataclass, field +from typing import TYPE_CHECKING, Any + +from langchain.agents.middleware.context_editing import ( + ClearToolUsesEdit, + ContextEdit, + ContextEditingMiddleware, + TokenCounter, +) +from langchain_core.messages import ( + AIMessage, + AnyMessage, + BaseMessage, + ToolMessage, +) +from langchain_core.messages.utils import count_tokens_approximately +from langgraph.config import get_config + +if TYPE_CHECKING: + from deepagents.backends.protocol import BackendProtocol + from langchain.agents.middleware.types import ( + ModelRequest, + ModelResponse, + ) + +logger = logging.getLogger(__name__) + +DEFAULT_SPILL_PREFIX = "/tool_outputs" + + +def _build_spill_placeholder(spill_path: str) -> str: + """Build the user-facing placeholder text shown to the model.""" + return ( + f"[cleared — full output at {spill_path}; " + f"ask the explore subagent to read it]" + ) + + +def _get_thread_id_or_session() -> str: + """Best-effort thread_id discovery for the spill path. + + Falls back to a process-stable string if no LangGraph config is + available (e.g. unit tests). The exact value doesn't matter as long + as it's stable within one stream so the placeholder paths line up + with the actual upload path. + """ + try: + config = get_config() + thread_id = config.get("configurable", {}).get("thread_id") + if thread_id is not None: + return str(thread_id) + except RuntimeError: + pass + return "no_thread" + + +@dataclass(slots=True) +class SpillToBackendEdit(ContextEdit): + """Capture-and-replace context edit that spills full tool output to the backend. + + Behaves like :class:`ClearToolUsesEdit` (same trigger / keep / exclude + semantics) **and** records the original ``ToolMessage.content`` in + :attr:`pending_spills` so the wrapping middleware can flush them + before the model call. + + Args: + trigger: Token threshold above which the edit fires. + clear_at_least: Minimum number of tokens to reclaim (best effort). + keep: Number of most-recent ``ToolMessage`` instances to leave + untouched. + exclude_tools: Names of tools whose output is NOT spilled. + clear_tool_inputs: Also clear the originating ``AIMessage.tool_calls`` + args when their pair is cleared. + path_prefix: Path under the backend where spills are written. + Default ``"/tool_outputs"``. + """ + + trigger: int = 100_000 + clear_at_least: int = 0 + keep: int = 3 + clear_tool_inputs: bool = False + exclude_tools: Sequence[str] = () + path_prefix: str = DEFAULT_SPILL_PREFIX + + pending_spills: list[tuple[str, bytes]] = field(default_factory=list) + _lock: threading.Lock = field(default_factory=threading.Lock) + + def drain_pending(self) -> list[tuple[str, bytes]]: + """Return and clear the pending-spill list atomically.""" + with self._lock: + out = list(self.pending_spills) + self.pending_spills.clear() + return out + + def apply( + self, + messages: list[AnyMessage], + *, + count_tokens: TokenCounter, + ) -> None: + """Mirror ``ClearToolUsesEdit.apply`` but capture originals first.""" + tokens = count_tokens(messages) + if tokens <= self.trigger: + return + + candidates = [ + (idx, msg) for idx, msg in enumerate(messages) if isinstance(msg, ToolMessage) + ] + if self.keep >= len(candidates): + return + if self.keep: + candidates = candidates[: -self.keep] + + thread_id = _get_thread_id_or_session() + excluded_tools = set(self.exclude_tools) + + for idx, tool_message in candidates: + if tool_message.response_metadata.get("context_editing", {}).get("cleared"): + continue + + ai_message = next( + (m for m in reversed(messages[:idx]) if isinstance(m, AIMessage)), + None, + ) + if ai_message is None: + continue + + tool_call = next( + ( + call + for call in ai_message.tool_calls + if call.get("id") == tool_message.tool_call_id + ), + None, + ) + if tool_call is None: + continue + + tool_name = tool_message.name or tool_call["name"] + if tool_name in excluded_tools: + continue + + message_id = tool_message.id or tool_message.tool_call_id or "unknown" + spill_path = f"{self.path_prefix}/{thread_id}/{message_id}.txt" + + original = tool_message.content + payload = self._encode_payload(original) + with self._lock: + self.pending_spills.append((spill_path, payload)) + + messages[idx] = tool_message.model_copy( + update={ + "artifact": None, + "content": _build_spill_placeholder(spill_path), + "response_metadata": { + **tool_message.response_metadata, + "context_editing": { + "cleared": True, + "strategy": "spill_to_backend", + "spill_path": spill_path, + }, + }, + } + ) + + if self.clear_tool_inputs: + ai_idx = messages.index(ai_message) + messages[ai_idx] = self._clear_input_args( + ai_message, tool_message.tool_call_id or "" + ) + + if self.clear_at_least > 0: + new_token_count = count_tokens(messages) + cleared_tokens = max(0, tokens - new_token_count) + if cleared_tokens >= self.clear_at_least: + break + + @staticmethod + def _encode_payload(content: Any) -> bytes: + """Serialize ``ToolMessage.content`` to bytes for upload.""" + if isinstance(content, bytes): + return content + if isinstance(content, str): + return content.encode("utf-8") + try: + import json + + return json.dumps(content, default=str).encode("utf-8") + except Exception: + return str(content).encode("utf-8") + + @staticmethod + def _clear_input_args(message: AIMessage, tool_call_id: str) -> AIMessage: + updated_tool_calls: list[dict[str, Any]] = [] + cleared_any = False + for tool_call in message.tool_calls: + updated = dict(tool_call) + if updated.get("id") == tool_call_id: + updated["args"] = {} + cleared_any = True + updated_tool_calls.append(updated) + + metadata = dict(getattr(message, "response_metadata", {})) + if cleared_any: + ctx = dict(metadata.get("context_editing", {})) + ids = set(ctx.get("cleared_tool_inputs", [])) + ids.add(tool_call_id) + ctx["cleared_tool_inputs"] = sorted(ids) + metadata["context_editing"] = ctx + return message.model_copy( + update={ + "tool_calls": updated_tool_calls, + "response_metadata": metadata, + } + ) + + +BackendResolver = "Callable[[Any], BackendProtocol] | BackendProtocol" + + +class SpillingContextEditingMiddleware(ContextEditingMiddleware): + """:class:`ContextEditingMiddleware` that flushes :class:`SpillToBackendEdit` writes. + + Runs the configured edits as the parent does, then flushes any + pending spills via the supplied backend resolver before delegating + to the model handler. Spill failures are logged but never abort the + model call — the placeholder text is already in the message, so the + worst case is the agent gets a placeholder it cannot follow up on. + """ + + def __init__( + self, + *, + edits: Sequence[ContextEdit], + backend_resolver: BackendResolver | None = None, + token_count_method: str = "approximate", + ) -> None: + super().__init__(edits=list(edits), token_count_method=token_count_method) # type: ignore[arg-type] + self._backend_resolver = backend_resolver + + def _resolve_backend(self, request: ModelRequest) -> BackendProtocol | None: + if self._backend_resolver is None: + return None + if callable(self._backend_resolver): + try: + from langchain.tools import ToolRuntime + + tool_runtime = ToolRuntime( + state=getattr(request, "state", {}), + context=getattr(request.runtime, "context", None), + stream_writer=getattr(request.runtime, "stream_writer", None), + store=getattr(request.runtime, "store", None), + config=getattr(request.runtime, "config", None) or {}, + tool_call_id=None, + ) + return self._backend_resolver(tool_runtime) + except Exception: + logger.exception("Failed to resolve spill backend") + return None + return self._backend_resolver # type: ignore[return-value] + + def _collect_pending(self) -> list[tuple[str, bytes]]: + out: list[tuple[str, bytes]] = [] + for edit in self.edits: + if isinstance(edit, SpillToBackendEdit): + out.extend(edit.drain_pending()) + return out + + async def awrap_model_call( # type: ignore[override] + self, + request: ModelRequest, + handler: Callable[[ModelRequest], Awaitable[ModelResponse]], + ) -> Any: + if not request.messages: + return await handler(request) + + if self.token_count_method == "approximate": + + def count_tokens(messages: Sequence[BaseMessage]) -> int: + return count_tokens_approximately(messages) + + else: + system_msg = [request.system_message] if request.system_message else [] + + def count_tokens(messages: Sequence[BaseMessage]) -> int: + return request.model.get_num_tokens_from_messages( + system_msg + list(messages), request.tools + ) + + edited_messages = deepcopy(list(request.messages)) + for edit in self.edits: + edit.apply(edited_messages, count_tokens=count_tokens) + + pending = self._collect_pending() + if pending: + backend = self._resolve_backend(request) + if backend is not None: + try: + await backend.aupload_files(pending) + except Exception: + logger.exception( + "Spill-to-backend upload failed (%d files); placeholders " + "remain in messages but content is unrecoverable", + len(pending), + ) + else: + logger.warning( + "SpillToBackendEdit produced %d pending spills but no backend " + "resolver was configured; content is unrecoverable", + len(pending), + ) + + return await handler(request.override(messages=edited_messages)) + + +__all__ = [ + "DEFAULT_SPILL_PREFIX", + "ClearToolUsesEdit", + "SpillToBackendEdit", + "SpillingContextEditingMiddleware", + "_build_spill_placeholder", +] diff --git a/surfsense_backend/app/agents/new_chat/middleware/dedup_tool_calls.py b/surfsense_backend/app/agents/new_chat/middleware/dedup_tool_calls.py index 61494ff1a..3aff524fe 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/dedup_tool_calls.py +++ b/surfsense_backend/app/agents/new_chat/middleware/dedup_tool_calls.py @@ -2,17 +2,28 @@ When the LLM emits multiple calls to the same HITL tool with the same primary argument (e.g. two ``delete_calendar_event("Doctor Appointment")``), -only the first call is kept. Non-HITL tools are never touched. +only the first call is kept. Non-HITL tools are never touched. This runs in the ``after_model`` hook — **before** any tool executes — so the duplicate call is stripped from the AIMessage that gets checkpointed. That means it is also safe across LangGraph ``interrupt()`` boundaries: the removed call will never appear on graph resume. + +Dedup-key resolution order (Tier 2.3 / cleanup in the OpenCode-port plan): + +1. :class:`ToolDefinition.dedup_key` — callable provided by the registry + entry. This is the canonical mechanism after the cleanup-tier removal + of the legacy ``PRIMARY_ARG`` map. +2. ``tool.metadata["hitl_dedup_key"]`` — string with a primary arg name; + used by MCP / Composio tools whose schemas the registry doesn't see. + +A tool with no resolver from either path simply opts out of dedup. """ from __future__ import annotations import logging +from collections.abc import Callable from typing import Any from langchain.agents.middleware import AgentMiddleware, AgentState @@ -20,81 +31,84 @@ from langgraph.runtime import Runtime logger = logging.getLogger(__name__) -_NATIVE_HITL_TOOL_DEDUP_KEYS: dict[str, str] = { - # Gmail - "send_gmail_email": "subject", - "create_gmail_draft": "subject", - "update_gmail_draft": "draft_subject_or_id", - "trash_gmail_email": "email_subject_or_id", - # Google Calendar - "create_calendar_event": "title", - "update_calendar_event": "event_title_or_id", - "delete_calendar_event": "event_title_or_id", - # Google Drive - "create_google_drive_file": "file_name", - "delete_google_drive_file": "file_name", - # OneDrive - "create_onedrive_file": "file_name", - "delete_onedrive_file": "file_name", - # Dropbox - "create_dropbox_file": "file_name", - "delete_dropbox_file": "file_name", - # Notion - "create_notion_page": "title", - "update_notion_page": "page_title", - "delete_notion_page": "page_title", - # Linear - "create_linear_issue": "title", - "update_linear_issue": "issue_ref", - "delete_linear_issue": "issue_ref", - # Jira - "create_jira_issue": "summary", - "update_jira_issue": "issue_title_or_key", - "delete_jira_issue": "issue_title_or_key", - # Confluence - "create_confluence_page": "title", - "update_confluence_page": "page_title_or_id", - "delete_confluence_page": "page_title_or_id", -} +# Resolver type — given the tool ``args`` dict returns a stable +# string used to dedupe consecutive calls. ``None`` means no dedup. +DedupResolver = Callable[[dict[str, Any]], str] + + +def wrap_dedup_key_by_arg_name(arg_name: str) -> DedupResolver: + """Adapt a string-arg name into a :data:`DedupResolver`. + + Convenience helper used by registry entries that just want to dedupe + on a single arg's lowercased value (the most common case for native + HITL tools like ``send_gmail_email`` keyed on ``subject``). + + Example:: + + ToolDefinition( + name="send_gmail_email", + ..., + dedup_key=wrap_dedup_key_by_arg_name("subject"), + ) + """ + + def _resolver(args: dict[str, Any]) -> str: + return str(args.get(arg_name, "")).lower() + + return _resolver + + +# Backwards-compatible alias for code that imported the original +# private name. New callers should use :func:`wrap_dedup_key_by_arg_name`. +_wrap_string_key = wrap_dedup_key_by_arg_name class DedupHITLToolCallsMiddleware(AgentMiddleware): # type: ignore[type-arg] """Remove duplicate HITL tool calls from a single LLM response. - Only the **first** occurrence of each (tool-name, primary-arg-value) + Only the **first** occurrence of each ``(tool-name, dedup_key)`` pair is kept; subsequent duplicates are silently dropped. - The dedup map is built from two sources: + The dedup-resolver map is built from two sources, in priority order: - 1. A comprehensive list of native HITL tools (hardcoded above). - 2. Any ``StructuredTool`` instances passed via *agent_tools* whose - ``metadata`` contains ``{"hitl": True, "hitl_dedup_key": "..."}``. - This is how MCP tools automatically get dedup support. + 1. ``tool.metadata["dedup_key"]`` — callable provided by the registry's + ``ToolDefinition.dedup_key`` (Tier 2.3). Receives the args dict + and returns a string signature. This is the canonical mechanism + after the cleanup-tier removal of the legacy ``PRIMARY_ARG`` map. + 2. ``tool.metadata["hitl_dedup_key"]`` — string with a primary arg + name; primarily used by MCP / Composio tools. """ tools = () def __init__(self, *, agent_tools: list[Any] | None = None) -> None: - self._dedup_keys: dict[str, str] = dict(_NATIVE_HITL_TOOL_DEDUP_KEYS) + self._resolvers: dict[str, DedupResolver] = {} + for t in agent_tools or []: meta = getattr(t, "metadata", None) or {} + callable_key = meta.get("dedup_key") + if callable(callable_key): + self._resolvers[t.name] = callable_key + continue if meta.get("hitl") and meta.get("hitl_dedup_key"): - self._dedup_keys[t.name] = meta["hitl_dedup_key"] + self._resolvers[t.name] = wrap_dedup_key_by_arg_name( + meta["hitl_dedup_key"] + ) def after_model( self, state: AgentState, runtime: Runtime[Any] ) -> dict[str, Any] | None: - return self._dedup(state, self._dedup_keys) + return self._dedup(state, self._resolvers) async def aafter_model( self, state: AgentState, runtime: Runtime[Any] ) -> dict[str, Any] | None: - return self._dedup(state, self._dedup_keys) + return self._dedup(state, self._resolvers) @staticmethod def _dedup( state: AgentState, - dedup_keys: dict[str, str], # type: ignore[type-arg] + resolvers: dict[str, DedupResolver], ) -> dict[str, Any] | None: messages = state.get("messages") if not messages: @@ -110,9 +124,16 @@ class DedupHITLToolCallsMiddleware(AgentMiddleware): # type: ignore[type-arg] for tc in tool_calls: name = tc.get("name", "") - dedup_key_arg = dedup_keys.get(name) - if dedup_key_arg is not None: - arg_val = str(tc.get("args", {}).get(dedup_key_arg, "")).lower() + resolver = resolvers.get(name) + if resolver is not None: + try: + arg_val = resolver(tc.get("args", {}) or {}) + except Exception: + logger.exception( + "Dedup resolver for tool %s raised; keeping call", name + ) + deduped.append(tc) + continue key = (name, arg_val) if key in seen: logger.info( diff --git a/surfsense_backend/app/agents/new_chat/middleware/doom_loop.py b/surfsense_backend/app/agents/new_chat/middleware/doom_loop.py new file mode 100644 index 000000000..49ac7dfa8 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/middleware/doom_loop.py @@ -0,0 +1,228 @@ +""" +DoomLoopMiddleware — pattern-based detector for repeated identical tool calls. + +Mirrors ``opencode/packages/opencode/src/session/processor.ts`` doom-loop +behavior. When the same tool with the same arguments is called N times +in a row, the agent has likely entered an infinite loop. We surface this +to the user as an interrupt with ``permission="doom_loop"`` so the UI +can render an "Are you stuck? Continue / cancel?" affordance. + +Tier 1.11 in the OpenCode-port plan. + +This ships **OFF by default** until the frontend explicitly handles +``context.permission == "doom_loop"`` interrupts (the plan flips +``SURFSENSE_ENABLE_DOOM_LOOP=true`` once the UI is ready). + +Wire format: uses SurfSense's existing ``interrupt()`` payload shape +(see ``app/agents/new_chat/tools/hitl.py``): + + { + "type": "permission_ask", + "action": {"tool": , "params": }, + "context": {"permission": "doom_loop", "recent_signatures": [...]}, + } + +so the frontend that already handles HITL prompts can render this with +no changes beyond a string check. +""" + +from __future__ import annotations + +import hashlib +import json +import logging +from collections import deque +from typing import Any + +from langchain.agents.middleware.types import ( + AgentMiddleware, + AgentState, + ContextT, + ResponseT, +) +from langchain_core.messages import AIMessage +from langgraph.config import get_config +from langgraph.runtime import Runtime +from langgraph.types import interrupt + +from app.observability import otel as ot + +logger = logging.getLogger(__name__) + + +def _signature(name: str, args: Any) -> str: + """Hash a tool call ``(name, args)`` to a short signature.""" + try: + canonical = json.dumps(args, sort_keys=True, default=str) + except (TypeError, ValueError): + canonical = repr(args) + digest = hashlib.sha1(f"{name}::{canonical}".encode()).hexdigest() + return digest[:16] + + +class DoomLoopMiddleware(AgentMiddleware[AgentState[ResponseT], ContextT, ResponseT]): + """Detect repeated identical tool calls and prompt the user. + + Tracks a sliding window of the most-recent ``threshold`` tool-call + signatures across the live request. When all entries match, raise + a SurfSense-style HITL interrupt with ``permission="doom_loop"``. + + Args: + threshold: How many consecutive identical signatures count as a + doom loop. Default 3 (opencode parity). + """ + + def __init__(self, *, threshold: int = 3) -> None: + super().__init__() + if threshold < 2: + raise ValueError("DoomLoopMiddleware threshold must be >= 2") + self._threshold = threshold + self.tools = [] + # Per-thread sliding windows. We can't put this in graph state + # without state-schema gymnastics; for one process-lifetime it's + # fine to keep an in-memory map keyed by thread_id. + self._windows: dict[str, deque[str]] = {} + + @staticmethod + def _thread_id_from_runtime(runtime: Runtime[ContextT]) -> str: + """Resolve the thread id for sliding-window keying. + + Prefer LangGraph's ``get_config()`` (the only way to read + ``RunnableConfig`` inside a node — :class:`Runtime` does NOT carry + a ``config`` attribute). Fall back to ``runtime.config`` for unit + tests that synthesize a config-bearing stub. Default + ``"no_thread"`` is intentionally only used when both lookups fail + — it would collapse all threads into one window so we keep the + debug log loud. + """ + + def _from_dict(cfg: Any) -> str | None: + if not isinstance(cfg, dict): + return None + tid = (cfg.get("configurable") or {}).get("thread_id") + return str(tid) if tid is not None else None + + try: + tid = _from_dict(get_config()) + except Exception: + tid = None + if tid is not None: + return tid + + tid = _from_dict(getattr(runtime, "config", None)) + if tid is not None: + return tid + + logger.debug( + "DoomLoopMiddleware: no thread_id resolved from RunnableConfig; " + "falling back to shared 'no_thread' window." + ) + return "no_thread" + + def _window(self, thread_id: str) -> deque[str]: + win = self._windows.get(thread_id) + if win is None: + win = deque(maxlen=self._threshold) + self._windows[thread_id] = win + return win + + def _detect( + self, message: AIMessage, runtime: Runtime[ContextT] + ) -> tuple[bool, list[str], dict[str, Any] | None]: + if not message.tool_calls: + return False, [], None + + thread_id = self._thread_id_from_runtime(runtime) + window = self._window(thread_id) + + triggered_call: dict[str, Any] | None = None + for call in message.tool_calls: + name = call.get("name") if isinstance(call, dict) else getattr(call, "name", None) + args = call.get("args") if isinstance(call, dict) else getattr(call, "args", {}) + if not isinstance(name, str): + continue + sig = _signature(name, args) + window.append(sig) + if ( + len(window) >= self._threshold + and len(set(window)) == 1 + ): + triggered_call = {"name": name, "params": args or {}} + break + + if triggered_call is None: + return False, list(window), None + return True, list(window), triggered_call + + def after_model( # type: ignore[override] + self, + state: AgentState[ResponseT], + runtime: Runtime[ContextT], + ) -> dict[str, Any] | None: + messages = state.get("messages") or [] + if not messages: + return None + last = messages[-1] + if not isinstance(last, AIMessage): + return None + + triggered, signatures, action = self._detect(last, runtime) + if not triggered: + return None + + logger.warning( + "Doom loop detected: tool %s called %d times in a row (sig=%s)", + action["name"] if action else "", + self._threshold, + signatures[-1] if signatures else "", + ) + + # Tier 3b: interrupt.raised span with permission=doom_loop attribute + # so dashboards can break out doom-loop interrupts from regular + # permission asks via the ``interrupt.permission`` attribute. + with ot.interrupt_span( + interrupt_type="permission_ask", + extra={ + "interrupt.permission": "doom_loop", + "interrupt.threshold": self._threshold, + "interrupt.tool": (action or {}).get("tool", ""), + }, + ): + decision = interrupt( + { + "type": "permission_ask", + "action": action or {"tool": "", "params": {}}, + "context": { + "permission": "doom_loop", + "recent_signatures": signatures, + "threshold": self._threshold, + }, + } + ) + + # Reset window so the next decision (continue/cancel) starts fresh. + thread_id = self._thread_id_from_runtime(runtime) + self._windows.pop(thread_id, None) + + # Decision shape mirrors ``tools/hitl.py``: {"decision_type": "..."} + # If the user cancelled, jump to end. Otherwise return ``None`` so the + # tool call proceeds. The frontend's exact reply names may differ — + # we tolerate any shape that contains a string with "reject"/"cancel". + if isinstance(decision, dict): + kind = str(decision.get("decision_type") or decision.get("type") or "").lower() + if "reject" in kind or "cancel" in kind: + return {"jump_to": "end"} + return None + + async def aafter_model( # type: ignore[override] + self, + state: AgentState[ResponseT], + runtime: Runtime[ContextT], + ) -> dict[str, Any] | None: + return self.after_model(state, runtime) + + +__all__ = [ + "DoomLoopMiddleware", + "_signature", +] diff --git a/surfsense_backend/app/agents/new_chat/middleware/knowledge_search.py b/surfsense_backend/app/agents/new_chat/middleware/knowledge_search.py index edd8c7af1..f39870df6 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/knowledge_search.py +++ b/surfsense_backend/app/agents/new_chat/middleware/knowledge_search.py @@ -31,14 +31,17 @@ from collections.abc import Sequence from datetime import UTC, datetime from typing import Any +from langchain.agents import create_agent from langchain.agents.middleware import AgentMiddleware, AgentState from langchain_core.language_models import BaseChatModel from langchain_core.messages import AIMessage, BaseMessage, HumanMessage, SystemMessage +from langchain_core.runnables import Runnable from langgraph.runtime import Runtime from litellm import token_counter from pydantic import BaseModel, Field, ValidationError from sqlalchemy import select +from app.agents.new_chat.feature_flags import get_flags from app.agents.new_chat.filesystem_selection import FilesystemMode from app.agents.new_chat.filesystem_state import SurfSenseFilesystemState from app.agents.new_chat.path_resolver import ( @@ -589,6 +592,53 @@ class KnowledgePriorityMiddleware(AgentMiddleware): # type: ignore[type-arg] self.available_document_types = available_document_types self.top_k = top_k self.mentioned_document_ids = mentioned_document_ids or [] + # Tier 4.2: build the kb-planner private Runnable ONCE here so we + # don't pay the create_agent compile cost (50–200ms) on every turn. + # Disabled by default behind ``enable_kb_planner_runnable``; when off + # the planner falls back to the legacy ``self.llm.ainvoke`` path. + self._planner: Runnable | None = None + self._planner_compile_failed = False + + def _build_kb_planner_runnable(self) -> Runnable | None: + """Compile the kb-planner private :class:`Runnable` once. + + Returns ``None`` when the feature flag is disabled, when the LLM is + unavailable, or when ``create_agent`` raises (we fall back to the + legacy ``self.llm.ainvoke`` path in that case). Compilation happens + lazily on first call, then memoized via ``self._planner``. + + The compiled agent is constructed without tools — the planner's + contract is "answer with structured JSON" — but with ``RetryAfter`` + + the OpenCode-port retry/limit middleware so it shares the parent + agent's resilience guarantees. + """ + if self._planner is not None or self._planner_compile_failed: + return self._planner + if self.llm is None: + return None + flags = get_flags() + if ( + not flags.enable_kb_planner_runnable + or flags.disable_new_agent_stack + ): + return None + + from app.agents.new_chat.middleware.retry_after import RetryAfterMiddleware + + try: + self._planner = create_agent( + self.llm, + tools=[], + middleware=[RetryAfterMiddleware(max_retries=2)], + ) + except Exception as exc: # pragma: no cover - defensive + logger.warning( + "kb-planner Runnable compile failed; falling back to llm.ainvoke: %s", + exc, + ) + self._planner_compile_failed = True + self._planner = None + return self._planner async def _plan_search_inputs( self, @@ -611,11 +661,32 @@ class KnowledgePriorityMiddleware(AgentMiddleware): # type: ignore[type-arg] loop = asyncio.get_running_loop() t0 = loop.time() + # Tier 4.2: prefer the compiled-once planner Runnable when enabled; + # otherwise fall back to ``self.llm.ainvoke``. The ``surfsense:internal`` + # tag is preserved on both paths so ``_stream_agent_events`` still + # suppresses the planner's intermediate events from the UI. + planner = self._build_kb_planner_runnable() try: - response = await self.llm.ainvoke( - [HumanMessage(content=prompt)], - config={"tags": ["surfsense:internal"]}, - ) + if planner is not None: + planner_state = await planner.ainvoke( + {"messages": [HumanMessage(content=prompt)]}, + config={"tags": ["surfsense:internal"]}, + ) + response_messages = ( + planner_state.get("messages", []) + if isinstance(planner_state, dict) + else [] + ) + response = ( + response_messages[-1] + if response_messages + else AIMessage(content="") + ) + else: + response = await self.llm.ainvoke( + [HumanMessage(content=prompt)], + config={"tags": ["surfsense:internal"]}, + ) plan = _parse_kb_search_plan_response(_extract_text_from_message(response)) optimized_query = ( re.sub(r"\s+", " ", plan.optimized_query).strip() or user_text diff --git a/surfsense_backend/app/agents/new_chat/middleware/noop_injection.py b/surfsense_backend/app/agents/new_chat/middleware/noop_injection.py new file mode 100644 index 000000000..f16084892 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/middleware/noop_injection.py @@ -0,0 +1,133 @@ +""" +``_noop`` provider-compatibility tool + injection middleware. + +OpenCode injects a ``_noop`` tool for LiteLLM/Bedrock/Copilot when the +model call has empty tools but message history includes prior +``tool_calls`` — some providers 400 in that shape (see +``opencode/packages/opencode/src/session/llm.ts:209-228``). SurfSense uses +LiteLLM, and the compaction summarize call (no tools, history full of +tool calls) hits this. Tier 1.5 in the OpenCode-port plan. + +Operation: a :class:`NoopInjectionMiddleware` ``wrap_model_call`` checks +if the request has zero tools but the last AI message in history includes +``tool_calls``. If yes, it injects the ``_noop`` tool only — never globally, +mirroring opencode's gating exactly. The :func:`noop_tool` returns empty +content when called (which it should never be in practice). +""" + +from __future__ import annotations + +import logging +from collections.abc import Awaitable, Callable +from typing import Any + +from langchain.agents.middleware.types import ( + AgentMiddleware, + AgentState, + ContextT, + ModelRequest, + ModelResponse, + ResponseT, +) +from langchain_core.messages import AIMessage +from langchain_core.tools import tool + +logger = logging.getLogger(__name__) + +NOOP_TOOL_NAME = "_noop" +NOOP_TOOL_DESCRIPTION = ( + "Do not call this tool. It exists only for API compatibility." +) + + +@tool(name_or_callable=NOOP_TOOL_NAME, description=NOOP_TOOL_DESCRIPTION) +def noop_tool() -> str: + """Return empty content. Never expected to be called.""" + return "" + + +# Provider markers that benefit from ``_noop`` injection. These match +# opencode's gating list. We also accept any string containing one of +# these substrings (so e.g. ``litellm`` matches ``ChatLiteLLM``). +_NOOP_NEEDED_PROVIDERS: tuple[str, ...] = ( + "litellm", + "bedrock", + "copilot", +) + + +def _provider_needs_noop(model: Any) -> bool: + """Heuristic: does this model's provider need the _noop injection?""" + try: + ls_params = model._get_ls_params() + provider = str(ls_params.get("ls_provider", "")).lower() + except Exception: + provider = "" + + if not provider: + cls_name = type(model).__name__.lower() + provider = cls_name + + return any(needle in provider for needle in _NOOP_NEEDED_PROVIDERS) + + +def _last_ai_has_tool_calls(messages: list[Any]) -> bool: + for msg in reversed(messages): + if isinstance(msg, AIMessage): + return bool(msg.tool_calls) + return False + + +class NoopInjectionMiddleware(AgentMiddleware[AgentState[ResponseT], ContextT, ResponseT]): + """Inject the ``_noop`` tool only when the provider would otherwise 400. + + The check fires per model call, not at agent build time, because the + summarization path generates a no-tool subcall at runtime. The + extra tool is appended to ``request.tools`` as an instance — the + actual ``langchain_core.tools.BaseTool`` is bound on every call site + that creates the agent. + """ + + def __init__(self, *, noop_tool_instance: Any | None = None) -> None: + super().__init__() + self._noop_tool = noop_tool_instance or noop_tool + self.tools = [] + + def _should_inject(self, request: ModelRequest[ContextT]) -> bool: + if request.tools: + return False + if not _last_ai_has_tool_calls(request.messages): + return False + return _provider_needs_noop(request.model) + + def _augmented(self, request: ModelRequest[ContextT]) -> ModelRequest[ContextT]: + return request.override(tools=[self._noop_tool]) + + def wrap_model_call( # type: ignore[override] + self, + request: ModelRequest[ContextT], + handler: Callable[[ModelRequest[ContextT]], ModelResponse[ResponseT]], + ) -> Any: + if self._should_inject(request): + logger.debug("Injecting _noop tool for provider compatibility") + return handler(self._augmented(request)) + return handler(request) + + async def awrap_model_call( # type: ignore[override] + self, + request: ModelRequest[ContextT], + handler: Callable[[ModelRequest[ContextT]], Awaitable[ModelResponse[ResponseT]]], + ) -> Any: + if self._should_inject(request): + logger.debug("Injecting _noop tool for provider compatibility") + return await handler(self._augmented(request)) + return await handler(request) + + +__all__ = [ + "NOOP_TOOL_DESCRIPTION", + "NOOP_TOOL_NAME", + "NoopInjectionMiddleware", + "_provider_needs_noop", + "noop_tool", +] diff --git a/surfsense_backend/app/agents/new_chat/middleware/otel_span.py b/surfsense_backend/app/agents/new_chat/middleware/otel_span.py new file mode 100644 index 000000000..5585cf7a2 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/middleware/otel_span.py @@ -0,0 +1,202 @@ +""" +OpenTelemetry span middleware for the SurfSense ``new_chat`` agent. + +Wraps both ``model.call`` (LLM invocations) and ``tool.call`` (tool +executions) with OTel spans, attaching low-cardinality span names and +high-cardinality identifiers as attributes (per the Tier 3b plan). + +This middleware is intentionally a thin adapter over +:mod:`app.observability.otel`; when OTel is not configured all spans +collapse to no-ops and the wrapper adds <1µs overhead per call. When +OTel **is** configured (``OTEL_EXPORTER_OTLP_ENDPOINT`` set), every +model and tool call gets a span with the standard attributes the +plan's dashboards expect. +""" + +from __future__ import annotations + +import logging +from collections.abc import Awaitable, Callable +from typing import TYPE_CHECKING, Any + +from langchain.agents.middleware import AgentMiddleware +from langchain_core.messages import AIMessage, ToolMessage + +from app.observability import otel as ot + +if TYPE_CHECKING: # pragma: no cover — type-only + from langchain.agents.middleware.types import ( + ModelRequest, + ModelResponse, + ToolCallRequest, + ) + from langgraph.types import Command + +logger = logging.getLogger(__name__) + + +class OtelSpanMiddleware(AgentMiddleware): + """Emit ``model.call`` and ``tool.call`` OTel spans for every invocation. + + Should be placed near the **outer** end of the middleware list so + that the spans encompass retry/fallback wrapper effects (i.e. ``N`` + model.call spans for ``N`` retry attempts) but inside any concurrency/ + auth gate. Empirically this means **between** ``BusyMutex`` and + ``RetryAfter``. + """ + + def __init__(self, *, instrumentation_name: str = "surfsense.new_chat") -> None: + super().__init__() + self._instrumentation_name = instrumentation_name + + # ------------------------------------------------------------------ + # Model call spans + # ------------------------------------------------------------------ + + async def awrap_model_call( + self, + request: ModelRequest, + handler: Callable[ + [ModelRequest], Awaitable[ModelResponse | AIMessage | Any] + ], + ) -> ModelResponse | AIMessage | Any: + if not ot.is_enabled(): + return await handler(request) + + model_id, provider = _resolve_model_attrs(request) + with ot.model_call_span(model_id=model_id, provider=provider) as sp: + try: + result = await handler(request) + except Exception: + # span context manager records + re-raises + raise + else: + _annotate_model_response(sp, result) + return result + + # ------------------------------------------------------------------ + # Tool call spans + # ------------------------------------------------------------------ + + async def awrap_tool_call( + self, + request: ToolCallRequest, + handler: Callable[ + [ToolCallRequest], Awaitable[ToolMessage | Command[Any]] + ], + ) -> ToolMessage | Command[Any]: + if not ot.is_enabled(): + return await handler(request) + + tool_name = _resolve_tool_name(request) + input_size = _resolve_input_size(request) + + with ot.tool_call_span(tool_name, input_size=input_size) as sp: + result = await handler(request) + _annotate_tool_result(sp, result) + return result + + +# --------------------------------------------------------------------------- +# Attribute helpers (kept defensive; we never want OTel bookkeeping to break +# a real model/tool call). +# --------------------------------------------------------------------------- + + +def _resolve_model_attrs(request: Any) -> tuple[str | None, str | None]: + """Extract ``model.id`` and ``model.provider`` from a ``ModelRequest``.""" + model_id: str | None = None + provider: str | None = None + try: + model = getattr(request, "model", None) + if model is None: + return None, None + # langchain BaseChatModel exposes a few different identifiers + for attr in ("model_name", "model", "model_id"): + value = getattr(model, attr, None) + if value: + model_id = str(value) + break + # provider sometimes lives on ``_llm_type`` (legacy) or ``provider`` + for attr in ("provider", "_llm_type"): + value = getattr(model, attr, None) + if value: + provider = str(value) + break + except Exception: # pragma: no cover — defensive + pass + return model_id, provider + + +def _resolve_tool_name(request: Any) -> str: + try: + tool = getattr(request, "tool", None) + if tool is not None: + name = getattr(tool, "name", None) + if isinstance(name, str) and name: + return name + # Fall back to the tool_call dict + call = getattr(request, "tool_call", None) or {} + name = call.get("name") if isinstance(call, dict) else None + if isinstance(name, str) and name: + return name + except Exception: # pragma: no cover — defensive + pass + return "unknown" + + +def _resolve_input_size(request: Any) -> int | None: + try: + call = getattr(request, "tool_call", None) + if not isinstance(call, dict) or not call: + return None + args = call.get("args") + if args is None: + return None + return len(repr(args)) + except Exception: # pragma: no cover — defensive + return None + + +def _annotate_model_response(span: Any, result: Any) -> None: + """Best-effort: attach prompt/completion token counts when available.""" + try: + # ModelResponse may be a dataclass with .result containing AIMessage + msg: Any + if isinstance(result, AIMessage): + msg = result + else: + inner = getattr(result, "result", None) + msg = inner[-1] if isinstance(inner, list) and inner else inner + if msg is None: + return + usage = getattr(msg, "usage_metadata", None) or {} + if isinstance(usage, dict): + if (n := usage.get("input_tokens")) is not None: + span.set_attribute("tokens.prompt", int(n)) + if (n := usage.get("output_tokens")) is not None: + span.set_attribute("tokens.completion", int(n)) + if (n := usage.get("total_tokens")) is not None: + span.set_attribute("tokens.total", int(n)) + tool_calls = getattr(msg, "tool_calls", None) or [] + span.set_attribute("model.tool_calls", len(tool_calls)) + except Exception: # pragma: no cover — defensive + pass + + +def _annotate_tool_result(span: Any, result: Any) -> None: + try: + if isinstance(result, ToolMessage): + content = result.content if isinstance(result.content, str) else repr(result.content) + span.set_attribute("tool.output.size", len(content)) + status = getattr(result, "status", None) + if isinstance(status, str): + span.set_attribute("tool.status", status) + kwargs = getattr(result, "additional_kwargs", None) or {} + if isinstance(kwargs, dict) and kwargs.get("error"): + span.set_attribute("tool.error", True) + except Exception: # pragma: no cover — defensive + pass + + +__all__ = ["OtelSpanMiddleware"] diff --git a/surfsense_backend/app/agents/new_chat/middleware/permission.py b/surfsense_backend/app/agents/new_chat/middleware/permission.py new file mode 100644 index 000000000..f59e70bc0 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/middleware/permission.py @@ -0,0 +1,344 @@ +""" +PermissionMiddleware — pattern-based allow/deny/ask with HITL fallback. + +Mirrors ``opencode/packages/opencode/src/permission/index.ts`` but uses +SurfSense's existing ``interrupt({type, action, context})`` payload shape +(see ``app/agents/new_chat/tools/hitl.py``) so the frontend keeps +working unchanged. Tier 2.1 in the OpenCode-port plan. + +Operation: +1. ``aafter_model`` inspects the latest ``AIMessage.tool_calls``. +2. For each call, the middleware builds a list of ``patterns`` (the + tool name plus any tool-specific patterns from the resolver). It + evaluates each pattern against the layered rulesets and aggregates + the results: ``deny`` > ``ask`` > ``allow``. +3. On ``deny``: replaces the call with a synthetic ``ToolMessage`` + containing a :class:`StreamingError`. +4. On ``ask``: raises a SurfSense-style ``interrupt(...)``. The reply + shape is ``{"decision_type": "once|always|reject", "feedback"?: str}``. + - ``once``: proceed. + - ``always``: also persist allow rules for ``request.always`` patterns. + - ``reject`` w/o feedback: raise :class:`RejectedError`. + - ``reject`` w/ feedback: raise :class:`CorrectedError`. +5. On ``allow``: proceed unchanged. + +The middleware also performs a *pre-model* tool-filter step (the +``before_model`` hook) so globally denied tools are stripped from the +exposed tool list before the model gets to see them. This is +opencode's ``Permission.disabled`` equivalent and dramatically reduces +the chance the model emits a deny-only call. +""" + +from __future__ import annotations + +import logging +from collections.abc import Callable +from typing import Any + +from langchain.agents.middleware.types import ( + AgentMiddleware, + AgentState, + ContextT, +) +from langchain_core.messages import AIMessage, ToolMessage +from langgraph.runtime import Runtime +from langgraph.types import interrupt + +from app.agents.new_chat.errors import ( + CorrectedError, + RejectedError, + StreamingError, +) +from app.agents.new_chat.permissions import ( + Rule, + Ruleset, + aggregate_action, + evaluate_many, +) +from app.observability import otel as ot + +logger = logging.getLogger(__name__) + + +# Mapping ``tool_name -> resolver`` that converts ``args`` to a list of +# patterns to evaluate. The first pattern is conventionally the bare +# tool name; later entries narrow down to specific resources. +PatternResolver = Callable[[dict[str, Any]], list[str]] + + +def _default_pattern_resolver(name: str) -> PatternResolver: + def _resolve(args: dict[str, Any]) -> list[str]: + # Bare name covers the default catch-all; primary-arg fallbacks + # are best added per-tool by callers. + del args + return [name] + + return _resolve + + +class PermissionMiddleware(AgentMiddleware): # type: ignore[type-arg] + """Allow/deny/ask layer over the agent's tool calls. + + Args: + rulesets: Layered rulesets to evaluate. Earlier entries are + overridden by later ones (last-match-wins). Typical layering: + ``defaults < global < space < thread < runtime_approved``. + pattern_resolvers: Optional per-tool callables that return a list + of patterns to evaluate. When a tool isn't listed, the bare + tool name is used as the only pattern. + runtime_ruleset: Mutable :class:`Ruleset` that the middleware + extends in-place when the user replies ``"always"`` to an + ask interrupt. Reused across all calls in the same agent + instance so newly-allowed rules apply to subsequent calls. + always_emit_interrupt_payload: If True, every ask uses the + SurfSense interrupt wire format (default). Set False to + disable interrupts and treat ``ask`` as ``deny`` for + non-interactive deployments. + """ + + tools = () + + def __init__( + self, + *, + rulesets: list[Ruleset] | None = None, + pattern_resolvers: dict[str, PatternResolver] | None = None, + runtime_ruleset: Ruleset | None = None, + always_emit_interrupt_payload: bool = True, + ) -> None: + super().__init__() + self._static_rulesets: list[Ruleset] = list(rulesets or []) + self._pattern_resolvers: dict[str, PatternResolver] = dict( + pattern_resolvers or {} + ) + self._runtime_ruleset: Ruleset = runtime_ruleset or Ruleset( + origin="runtime_approved" + ) + self._emit_interrupt = always_emit_interrupt_payload + + # ------------------------------------------------------------------ + # Tool-filter step (opencode `Permission.disabled` equivalent) + # ------------------------------------------------------------------ + + def _globally_denied(self, tool_name: str) -> bool: + """Return True if a deny rule with no narrowing pattern matches.""" + rules = evaluate_many(tool_name, ["*"], *self._all_rulesets()) + return aggregate_action(rules) == "deny" + + def _all_rulesets(self) -> list[Ruleset]: + return [*self._static_rulesets, self._runtime_ruleset] + + # NOTE: ``before_model`` filtering of the tools list is left to the + # agent factory. This middleware only blocks at execution time — and + # only via the rule-evaluator path, not by mutating ``request.tools``. + # Mutating ``request.tools`` per-call would invalidate provider + # prompt-cache prefixes (see Operational risks: prompt-cache regression). + + # ------------------------------------------------------------------ + # Tool-call evaluation + # ------------------------------------------------------------------ + + def _resolve_patterns(self, tool_name: str, args: dict[str, Any]) -> list[str]: + resolver = self._pattern_resolvers.get( + tool_name, _default_pattern_resolver(tool_name) + ) + try: + patterns = resolver(args or {}) + except Exception: + logger.exception("Pattern resolver for %s raised; using bare name", tool_name) + patterns = [tool_name] + if not patterns: + patterns = [tool_name] + return patterns + + def _evaluate( + self, tool_name: str, args: dict[str, Any] + ) -> tuple[str, list[str], list[Rule]]: + patterns = self._resolve_patterns(tool_name, args) + rules = evaluate_many(tool_name, patterns, *self._all_rulesets()) + action = aggregate_action(rules) + return action, patterns, rules + + # ------------------------------------------------------------------ + # HITL ask flow — SurfSense wire format + # ------------------------------------------------------------------ + + def _raise_interrupt( + self, + *, + tool_name: str, + args: dict[str, Any], + patterns: list[str], + rules: list[Rule], + ) -> dict[str, Any]: + """Block on user approval via SurfSense's ``interrupt`` shape.""" + if not self._emit_interrupt: + return {"decision_type": "reject"} + + # ``params`` (NOT ``args``) is what SurfSense's streaming + # normalizer forwards. Other fields move into ``context``. + payload = { + "type": "permission_ask", + "action": {"tool": tool_name, "params": args or {}}, + "context": { + "patterns": patterns, + "rules": [ + { + "permission": r.permission, + "pattern": r.pattern, + "action": r.action, + } + for r in rules + ], + # Rules of thumb for the frontend: surface the patterns + # the user can promote to "always" with a single reply. + "always": patterns, + }, + } + # Tier 3b: permission.asked + interrupt.raised spans (no-op when + # OTel is disabled). Both fire here so dashboards can correlate + # "we asked X" with "interrupt was actually delivered". + with ot.permission_asked_span( + permission=tool_name, + pattern=patterns[0] if patterns else None, + extra={"permission.patterns": list(patterns)}, + ), ot.interrupt_span(interrupt_type="permission_ask"): + decision = interrupt(payload) + if isinstance(decision, dict): + return decision + # Tolerate a plain string reply ("once", "always", "reject") + if isinstance(decision, str): + return {"decision_type": decision} + return {"decision_type": "reject"} + + def _persist_always( + self, tool_name: str, patterns: list[str] + ) -> None: + """Promote ``always`` reply into runtime allow rules. + + Persistence to ``agent_permission_rules`` is done by the + streaming layer (``stream_new_chat``) once it observes the + ``always`` reply — the middleware just keeps an in-memory + copy so subsequent calls in the same stream see the rule. + """ + for pattern in patterns: + self._runtime_ruleset.rules.append( + Rule(permission=tool_name, pattern=pattern, action="allow") + ) + + # ------------------------------------------------------------------ + # Synthesizing deny -> ToolMessage + # ------------------------------------------------------------------ + + @staticmethod + def _deny_message( + tool_call: dict[str, Any], + rule: Rule, + ) -> ToolMessage: + err = StreamingError( + code="permission_denied", + retryable=False, + suggestion=( + f"rule permission={rule.permission!r} pattern={rule.pattern!r} " + f"blocked this call" + ), + ) + return ToolMessage( + content=( + f"Permission denied: rule {rule.permission}/{rule.pattern} " + f"blocked tool {tool_call.get('name')!r}." + ), + tool_call_id=tool_call.get("id") or "", + name=tool_call.get("name"), + status="error", + additional_kwargs={"error": err.model_dump()}, + ) + + # ------------------------------------------------------------------ + # The hook: aafter_model + # ------------------------------------------------------------------ + + def _process( + self, + state: AgentState, + runtime: Runtime[Any], + ) -> dict[str, Any] | None: + del runtime # unused + messages = state.get("messages") or [] + if not messages: + return None + last = messages[-1] + if not isinstance(last, AIMessage) or not last.tool_calls: + return None + + deny_messages: list[ToolMessage] = [] + kept_calls: list[dict[str, Any]] = [] + any_change = False + + for raw in last.tool_calls: + call = dict(raw) if isinstance(raw, dict) else { + "name": getattr(raw, "name", None), + "args": getattr(raw, "args", {}), + "id": getattr(raw, "id", None), + "type": "tool_call", + } + name = call.get("name") or "" + args = call.get("args") or {} + action, patterns, rules = self._evaluate(name, args) + + if action == "deny": + # Find the deny rule for the suggestion text + deny_rule = next((r for r in rules if r.action == "deny"), rules[0]) + deny_messages.append(self._deny_message(call, deny_rule)) + any_change = True + continue + + if action == "ask": + decision = self._raise_interrupt( + tool_name=name, args=args, patterns=patterns, rules=rules + ) + kind = str(decision.get("decision_type") or "reject").lower() + if kind == "once": + kept_calls.append(call) + elif kind == "always": + self._persist_always(name, patterns) + kept_calls.append(call) + elif kind == "reject": + feedback = decision.get("feedback") + if isinstance(feedback, str) and feedback.strip(): + raise CorrectedError(feedback, tool=name) + raise RejectedError(tool=name, pattern=patterns[0] if patterns else None) + else: + logger.warning( + "Unknown permission decision %r; treating as reject", kind + ) + raise RejectedError(tool=name) + continue + + # allow + kept_calls.append(call) + + if not any_change and len(kept_calls) == len(last.tool_calls): + return None + + updated = last.model_copy(update={"tool_calls": kept_calls}) + result_messages: list[Any] = [updated] + if deny_messages: + result_messages.extend(deny_messages) + return {"messages": result_messages} + + def after_model( # type: ignore[override] + self, state: AgentState, runtime: Runtime[ContextT] + ) -> dict[str, Any] | None: + return self._process(state, runtime) + + async def aafter_model( # type: ignore[override] + self, state: AgentState, runtime: Runtime[ContextT] + ) -> dict[str, Any] | None: + return self._process(state, runtime) + + +__all__ = [ + "PatternResolver", + "PermissionMiddleware", +] diff --git a/surfsense_backend/app/agents/new_chat/middleware/retry_after.py b/surfsense_backend/app/agents/new_chat/middleware/retry_after.py new file mode 100644 index 000000000..82da6a97c --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/middleware/retry_after.py @@ -0,0 +1,245 @@ +""" +RetryAfterMiddleware — Header-aware retry with custom backoff and SSE eventing. + +Why standalone instead of subclassing ``ModelRetryMiddleware``: the upstream +class calls module-level ``calculate_delay`` inline (no overridable +``_calculate_delay`` hook), so a subclass cannot inject Retry-After header +delays without rewriting the loop. Tier 1.4 in the OpenCode-port plan. + +Behaviour: +- Extracts ``Retry-After`` / ``retry-after-ms`` from + ``litellm.exceptions.RateLimitError.response.headers`` (or any exception + exposing a similar shape). +- Sleeps ``max(exponential_backoff, header_delay)`` between retries. +- Returns ``False`` from ``retry_on`` for ``ContextWindowExceededError`` / + ``ContextOverflowError`` so :class:`SurfSenseCompactionMiddleware` (or + the LangChain summarization fallback path) handles those instead. +- Emits ``surfsense.retrying`` via ``adispatch_custom_event`` on each retry + so ``stream_new_chat`` can forward it to clients as an SSE event. +""" + +from __future__ import annotations + +import asyncio +import logging +import random +import re +import time +from collections.abc import Awaitable, Callable +from typing import Any + +from langchain.agents.middleware.types import ( + AgentMiddleware, + AgentState, + ContextT, + ModelRequest, + ModelResponse, + ResponseT, +) +from langchain_core.callbacks import adispatch_custom_event, dispatch_custom_event +from langchain_core.messages import AIMessage + +logger = logging.getLogger(__name__) + +# Names of exception classes for which a retry would not help — context +# overflow needs compaction, auth needs human intervention, etc. Detected +# by class-name substring so we don't have to import LiteLLM/Anthropic +# here (which would tie this module to optional deps). +_NON_RETRYABLE_NAME_HINTS: tuple[str, ...] = ( + "ContextWindowExceeded", + "ContextOverflow", + "AuthenticationError", + "InvalidRequestError", + "PermissionDenied", + "InvalidApiKey", + "ContextLimit", +) + + +def _is_non_retryable(exc: BaseException) -> bool: + name = type(exc).__name__ + return any(hint in name for hint in _NON_RETRYABLE_NAME_HINTS) + + +def _extract_retry_after_seconds(exc: BaseException) -> float | None: + """Return seconds-to-wait suggested by the provider, if any. + + Looks at ``exc.response.headers`` or ``exc.headers`` for the standard + HTTP ``Retry-After`` header (in seconds) or its millisecond cousin + ``retry-after-ms`` (sometimes used by Anthropic / OpenAI). Falls back + to a regex on the exception message for shapes like + ``"Please retry after 30s"``. + """ + headers: dict[str, Any] | None = None + response = getattr(exc, "response", None) + if response is not None: + headers = getattr(response, "headers", None) + if headers is None: + headers = getattr(exc, "headers", None) + + if isinstance(headers, dict): + # Normalize keys to lowercase for case-insensitive matching + norm = {str(k).lower(): v for k, v in headers.items()} + ms = norm.get("retry-after-ms") + if ms is not None: + try: + return float(ms) / 1000.0 + except (TypeError, ValueError): + pass + seconds = norm.get("retry-after") + if seconds is not None: + try: + return float(seconds) + except (TypeError, ValueError): + pass + + # Last resort: scan the message for "retry after Xs" or "X seconds" + msg = str(exc) + match = re.search(r"retry\s+after\s+([0-9]+(?:\.[0-9]+)?)", msg, re.IGNORECASE) + if match: + try: + return float(match.group(1)) + except ValueError: + return None + return None + + +def _exponential_delay( + attempt: int, + *, + initial_delay: float, + backoff_factor: float, + max_delay: float, + jitter: bool, +) -> float: + """Compute an exponential-backoff delay with optional ±25% jitter.""" + delay = initial_delay * (backoff_factor**attempt) if backoff_factor else initial_delay + delay = min(delay, max_delay) + if jitter and delay > 0: + delay *= 1 + random.uniform(-0.25, 0.25) + return max(delay, 0.0) + + +class RetryAfterMiddleware(AgentMiddleware[AgentState[ResponseT], ContextT, ResponseT]): + """Retry middleware that honors provider-issued Retry-After hints. + + Drop-in replacement for :class:`langchain.agents.middleware.ModelRetryMiddleware` + when working with LiteLLM/Anthropic/OpenAI providers that surface + rate-limit hints in headers. Always emits ``surfsense.retrying`` SSE + events so the UI can show a friendly "rate limited, retrying in Xs" + indicator. + + Args: + max_retries: Maximum retries after the initial attempt (default 3). + initial_delay: Initial backoff delay in seconds. + backoff_factor: Exponential growth factor for backoff. + max_delay: Cap on per-attempt delay in seconds. + jitter: Whether to add ±25% jitter. + retry_on: Optional callable that returns True for retryable + exceptions. The default retries everything except known + non-retryable classes (context overflow, auth, etc.). + """ + + def __init__( + self, + *, + max_retries: int = 3, + initial_delay: float = 1.0, + backoff_factor: float = 2.0, + max_delay: float = 60.0, + jitter: bool = True, + retry_on: Callable[[BaseException], bool] | None = None, + ) -> None: + super().__init__() + self.max_retries = max_retries + self.initial_delay = initial_delay + self.backoff_factor = backoff_factor + self.max_delay = max_delay + self.jitter = jitter + self._retry_on: Callable[[BaseException], bool] = retry_on or ( + lambda exc: not _is_non_retryable(exc) + ) + + def _should_retry(self, exc: BaseException) -> bool: + try: + return bool(self._retry_on(exc)) + except Exception: + logger.exception("retry_on callable raised; defaulting to False") + return False + + def _delay_for_attempt(self, attempt: int, exc: BaseException) -> float: + backoff = _exponential_delay( + attempt, + initial_delay=self.initial_delay, + backoff_factor=self.backoff_factor, + max_delay=self.max_delay, + jitter=self.jitter, + ) + header = _extract_retry_after_seconds(exc) or 0.0 + return max(backoff, header) + + def wrap_model_call( # type: ignore[override] + self, + request: ModelRequest[ContextT], + handler: Callable[[ModelRequest[ContextT]], ModelResponse[ResponseT]], + ) -> ModelResponse[ResponseT] | AIMessage: + for attempt in range(self.max_retries + 1): + try: + return handler(request) + except Exception as exc: + if not self._should_retry(exc) or attempt >= self.max_retries: + raise + delay = self._delay_for_attempt(attempt, exc) + try: + dispatch_custom_event( + "surfsense.retrying", + { + "attempt": attempt + 1, + "max_retries": self.max_retries, + "delay_ms": int(delay * 1000), + "reason": type(exc).__name__, + }, + ) + except Exception: + logger.debug("dispatch_custom_event failed; suppressed", exc_info=True) + if delay > 0: + time.sleep(delay) + # Unreachable + raise RuntimeError("RetryAfterMiddleware: retry loop exited without resolution") + + async def awrap_model_call( # type: ignore[override] + self, + request: ModelRequest[ContextT], + handler: Callable[[ModelRequest[ContextT]], Awaitable[ModelResponse[ResponseT]]], + ) -> ModelResponse[ResponseT] | AIMessage: + for attempt in range(self.max_retries + 1): + try: + return await handler(request) + except Exception as exc: + if not self._should_retry(exc) or attempt >= self.max_retries: + raise + delay = self._delay_for_attempt(attempt, exc) + try: + await adispatch_custom_event( + "surfsense.retrying", + { + "attempt": attempt + 1, + "max_retries": self.max_retries, + "delay_ms": int(delay * 1000), + "reason": type(exc).__name__, + }, + ) + except Exception: + logger.debug( + "adispatch_custom_event failed; suppressed", exc_info=True + ) + if delay > 0: + await asyncio.sleep(delay) + raise RuntimeError("RetryAfterMiddleware: retry loop exited without resolution") + + +__all__ = [ + "RetryAfterMiddleware", + "_extract_retry_after_seconds", + "_is_non_retryable", +] diff --git a/surfsense_backend/app/agents/new_chat/middleware/safe_summarization.py b/surfsense_backend/app/agents/new_chat/middleware/safe_summarization.py deleted file mode 100644 index 4ddcf334f..000000000 --- a/surfsense_backend/app/agents/new_chat/middleware/safe_summarization.py +++ /dev/null @@ -1,123 +0,0 @@ -"""Safe wrapper around deepagents' SummarizationMiddleware. - -Upstream issue --------------- -`deepagents.middleware.summarization.SummarizationMiddleware._aoffload_to_backend` -(and its sync counterpart) call -``get_buffer_string(filtered_messages)`` before writing the evicted history -to the backend file. In recent ``langchain-core`` versions, ``get_buffer_string`` -accesses ``m.text`` which iterates ``self.content`` — this raises -``TypeError: 'NoneType' object is not iterable`` whenever an ``AIMessage`` -has ``content=None`` (common when a model returns *only* tool_calls, seen -frequently with Azure OpenAI ``gpt-5.x`` responses streamed through -LiteLLM). - -The exception aborts the whole agent turn, so the user just sees "Error during -chat" with no assistant response. - -Fix ---- -We subclass ``SummarizationMiddleware`` and override -``_filter_summary_messages`` — the only call site that feeds messages into -``get_buffer_string`` — to return *copies* of messages whose ``content`` is -``None`` with ``content=""``. The originals flowing through the rest of the -agent state are untouched. - -We also expose a drop-in ``create_safe_summarization_middleware`` factory -that mirrors ``deepagents.middleware.summarization.create_summarization_middleware`` -but instantiates our safe subclass. -""" - -from __future__ import annotations - -import logging -from typing import TYPE_CHECKING - -from deepagents.middleware.summarization import ( - SummarizationMiddleware, - compute_summarization_defaults, -) - -if TYPE_CHECKING: - from deepagents.backends.protocol import BACKEND_TYPES - from langchain_core.language_models import BaseChatModel - from langchain_core.messages import AnyMessage - -logger = logging.getLogger(__name__) - - -def _sanitize_message_content(msg: AnyMessage) -> AnyMessage: - """Return ``msg`` with ``content`` coerced to a non-``None`` value. - - ``get_buffer_string`` reads ``m.text`` which iterates ``self.content``; - when a provider streams back an ``AIMessage`` with only tool_calls and - no text, ``content`` can be ``None`` and the iteration explodes. We - replace ``None`` with an empty string so downstream consumers that only - care about text see an empty body. - - The original message is left untouched — we return a copy via - pydantic's ``model_copy`` when available, otherwise we fall back to - re-setting the attribute on a shallow copy. - """ - - if getattr(msg, "content", "not-missing") is not None: - return msg - - try: - return msg.model_copy(update={"content": ""}) - except AttributeError: - import copy - - new_msg = copy.copy(msg) - try: - new_msg.content = "" - except Exception: # pragma: no cover - defensive - logger.debug( - "Could not sanitize content=None on message of type %s", - type(msg).__name__, - ) - return msg - return new_msg - - -class SafeSummarizationMiddleware(SummarizationMiddleware): - """`SummarizationMiddleware` that tolerates messages with ``content=None``. - - Only ``_filter_summary_messages`` is overridden — this is the single - helper invoked by both the sync and async offload paths immediately - before ``get_buffer_string``. Normalising here means we get coverage - for both without having to copy the (long, rapidly-changing) offload - implementations from upstream. - """ - - def _filter_summary_messages(self, messages: list[AnyMessage]) -> list[AnyMessage]: - filtered = super()._filter_summary_messages(messages) - return [_sanitize_message_content(m) for m in filtered] - - -def create_safe_summarization_middleware( - model: BaseChatModel, - backend: BACKEND_TYPES, -) -> SafeSummarizationMiddleware: - """Drop-in replacement for ``create_summarization_middleware``. - - Mirrors the defaults computed by ``deepagents`` but returns our - ``SafeSummarizationMiddleware`` subclass so the - ``content=None`` crash in ``get_buffer_string`` is avoided. - """ - - defaults = compute_summarization_defaults(model) - return SafeSummarizationMiddleware( - model=model, - backend=backend, - trigger=defaults["trigger"], - keep=defaults["keep"], - trim_tokens_to_summarize=None, - truncate_args_settings=defaults["truncate_args_settings"], - ) - - -__all__ = [ - "SafeSummarizationMiddleware", - "create_safe_summarization_middleware", -] diff --git a/surfsense_backend/app/agents/new_chat/middleware/skills_backends.py b/surfsense_backend/app/agents/new_chat/middleware/skills_backends.py new file mode 100644 index 000000000..4c3791c87 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/middleware/skills_backends.py @@ -0,0 +1,332 @@ +"""Skills backends for SurfSense. + +Implements two minimal :class:`deepagents.backends.protocol.BackendProtocol` +subclasses tailored for use with :class:`deepagents.middleware.skills.SkillsMiddleware`. + +The middleware only needs four methods to load skills from a backend: + +* ``ls_info`` / ``als_info`` — list directories under a source path. +* ``download_files`` / ``adownload_files`` — fetch ``SKILL.md`` bytes. + +Other ``BackendProtocol`` methods (``read``/``write``/``edit``/``grep_raw`` …) +default to ``NotImplementedError`` from the base class. They are never reached +by the skills middleware because skill content is rendered into the system +prompt at agent build time, not edited at runtime. + +Two backends are provided: + +* :class:`BuiltinSkillsBackend` — disk-backed read of bundled skills from + ``app/agents/new_chat/skills/builtin/``. +* :class:`SearchSpaceSkillsBackend` — a thin read-only wrapper over + :class:`KBPostgresBackend` that filters notes under the privileged folder + ``/documents/_skills/``. + +Both backends are intentionally read-only: skill authoring happens out of band +(via filesystem or a search-space-admin route), so we never expose +``write`` / ``edit`` / ``upload_files``. The base class' ``NotImplementedError`` +gives a clean failure mode if anything tries. +""" + +from __future__ import annotations + +import logging +from collections.abc import Callable +from dataclasses import replace +from pathlib import Path +from typing import TYPE_CHECKING + +from deepagents.backends.composite import CompositeBackend +from deepagents.backends.protocol import ( + BackendProtocol, + FileDownloadResponse, + FileInfo, +) +from deepagents.backends.state import StateBackend + +if TYPE_CHECKING: + from langchain.tools import ToolRuntime + + from app.agents.new_chat.middleware.kb_postgres_backend import KBPostgresBackend + +logger = logging.getLogger(__name__) + + +# Limit per Agent Skills spec; matches deepagents.middleware.skills.MAX_SKILL_FILE_SIZE. +_MAX_SKILL_FILE_SIZE = 10 * 1024 * 1024 + + +def _default_builtin_root() -> Path: + """Return the absolute path to the bundled builtin skills directory. + + Located at ``app/agents/new_chat/skills/builtin/`` relative to this module. + """ + return (Path(__file__).resolve().parent.parent / "skills" / "builtin").resolve() + + +class BuiltinSkillsBackend(BackendProtocol): + """Read-only disk-backed skills source. + + Maps a virtual ``/skills/builtin/`` namespace onto a directory on local disk, + where each skill is its own subdirectory containing a ``SKILL.md`` file:: + + //SKILL.md + + The middleware calls :meth:`als_info` with the source path and expects a + ``list[FileInfo]`` whose ``is_dir=True`` entries are descended into. Then it + calls :meth:`adownload_files` with the synthesized ``SKILL.md`` paths and + parses YAML frontmatter from the returned ``content`` bytes. + + Mounting under :class:`~deepagents.backends.composite.CompositeBackend` at + prefix ``/skills/builtin/`` means the middleware can issue paths like + ``/skills/builtin/kb-research/SKILL.md`` which the composite strips down to + ``/kb-research/SKILL.md`` before forwarding here. We treat any leading + slash as anchoring at :attr:`root`. + """ + + def __init__(self, root: Path | str | None = None) -> None: + self.root: Path = Path(root).resolve() if root else _default_builtin_root() + if not self.root.exists(): + logger.info( + "BuiltinSkillsBackend root %s does not exist; skills will be empty.", + self.root, + ) + + def _resolve(self, path: str) -> Path: + """Resolve a virtual posix path under :attr:`root`, refusing escapes.""" + bare = path.lstrip("/") + candidate = (self.root / bare).resolve() if bare else self.root + # Refuse symlink/.. traversal that escapes the root. + try: + candidate.relative_to(self.root) + except ValueError as exc: + raise ValueError(f"path {path!r} escapes builtin skills root") from exc + return candidate + + def ls_info(self, path: str) -> list[FileInfo]: + try: + target = self._resolve(path) + except ValueError as exc: + logger.warning("BuiltinSkillsBackend.ls_info refused: %s", exc) + return [] + if not target.exists() or not target.is_dir(): + return [] + + infos: list[FileInfo] = [] + # Build virtual paths anchored at "/" because CompositeBackend already + # stripped the route prefix before calling us. + target_virtual = "/" if target == self.root else ( + "/" + str(target.relative_to(self.root)).replace("\\", "/") + ) + for child in sorted(target.iterdir()): + child_virtual = ( + target_virtual.rstrip("/") + "/" + child.name + if target_virtual != "/" + else "/" + child.name + ) + info: FileInfo = { + "path": child_virtual, + "is_dir": child.is_dir(), + } + if child.is_file(): + try: + info["size"] = child.stat().st_size + except OSError: # pragma: no cover - defensive + pass + infos.append(info) + return infos + + def download_files(self, paths: list[str]) -> list[FileDownloadResponse]: + responses: list[FileDownloadResponse] = [] + for p in paths: + try: + target = self._resolve(p) + except ValueError: + responses.append(FileDownloadResponse(path=p, error="invalid_path")) + continue + if not target.exists(): + responses.append(FileDownloadResponse(path=p, error="file_not_found")) + continue + if target.is_dir(): + responses.append(FileDownloadResponse(path=p, error="is_directory")) + continue + try: + # Hard cap to avoid loading rogue mega-files into memory. + size = target.stat().st_size + if size > _MAX_SKILL_FILE_SIZE: + logger.warning( + "Builtin skill file %s exceeds %d bytes; truncating.", + target, + _MAX_SKILL_FILE_SIZE, + ) + with target.open("rb") as fh: + content = fh.read(_MAX_SKILL_FILE_SIZE) + else: + content = target.read_bytes() + except PermissionError: + responses.append(FileDownloadResponse(path=p, error="permission_denied")) + continue + except OSError as exc: # pragma: no cover - defensive + logger.warning("Builtin skill read failed %s: %s", target, exc) + responses.append(FileDownloadResponse(path=p, error="file_not_found")) + continue + responses.append(FileDownloadResponse(path=p, content=content, error=None)) + return responses + + +class SearchSpaceSkillsBackend(BackendProtocol): + """Read-only view of search-space-authored skills. + + Wraps a :class:`KBPostgresBackend` and only ever reads under the privileged + folder ``/documents/_skills/`` (configurable). The folder is intended to be + writable only by search-space admins; this backend never writes. + + The skills middleware expects a layout like:: + + ///SKILL.md + + But the KB stores documents like ``/documents/_skills//SKILL.md``. + We expose the inner namespace by remapping each path. When mounted under + :class:`CompositeBackend` at prefix ``/skills/space/`` the paths the + middleware sees become ``/skills/space//SKILL.md``; the composite + strips ``/skills/space/`` and hands us ``//SKILL.md``, which we + rewrite to ``/documents/_skills//SKILL.md`` before forwarding to the + KB. + + No new database table is needed: the privileged folder convention is + enforced server-side outside of this class. We intentionally swallow any + write/edit attempts (the base class raises ``NotImplementedError``). + """ + + DEFAULT_KB_ROOT: str = "/documents/_skills" + + def __init__( + self, + kb_backend: KBPostgresBackend, + *, + kb_root: str = DEFAULT_KB_ROOT, + ) -> None: + self._kb = kb_backend + # Normalize trailing slash off so we can join cleanly. + self._kb_root = kb_root.rstrip("/") or "/" + + def _to_kb(self, path: str) -> str: + """Rewrite a virtual path into the underlying KB namespace.""" + bare = path.lstrip("/") + if not bare: + return self._kb_root + return f"{self._kb_root}/{bare}" + + def _from_kb(self, kb_path: str) -> str: + """Rewrite a KB path back into our virtual namespace.""" + if not kb_path.startswith(self._kb_root): + return kb_path # pragma: no cover - defensive + rel = kb_path[len(self._kb_root) :] + return rel if rel.startswith("/") else "/" + rel + + def ls_info(self, path: str) -> list[FileInfo]: + # KBPostgresBackend exposes only the async API meaningfully; the sync + # path falls back to ``asyncio.to_thread(...)`` in the base class. We + # keep this stub to satisfy abstract resolution; the middleware calls + # ``als_info``. + raise NotImplementedError("SearchSpaceSkillsBackend is async-only") + + async def als_info(self, path: str) -> list[FileInfo]: + kb_path = self._to_kb(path) + try: + infos = await self._kb.als_info(kb_path) + except Exception as exc: # pragma: no cover - defensive + logger.warning("SearchSpaceSkillsBackend.als_info failed: %s", exc) + return [] + remapped: list[FileInfo] = [] + for info in infos: + kb_p = info.get("path", "") + if not kb_p.startswith(self._kb_root): + continue + remapped.append({**info, "path": self._from_kb(kb_p)}) + return remapped + + def download_files(self, paths: list[str]) -> list[FileDownloadResponse]: + raise NotImplementedError("SearchSpaceSkillsBackend is async-only") + + async def adownload_files(self, paths: list[str]) -> list[FileDownloadResponse]: + kb_paths = [self._to_kb(p) for p in paths] + responses = await self._kb.adownload_files(kb_paths) + # Re-map response paths back to the virtual namespace so the middleware + # correlates them to the input list correctly. + remapped: list[FileDownloadResponse] = [] + for original, resp in zip(paths, responses, strict=True): + remapped.append(replace(resp, path=original)) + return remapped + + +SKILLS_BUILTIN_PREFIX = "/skills/builtin/" +SKILLS_SPACE_PREFIX = "/skills/space/" + + +def build_skills_backend_factory( + *, + builtin_root: Path | str | None = None, + search_space_id: int | None = None, +) -> Callable[[ToolRuntime], BackendProtocol]: + """Return a runtime-aware factory for the skills :class:`CompositeBackend`. + + When ``search_space_id`` is provided the composite includes a + :class:`SearchSpaceSkillsBackend` route at ``/skills/space/`` over a fresh + per-runtime :class:`KBPostgresBackend`, mirroring how + :func:`build_backend_resolver` constructs the main filesystem backend. + + When ``search_space_id`` is ``None`` (e.g., desktop-local mode or unit + tests) only the bundled :class:`BuiltinSkillsBackend` is exposed. + + Returning a factory rather than a fixed instance is intentional: the + underlying KB backend depends on per-call ``ToolRuntime`` state + (``staged_dirs``, ``files`` cache, runtime config), so a single shared + instance cannot serve multiple concurrent agent runs. + """ + builtin = BuiltinSkillsBackend(builtin_root) + + if search_space_id is None: + def _factory_builtin_only(runtime: ToolRuntime) -> BackendProtocol: + # Default StateBackend is intentionally inert: any path outside the + # ``/skills/builtin/`` route resolves to an empty per-runtime state + # so the SkillsMiddleware can iterate sources without raising. + return CompositeBackend( + default=StateBackend(runtime), + routes={SKILLS_BUILTIN_PREFIX: builtin}, + ) + return _factory_builtin_only + + def _factory_with_space(runtime: ToolRuntime) -> BackendProtocol: + # Imported lazily to avoid a hard dependency at module import time: + # ``KBPostgresBackend`` pulls in DB models, which are unnecessary for + # the unit-tested builtin path. + from app.agents.new_chat.middleware.kb_postgres_backend import ( + KBPostgresBackend, + ) + + kb = KBPostgresBackend(search_space_id, runtime) + space = SearchSpaceSkillsBackend(kb) + return CompositeBackend( + default=StateBackend(runtime), + routes={ + SKILLS_BUILTIN_PREFIX: builtin, + SKILLS_SPACE_PREFIX: space, + }, + ) + + return _factory_with_space + + +def default_skills_sources() -> list[str]: + """Return the canonical source list for SkillsMiddleware (built-in then space).""" + return [SKILLS_BUILTIN_PREFIX, SKILLS_SPACE_PREFIX] + + +__all__ = [ + "SKILLS_BUILTIN_PREFIX", + "SKILLS_SPACE_PREFIX", + "BuiltinSkillsBackend", + "SearchSpaceSkillsBackend", + "build_skills_backend_factory", + "default_skills_sources", +] diff --git a/surfsense_backend/app/agents/new_chat/middleware/tool_call_repair.py b/surfsense_backend/app/agents/new_chat/middleware/tool_call_repair.py new file mode 100644 index 000000000..6c3bc674d --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/middleware/tool_call_repair.py @@ -0,0 +1,190 @@ +""" +ToolCallNameRepairMiddleware — two-stage tool-name repair. + +Mirrors ``opencode/packages/opencode/src/session/llm.ts:339-358`` plus +``opencode/packages/opencode/src/tool/invalid.ts``. Tier 1.7 in the +OpenCode-port plan. + +Operation: +1. **Stage 1 — lowercase repair:** if a tool call's ``name`` is not in + the registry but ``name.lower()`` is, rewrite in place. Catches + models that emit ``Search`` instead of ``search``. +2. **Stage 2 — invalid fallback:** if still unmatched, rewrite the call + to ``invalid`` with ``args={"tool": original_name, "error": }`` + so the registered :func:`invalid_tool` returns the error to the model + for self-correction. + +Distinct from :class:`deepagents.middleware.PatchToolCallsMiddleware`, +which patches *dangling* tool calls (no matching ToolMessage) — that +class does not handle the wrong-name case at all. +""" + +from __future__ import annotations + +import difflib +import logging +from typing import Any + +from langchain.agents.middleware.types import ( + AgentMiddleware, + AgentState, + ContextT, + ResponseT, +) +from langchain_core.messages import AIMessage +from langgraph.runtime import Runtime + +from app.agents.new_chat.tools.invalid_tool import INVALID_TOOL_NAME + +logger = logging.getLogger(__name__) + + +def _coerce_existing_tool_call(call: Any) -> dict[str, Any]: + """Normalize a tool call entry to a mutable dict.""" + if isinstance(call, dict): + return dict(call) + return { + "name": getattr(call, "name", None), + "args": getattr(call, "args", {}), + "id": getattr(call, "id", None), + "type": "tool_call", + } + + +class ToolCallNameRepairMiddleware(AgentMiddleware[AgentState[ResponseT], ContextT, ResponseT]): + """Two-stage tool-name repair on the most recent ``AIMessage``. + + Args: + registered_tool_names: Set of canonically-registered tool names. + ``invalid`` should be in this set so the fallback dispatches. + fuzzy_match_threshold: Optional ``difflib`` ratio (0–1) for the + fuzzy-match step that runs *between* lowercase and invalid. + Set to ``None`` to disable fuzzy matching (opencode parity). + """ + + def __init__( + self, + *, + registered_tool_names: set[str], + fuzzy_match_threshold: float | None = 0.85, + ) -> None: + super().__init__() + self._registered = set(registered_tool_names) + self._registered_lower = {name.lower(): name for name in self._registered} + self._fuzzy_threshold = fuzzy_match_threshold + self.tools = [] + + def _registered_for_runtime(self, runtime: Runtime[ContextT]) -> set[str]: + """Allow runtime overrides to expand the set (e.g. dynamic MCP tools).""" + ctx_tools = getattr(runtime.context, "registered_tool_names", None) + if isinstance(ctx_tools, (set, frozenset)): + return self._registered | set(ctx_tools) + if isinstance(ctx_tools, (list, tuple)): + return self._registered | set(ctx_tools) + return self._registered + + def _repair_one( + self, + call: dict[str, Any], + registered: set[str], + ) -> dict[str, Any]: + name = call.get("name") + if not isinstance(name, str): + return call + + if name in registered: + return call + + # Stage 1 — lowercase + lowered = name.lower() + if lowered in registered: + call["name"] = lowered + metadata = dict(call.get("response_metadata") or {}) + metadata.setdefault("repair", "lowercase") + call["response_metadata"] = metadata + return call + + # Optional fuzzy step (off by default for opencode parity) + if self._fuzzy_threshold is not None: + close = difflib.get_close_matches( + name, registered, n=1, cutoff=self._fuzzy_threshold + ) + if close: + call["name"] = close[0] + metadata = dict(call.get("response_metadata") or {}) + metadata.setdefault("repair", f"fuzzy:{name}->{close[0]}") + call["response_metadata"] = metadata + return call + + # Stage 2 — invalid fallback + if INVALID_TOOL_NAME in registered: + original_args = call.get("args") or {} + error_msg = ( + f"Tool name '{name}' is not registered. " + f"Original arguments were: {original_args!r}." + ) + call["name"] = INVALID_TOOL_NAME + call["args"] = {"tool": name, "error": error_msg} + metadata = dict(call.get("response_metadata") or {}) + metadata.setdefault("repair", f"invalid_fallback:{name}") + call["response_metadata"] = metadata + else: + logger.warning( + "Could not repair unknown tool call %r; 'invalid' tool not registered", + name, + ) + return call + + def _maybe_repair( + self, + message: AIMessage, + registered: set[str], + ) -> AIMessage | None: + if not message.tool_calls: + return None + + new_calls: list[dict[str, Any]] = [] + any_changed = False + for raw in message.tool_calls: + call = _coerce_existing_tool_call(raw) + before = (call.get("name"), call.get("args")) + repaired = self._repair_one(call, registered) + after = (repaired.get("name"), repaired.get("args")) + if before != after: + any_changed = True + new_calls.append(repaired) + + if not any_changed: + return None + + return message.model_copy(update={"tool_calls": new_calls}) + + def after_model( # type: ignore[override] + self, + state: AgentState[ResponseT], + runtime: Runtime[ContextT], + ) -> dict[str, Any] | None: + messages = state.get("messages") or [] + if not messages: + return None + last = messages[-1] + if not isinstance(last, AIMessage): + return None + + registered = self._registered_for_runtime(runtime) + repaired = self._maybe_repair(last, registered) + if repaired is None: + return None + return {"messages": [repaired]} + + async def aafter_model( # type: ignore[override] + self, + state: AgentState[ResponseT], + runtime: Runtime[ContextT], + ) -> dict[str, Any] | None: + return self.after_model(state, runtime) + + +__all__ = [ + "ToolCallNameRepairMiddleware", +] diff --git a/surfsense_backend/app/agents/new_chat/permissions.py b/surfsense_backend/app/agents/new_chat/permissions.py new file mode 100644 index 000000000..50a0cfbdc --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/permissions.py @@ -0,0 +1,204 @@ +""" +Wildcard pattern matching + rule evaluation for the SurfSense permission system. + +Mirrors ``opencode/packages/opencode/src/permission/evaluate.ts`` and +``opencode/packages/opencode/src/util/wildcard.ts`` precisely: + +- ``Wildcard.match`` matches both the ``permission`` and the ``pattern`` + fields of a rule against the requested ``(permission, pattern)`` pair. + ``*`` matches any segment, ``**`` matches across separators. +- The evaluator runs ``findLast`` over the **flattened** list of rules + from all rulesets — last matching rule wins. +- The default fallback is ``ask`` (NOT deny), matching opencode. +- Multi-pattern requests AND together: if ANY pattern resolves to + ``deny``, the whole request is denied; if ANY needs ``ask``, an + interrupt is raised; only when all patterns ``allow`` does the + request proceed. + +Tier 2.1 in the OpenCode-port plan. +""" + +from __future__ import annotations + +import re +from collections.abc import Iterable +from dataclasses import dataclass, field +from typing import Literal + +RuleAction = Literal["allow", "deny", "ask"] + + +@dataclass(frozen=True) +class Rule: + """A single permission rule. + + Attributes: + permission: A wildcard-matched permission identifier + (e.g. ``"edit"``, ``"linear_*"``, ``"mcp:*"``, + ``"doom_loop"``). Anchored at start AND end of the input. + pattern: A wildcard-matched pattern over the request payload + (e.g. ``"/documents/secrets/**"``, ``"page_id=123"``, + ``"*"``). Anchored at start AND end. + action: One of ``"allow"`` / ``"deny"`` / ``"ask"``. + """ + + permission: str + pattern: str + action: RuleAction + + +@dataclass +class Ruleset: + """A list of rules with an associated origin used for debugging.""" + + rules: list[Rule] = field(default_factory=list) + origin: str = "unknown" # e.g. "defaults", "global", "space", "thread", "runtime" + + +# ----------------------------------------------------------------------------- +# Wildcard matcher +# ----------------------------------------------------------------------------- + + +_GLOB_TOKEN = re.compile(r"\*\*|\*|[^*]+") + + +def _wildcard_to_regex(pattern: str) -> re.Pattern[str]: + """Translate an opencode-style wildcard pattern to a compiled regex. + + Rules: + - ``**`` matches any sequence of any characters (including separators). + - ``*`` matches any sequence of characters that does **not** include + the path separator ``/`` — same as glob. + - All other characters match literally. + - The pattern is anchored at both ends (``^...$``). + """ + parts: list[str] = ["^"] + for token in _GLOB_TOKEN.findall(pattern): + if token == "**": + parts.append(r".*") + elif token == "*": + parts.append(r"[^/]*") + else: + parts.append(re.escape(token)) + parts.append("$") + return re.compile("".join(parts)) + + +_REGEX_CACHE: dict[str, re.Pattern[str]] = {} + + +def wildcard_match(value: str, pattern: str) -> bool: + """Return True if ``value`` matches the wildcard ``pattern``. + + Special case: a bare ``"*"`` pattern matches any value, including + those containing ``/`` separators. This mirrors opencode's + ``Wildcard.match`` short-circuit and matches the convention that + ``pattern="*"`` means "any pattern" in permission rules. + """ + if pattern == "*": + return True + compiled = _REGEX_CACHE.get(pattern) + if compiled is None: + compiled = _wildcard_to_regex(pattern) + _REGEX_CACHE[pattern] = compiled + return compiled.match(value) is not None + + +# ----------------------------------------------------------------------------- +# Evaluator +# ----------------------------------------------------------------------------- + + +def evaluate( + permission: str, + pattern: str, + *rulesets: Ruleset | Iterable[Rule], +) -> Rule: + """Find the last rule matching ``(permission, pattern)`` from ``rulesets``. + + Mirrors opencode ``permission/evaluate.ts:9-15`` precisely: + - Flatten rulesets in argument order. + - Walk the flat list **in reverse**. + - First reverse-match wins (i.e. the last specified rule wins). + - When no rule matches, default to ``Rule(permission, "*", "ask")``. + + Args: + permission: The permission identifier being requested + (e.g. tool name, ``"edit"``, ``"doom_loop"``). + pattern: The request-specific pattern (e.g. file path, + primary arg value). Use ``"*"`` when no specific pattern + applies. + *rulesets: Layered rulesets, applied earliest to latest. Later + rulesets override earlier ones. + + Returns: + The matched :class:`Rule`, or the default ask fallback. + """ + flat: list[Rule] = [] + for rs in rulesets: + if isinstance(rs, Ruleset): + flat.extend(rs.rules) + else: + flat.extend(rs) + + for rule in reversed(flat): + if wildcard_match(permission, rule.permission) and wildcard_match( + pattern, rule.pattern + ): + return rule + + return Rule(permission=permission, pattern="*", action="ask") + + +def evaluate_many( + permission: str, + patterns: Iterable[str], + *rulesets: Ruleset | Iterable[Rule], +) -> list[Rule]: + """Evaluate ``permission`` against each of ``patterns`` (multi-pattern AND). + + Returns the list of resolved rules in the same order as ``patterns``. + The caller is responsible for combining the results — opencode-style + multi-pattern AND collapses ``deny`` first, then ``ask``, then + ``allow``. + """ + return [evaluate(permission, p, *rulesets) for p in patterns] + + +def aggregate_action(rules: Iterable[Rule]) -> RuleAction: + """Collapse a list of per-pattern rules into one action. + + Order: + 1. If any rule is ``deny`` -> ``deny``. + 2. Else if any rule is ``ask`` -> ``ask``. + 3. Else if at least one rule is ``allow`` -> ``allow``. + 4. Else (empty input) -> ``ask`` (safe default mirroring ``evaluate``). + + Mirrors opencode's behavior in ``permission/index.ts:180-272``. + """ + saw_ask = False + saw_allow = False + for rule in rules: + if rule.action == "deny": + return "deny" + if rule.action == "ask": + saw_ask = True + elif rule.action == "allow": + saw_allow = True + if saw_ask: + return "ask" + if saw_allow: + return "allow" + return "ask" + + +__all__ = [ + "Rule", + "RuleAction", + "Ruleset", + "aggregate_action", + "evaluate", + "evaluate_many", + "wildcard_match", +] diff --git a/surfsense_backend/app/agents/new_chat/plugin_loader.py b/surfsense_backend/app/agents/new_chat/plugin_loader.py new file mode 100644 index 000000000..426e28041 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/plugin_loader.py @@ -0,0 +1,157 @@ +"""Entry-point based plugin loader for SurfSense agent middleware. + +The realization in the Tier 6 plan: LangChain's :class:`AgentMiddleware` ABC +already covers the practical surface most plugins need (``before_agent`` / +``before_model`` / ``wrap_tool_call`` / their async counterparts), so a +SurfSense-specific plugin protocol is unnecessary. + +A plugin is therefore just an installable Python package that registers a +factory callable under the ``surfsense.plugins`` entry-point group: + +.. code-block:: toml + + # in a plugin package's pyproject.toml + [project.entry-points."surfsense.plugins"] + year_substituter = "my_plugin:make_middleware" + +The factory has the signature ``Callable[[PluginContext], AgentMiddleware]``. +It receives a small, sanitized :class:`PluginContext` with the IDs and the +LLM the plugin is allowed to talk to — and **never** raw secrets, DB +sessions, or other connectors. + +## Trust model + +Plugins are loaded **only if** their entry-point ``name`` appears in +``allowed_plugins`` (admin-controlled, sourced from +``global_llm_config.yaml`` or :func:`load_allowed_plugin_names_from_env`). +There is **no env-driven auto-load**. A plugin failure is logged and +isolated; it does not break agent construction. +""" + +from __future__ import annotations + +import logging +import os +from collections.abc import Iterable +from importlib.metadata import entry_points +from typing import TYPE_CHECKING + +from langchain.agents.middleware import AgentMiddleware + +if TYPE_CHECKING: # pragma: no cover - type-only + from langchain_core.language_models import BaseChatModel + + from app.db import ChatVisibility + + +logger = logging.getLogger(__name__) + + +PLUGIN_ENTRY_POINT_GROUP = "surfsense.plugins" + + +class PluginContext(dict): + """Sanitized DI bag handed to each plugin factory. + + Backed by ``dict`` so plugins can inspect the keys they care about + without coupling to a concrete dataclass shape. Required keys: + + * ``search_space_id`` (int) + * ``user_id`` (str | None) + * ``thread_visibility`` (:class:`app.db.ChatVisibility`) + * ``llm`` (:class:`langchain_core.language_models.BaseChatModel`) + + The context **never** carries DB sessions, raw secrets, or other + connectors. If a future plugin genuinely needs DB access, that + integration goes through a rate-limited service interface, not + through this bag. + """ + + @classmethod + def build( + cls, + *, + search_space_id: int, + user_id: str | None, + thread_visibility: ChatVisibility, + llm: BaseChatModel, + ) -> PluginContext: + return cls( + search_space_id=search_space_id, + user_id=user_id, + thread_visibility=thread_visibility, + llm=llm, + ) + + +def load_plugin_middlewares( + ctx: PluginContext, + allowed_plugin_names: Iterable[str], +) -> list[AgentMiddleware]: + """Discover, allowlist-filter, and instantiate plugin middleware. + + For each entry-point in :data:`PLUGIN_ENTRY_POINT_GROUP` whose name is + in ``allowed_plugin_names``, load the factory and call it with ``ctx``. + The factory's return value must be an :class:`AgentMiddleware` instance; + anything else is logged and skipped. + + Errors are isolated — a plugin that raises during ``ep.load()`` or + factory invocation is logged at ``ERROR`` and ignored. Agent + construction continues with whatever plugins did succeed. + """ + allowed = {name for name in allowed_plugin_names if name} + if not allowed: + return [] + + out: list[AgentMiddleware] = [] + try: + eps = entry_points(group=PLUGIN_ENTRY_POINT_GROUP) + except Exception: # pragma: no cover - defensive (entry_points is robust) + logger.exception("Failed to enumerate plugin entry points") + return [] + + for ep in eps: + if ep.name not in allowed: + logger.info("Skipping non-allowlisted plugin %s", ep.name) + continue + try: + factory = ep.load() + except Exception: + logger.exception("Failed to load plugin %s", ep.name) + continue + try: + mw = factory(ctx) + except Exception: + logger.exception("Plugin %s factory raised", ep.name) + continue + if not isinstance(mw, AgentMiddleware): + logger.warning( + "Plugin %s returned %s, expected AgentMiddleware; skipping", + ep.name, + type(mw).__name__, + ) + continue + out.append(mw) + logger.info("Loaded plugin %s as %s", ep.name, type(mw).__name__) + return out + + +def load_allowed_plugin_names_from_env() -> set[str]: + """Read ``SURFSENSE_ALLOWED_PLUGINS`` (comma-separated) into a set. + + Provided as a thin convenience for deployments that don't surface plugins + through ``global_llm_config.yaml`` yet. Whitespace is stripped and empty + entries are dropped. + """ + raw = os.environ.get("SURFSENSE_ALLOWED_PLUGINS", "").strip() + if not raw: + return set() + return {token.strip() for token in raw.split(",") if token.strip()} + + +__all__ = [ + "PLUGIN_ENTRY_POINT_GROUP", + "PluginContext", + "load_allowed_plugin_names_from_env", + "load_plugin_middlewares", +] diff --git a/surfsense_backend/app/agents/new_chat/plugins/__init__.py b/surfsense_backend/app/agents/new_chat/plugins/__init__.py new file mode 100644 index 000000000..cef6bd367 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/plugins/__init__.py @@ -0,0 +1,6 @@ +"""Reference plugins bundled with SurfSense. + +These plugins are intentionally small and demonstrative. They are NOT +auto-loaded — they ship as examples that a deployment can opt into via +``global_llm_config.yaml`` or ``SURFSENSE_ALLOWED_PLUGINS``. +""" diff --git a/surfsense_backend/app/agents/new_chat/plugins/year_substituter.py b/surfsense_backend/app/agents/new_chat/plugins/year_substituter.py new file mode 100644 index 000000000..927d533d5 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/plugins/year_substituter.py @@ -0,0 +1,87 @@ +"""Reference plugin: substitute ``{{year}}`` in tool descriptions. + +Mirrors the OpenCode ``chat.system.transform`` example. Demonstrates the +:meth:`AgentMiddleware.awrap_tool_call` hook -- the plugin sees every tool +invocation and can rewrite the request *or* the result. This particular +plugin is read-only and only transforms the *description* the user might +see in error messages (no request mutation). + +The plugin is built as a factory function so the entry-point loader can +inject :class:`PluginContext` (containing the agent's LLM, search-space +ID, etc.). The factory signature +``Callable[[PluginContext], AgentMiddleware]`` is the only contract -- +SurfSense doesn't define a custom plugin protocol on top of LangChain's +:class:`AgentMiddleware`. + +Wire-up in ``pyproject.toml`` (illustrative; the in-repo plugin doesn't +need this -- it's already on the import path):: + + [project.entry-points."surfsense.plugins"] + year_substituter = "app.agents.new_chat.plugins.year_substituter:make_middleware" +""" + +from __future__ import annotations + +import logging +from collections.abc import Awaitable, Callable +from datetime import UTC, datetime +from typing import TYPE_CHECKING, Any + +from langchain.agents.middleware import AgentMiddleware + +if TYPE_CHECKING: # pragma: no cover - type-only + from langchain.agents.middleware.types import ToolCallRequest + from langchain_core.messages import ToolMessage + from langgraph.types import Command + + from app.agents.new_chat.plugin_loader import PluginContext + + +logger = logging.getLogger(__name__) + + +class _YearSubstituterMiddleware(AgentMiddleware): + """Replace ``{{year}}`` in the result text with the current UTC year.""" + + tools = () + + def __init__(self, year: int | None = None) -> None: + super().__init__() + self._year = str(year if year is not None else datetime.now(UTC).year) + + async def awrap_tool_call( + self, + request: ToolCallRequest, + handler: Callable[ + [ToolCallRequest], Awaitable[ToolMessage | Command[Any]] + ], + ) -> ToolMessage | Command[Any]: + result = await handler(request) + try: + from langchain_core.messages import ToolMessage + + if isinstance(result, ToolMessage) and isinstance(result.content, str): + if "{{year}}" in result.content: + new_text = result.content.replace("{{year}}", self._year) + result = ToolMessage( + content=new_text, + tool_call_id=result.tool_call_id, + id=result.id, + name=result.name, + status=result.status, + artifact=result.artifact, + ) + except Exception: # pragma: no cover - defensive + logger.exception("year_substituter plugin failed; passing original result") + return result + + +def make_middleware(ctx: PluginContext) -> AgentMiddleware: + """Plugin factory used by :func:`load_plugin_middlewares`.""" + # Plugin is intentionally small so it has no state to threading-protect + # and ignores ``ctx`` beyond demonstrating that the loader passes it in. + _ = ctx + return _YearSubstituterMiddleware() + + +__all__ = ["make_middleware"] diff --git a/surfsense_backend/app/agents/new_chat/prompts/__init__.py b/surfsense_backend/app/agents/new_chat/prompts/__init__.py new file mode 100644 index 000000000..c91bb8a0b --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/__init__.py @@ -0,0 +1,7 @@ +"""SurfSense agent prompt fragments. + +The prompt is composed at runtime by :mod:`composer` from the markdown +fragments under ``base/``, ``providers/``, ``tools/``, ``examples/``, and +``routing/``. ``system_prompt.py`` is now a thin wrapper that delegates +to :func:`composer.compose_system_prompt`. +""" diff --git a/surfsense_backend/app/agents/new_chat/prompts/base/__init__.py b/surfsense_backend/app/agents/new_chat/prompts/base/__init__.py new file mode 100644 index 000000000..8b1378917 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/base/__init__.py @@ -0,0 +1 @@ + diff --git a/surfsense_backend/app/agents/new_chat/prompts/base/agent_private.md b/surfsense_backend/app/agents/new_chat/prompts/base/agent_private.md new file mode 100644 index 000000000..88554ad4e --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/base/agent_private.md @@ -0,0 +1,7 @@ +You are SurfSense, a reasoning and acting AI agent designed to answer user questions using the user's personal knowledge base. + +Today's date (UTC): {resolved_today} + +When writing mathematical formulas or equations, ALWAYS use LaTeX notation. NEVER use backtick code spans or Unicode symbols for math. + +NEVER expose internal tool parameter names, backend IDs, or implementation details to the user. Always use natural, user-friendly language instead. diff --git a/surfsense_backend/app/agents/new_chat/prompts/base/agent_team.md b/surfsense_backend/app/agents/new_chat/prompts/base/agent_team.md new file mode 100644 index 000000000..5fd56ae1b --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/base/agent_team.md @@ -0,0 +1,9 @@ +You are SurfSense, a reasoning and acting AI agent designed to answer questions in this team space using the team's shared knowledge base. + +In this team thread, each message is prefixed with **[DisplayName of the author]**. Use this to attribute and reference the author of anything in the discussion (who asked a question, made a suggestion, or contributed an idea) and to cite who said what in your answers. + +Today's date (UTC): {resolved_today} + +When writing mathematical formulas or equations, ALWAYS use LaTeX notation. NEVER use backtick code spans or Unicode symbols for math. + +NEVER expose internal tool parameter names, backend IDs, or implementation details to the user. Always use natural, user-friendly language instead. diff --git a/surfsense_backend/app/agents/new_chat/prompts/base/citations_off.md b/surfsense_backend/app/agents/new_chat/prompts/base/citations_off.md new file mode 100644 index 000000000..8288886e9 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/base/citations_off.md @@ -0,0 +1,16 @@ + +IMPORTANT: Citations are DISABLED for this configuration. + +DO NOT include any citations in your responses. Specifically: +1. Do NOT use the [citation:chunk_id] format anywhere in your response. +2. Do NOT reference document IDs, chunk IDs, or source IDs. +3. Simply provide the information naturally without any citation markers. +4. Write your response as if you're having a normal conversation, incorporating the information from your knowledge seamlessly. + +When answering questions based on documents from the knowledge base: +- Present the information directly and confidently +- Do not mention that information comes from specific documents or chunks +- Integrate facts naturally into your response without attribution markers + +Your goal is to provide helpful, informative answers in a clean, readable format without any citation notation. + diff --git a/surfsense_backend/app/agents/new_chat/prompts/base/citations_on.md b/surfsense_backend/app/agents/new_chat/prompts/base/citations_on.md new file mode 100644 index 000000000..56291bf3e --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/base/citations_on.md @@ -0,0 +1,90 @@ + +CRITICAL CITATION REQUIREMENTS: + +1. For EVERY piece of information you include from the documents, add a citation in the format [citation:chunk_id] where chunk_id is the exact value from the `` tag inside ``. +2. Make sure ALL factual statements from the documents have proper citations. +3. If multiple chunks support the same point, include all relevant citations [citation:chunk_id1], [citation:chunk_id2]. +4. You MUST use the exact chunk_id values from the `` attributes. Do not create your own citation numbers. +5. Every citation MUST be in the format [citation:chunk_id] where chunk_id is the exact chunk id value. +6. Never modify or change the chunk_id - always use the original values exactly as provided in the chunk tags. +7. Do not return citations as clickable links. +8. Never format citations as markdown links like "([citation:5](https://example.com))". Always use plain square brackets only. +9. Citations must ONLY appear as [citation:chunk_id] or [citation:chunk_id1], [citation:chunk_id2] format - never with parentheses, hyperlinks, or other formatting. +10. Never make up chunk IDs. Only use chunk_id values that are explicitly provided in the `` tags. +11. If you are unsure about a chunk_id, do not include a citation rather than guessing or making one up. + + +The documents you receive are structured like this: + +**Knowledge base documents (numeric chunk IDs):** + + + 42 + GITHUB_CONNECTOR + <![CDATA[Some repo / file / issue title]]> + + + + + + + + + + +**Web search results (URL chunk IDs):** + + + WEB_SEARCH + <![CDATA[Some web search result]]> + + + + + + + + +IMPORTANT: You MUST cite using the EXACT chunk ids from the `` tags. +- For knowledge base documents, chunk ids are numeric (e.g. 123, 124) or prefixed (e.g. doc-45). +- For live web search results, chunk ids are URLs (e.g. https://example.com/article). +Do NOT cite document_id. Always use the chunk id. + + + +- Every fact from the documents must have a citation in the format [citation:chunk_id] where chunk_id is the EXACT id value from a `` tag +- Citations should appear at the end of the sentence containing the information they support +- Multiple citations should be separated by commas: [citation:chunk_id1], [citation:chunk_id2], [citation:chunk_id3] +- No need to return references section. Just citations in answer. +- NEVER create your own citation format - use the exact chunk_id values from the documents in the [citation:chunk_id] format +- NEVER format citations as clickable links or as markdown links like "([citation:5](https://example.com))". Always use plain square brackets only +- NEVER make up chunk IDs if you are unsure about the chunk_id. It is better to omit the citation than to guess +- Copy the EXACT chunk id from the XML - if it says ``, use [citation:doc-123] +- If the chunk id is a URL like ``, use [citation:https://example.com/page] + + + +CORRECT citation formats: +- [citation:5] (numeric chunk ID from knowledge base) +- [citation:doc-123] (for Surfsense documentation chunks) +- [citation:https://example.com/article] (URL chunk ID from web search results) +- [citation:chunk_id1], [citation:chunk_id2], [citation:chunk_id3] (multiple citations) + +INCORRECT citation formats (DO NOT use): +- Using parentheses and markdown links: ([citation:5](https://github.com/MODSetter/SurfSense)) +- Using parentheses around brackets: ([citation:5]) +- Using hyperlinked text: [link to source 5](https://example.com) +- Using footnote style: ... library¹ +- Making up source IDs when source_id is unknown +- Using old IEEE format: [1], [2], [3] +- Using source types instead of IDs: [citation:GITHUB_CONNECTOR] instead of [citation:5] + + + +Based on your GitHub repositories and video content, Python's asyncio library provides tools for writing concurrent code using the async/await syntax [citation:5]. It's particularly useful for I/O-bound and high-level structured network code [citation:5]. + +According to web search results, the key advantage of asyncio is that it can improve performance by allowing other code to run while waiting for I/O operations to complete [citation:https://docs.python.org/3/library/asyncio.html]. This makes it excellent for scenarios like web scraping, API calls, database operations, or any situation where your program spends time waiting for external resources. + +However, from your video learning, it's important to note that asyncio is not suitable for CPU-bound tasks as it runs on a single thread [citation:12]. For computationally intensive work, you'd want to use multiprocessing instead. + + diff --git a/surfsense_backend/app/agents/new_chat/prompts/base/kb_only_policy_private.md b/surfsense_backend/app/agents/new_chat/prompts/base/kb_only_policy_private.md new file mode 100644 index 000000000..9cc767e7e --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/base/kb_only_policy_private.md @@ -0,0 +1,15 @@ + +CRITICAL RULE — KNOWLEDGE BASE FIRST, NEVER DEFAULT TO GENERAL KNOWLEDGE: +- You MUST answer questions ONLY using information retrieved from the user's knowledge base, web search results, scraped webpages, or other tool outputs. +- You MUST NOT answer factual or informational questions from your own training data or general knowledge unless the user explicitly grants permission. +- If the knowledge base search returns no relevant results AND no other tool provides the answer, you MUST: + 1. Inform the user that you could not find relevant information in their knowledge base. + 2. Ask the user: "Would you like me to answer from my general knowledge instead?" + 3. ONLY provide a general-knowledge answer AFTER the user explicitly says yes. +- This policy does NOT apply to: + * Casual conversation, greetings, or meta-questions about SurfSense itself (e.g., "what can you do?") + * Formatting, summarization, or analysis of content already present in the conversation + * Following user instructions that are clearly task-oriented (e.g., "rewrite this in bullet points") + * Tool-usage actions like generating reports, podcasts, images, or scraping webpages + * Queries about services that have direct tools (Linear, ClickUp, Jira, Slack, Airtable) — see below + diff --git a/surfsense_backend/app/agents/new_chat/prompts/base/kb_only_policy_team.md b/surfsense_backend/app/agents/new_chat/prompts/base/kb_only_policy_team.md new file mode 100644 index 000000000..1d806dbae --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/base/kb_only_policy_team.md @@ -0,0 +1,15 @@ + +CRITICAL RULE — KNOWLEDGE BASE FIRST, NEVER DEFAULT TO GENERAL KNOWLEDGE: +- You MUST answer questions ONLY using information retrieved from the team's shared knowledge base, web search results, scraped webpages, or other tool outputs. +- You MUST NOT answer factual or informational questions from your own training data or general knowledge unless a team member explicitly grants permission. +- If the knowledge base search returns no relevant results AND no other tool provides the answer, you MUST: + 1. Inform the team that you could not find relevant information in the shared knowledge base. + 2. Ask: "Would you like me to answer from my general knowledge instead?" + 3. ONLY provide a general-knowledge answer AFTER a team member explicitly says yes. +- This policy does NOT apply to: + * Casual conversation, greetings, or meta-questions about SurfSense itself (e.g., "what can you do?") + * Formatting, summarization, or analysis of content already present in the conversation + * Following user instructions that are clearly task-oriented (e.g., "rewrite this in bullet points") + * Tool-usage actions like generating reports, podcasts, images, or scraping webpages + * Queries about services that have direct tools (Linear, ClickUp, Jira, Slack, Airtable) — see below + diff --git a/surfsense_backend/app/agents/new_chat/prompts/base/memory_protocol_private.md b/surfsense_backend/app/agents/new_chat/prompts/base/memory_protocol_private.md new file mode 100644 index 000000000..8f7da14f8 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/base/memory_protocol_private.md @@ -0,0 +1,6 @@ + +IMPORTANT — After understanding each user message, ALWAYS check: does this message +reveal durable facts about the user (role, interests, preferences, projects, +background, or standing instructions)? If yes, you MUST call update_memory +alongside your normal response — do not defer this to a later turn. + diff --git a/surfsense_backend/app/agents/new_chat/prompts/base/memory_protocol_team.md b/surfsense_backend/app/agents/new_chat/prompts/base/memory_protocol_team.md new file mode 100644 index 000000000..61d89cc5d --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/base/memory_protocol_team.md @@ -0,0 +1,6 @@ + +IMPORTANT — After understanding each user message, ALWAYS check: does this message +reveal durable facts about the team (decisions, conventions, architecture, processes, +or key facts)? If yes, you MUST call update_memory alongside your normal response — +do not defer this to a later turn. + diff --git a/surfsense_backend/app/agents/new_chat/prompts/base/parameter_resolution.md b/surfsense_backend/app/agents/new_chat/prompts/base/parameter_resolution.md new file mode 100644 index 000000000..77be4d87c --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/base/parameter_resolution.md @@ -0,0 +1,39 @@ + +Some service tools require identifiers or context you do not have (account IDs, +workspace names, channel IDs, project keys, etc.). NEVER ask the user for raw +IDs or technical identifiers — they cannot memorise them. + +Instead, follow this discovery pattern: +1. Call a listing/discovery tool to find available options. +2. ONE result → use it silently, no question to the user. +3. MULTIPLE results → present the options by their display names and let the + user choose. Never show raw UUIDs — always use friendly names. + +Discovery tools by level: +- Which account/workspace? → get_connected_accounts("") +- Which Jira site (cloudId)? → getAccessibleAtlassianResources +- Which Jira project? → getVisibleJiraProjects (after resolving cloudId) +- Which Jira issue type? → getJiraProjectIssueTypesMetadata (after resolving project) +- Which channel? → slack_search_channels +- Which base? → list_bases +- Which table? → list_tables_for_base (after resolving baseId) +- Which task? → clickup_search +- Which issue? → list_issues (Linear) or searchJiraIssuesUsingJql (Jira) + +For Jira specifically: ALWAYS call getAccessibleAtlassianResources first to +obtain the cloudId, then pass it to other Jira tools. When creating an issue, +chain: getAccessibleAtlassianResources → getVisibleJiraProjects → createJiraIssue. +If there is only one option at each step, use it silently. If multiple, present +friendly names. + +Chain discovery when needed — e.g. for Airtable records: list_bases → pick +base → list_tables_for_base → pick table → list_records_for_table. + +MULTI-ACCOUNT TOOL NAMING: When the user has multiple accounts connected for +the same service, tool names are prefixed to avoid collisions — e.g. +linear_25_list_issues and linear_30_list_issues instead of two list_issues. +Each prefixed tool's description starts with [Account: ] so you +know which account it targets. Use get_connected_accounts("") to see +the full list of accounts with their connector IDs and display names. +When only one account is connected, tools have their normal unprefixed names. + diff --git a/surfsense_backend/app/agents/new_chat/prompts/base/tool_routing_private.md b/surfsense_backend/app/agents/new_chat/prompts/base/tool_routing_private.md new file mode 100644 index 000000000..ec667bf88 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/base/tool_routing_private.md @@ -0,0 +1,16 @@ + +CRITICAL — You have direct tools for these services: Linear, ClickUp, Jira, Slack, Airtable. +Their data is NEVER in the knowledge base. You MUST call their tools immediately — never +say "I don't see it in the knowledge base" or ask the user if they want you to check. +Ignore any knowledge base results for these services. + +When to use which tool: +- Linear (issues) → list_issues, get_issue, save_issue (create/update) +- ClickUp (tasks) → clickup_search, clickup_get_task +- Jira (issues) → getAccessibleAtlassianResources (cloudId discovery), getVisibleJiraProjects (project discovery), getJiraProjectIssueTypesMetadata (issue type discovery), searchJiraIssuesUsingJql, createJiraIssue, editJiraIssue +- Slack (messages, channels) → slack_search_channels, slack_read_channel, slack_read_thread +- Airtable (bases, tables, records) → list_bases, list_tables_for_base, list_records_for_table +- Knowledge base content (Notion, GitHub, files, notes) → automatically searched +- Real-time public web data → call web_search +- Reading a specific webpage → call scrape_webpage + diff --git a/surfsense_backend/app/agents/new_chat/prompts/base/tool_routing_team.md b/surfsense_backend/app/agents/new_chat/prompts/base/tool_routing_team.md new file mode 100644 index 000000000..48b7a990b --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/base/tool_routing_team.md @@ -0,0 +1,16 @@ + +CRITICAL — You have direct tools for these services: Linear, ClickUp, Jira, Slack, Airtable. +Their data is NEVER in the knowledge base. You MUST call their tools immediately — never +say "I don't see it in the knowledge base" or ask if they want you to check. +Ignore any knowledge base results for these services. + +When to use which tool: +- Linear (issues) → list_issues, get_issue, save_issue (create/update) +- ClickUp (tasks) → clickup_search, clickup_get_task +- Jira (issues) → getAccessibleAtlassianResources (cloudId discovery), getVisibleJiraProjects (project discovery), getJiraProjectIssueTypesMetadata (issue type discovery), searchJiraIssuesUsingJql, createJiraIssue, editJiraIssue +- Slack (messages, channels) → slack_search_channels, slack_read_channel, slack_read_thread +- Airtable (bases, tables, records) → list_bases, list_tables_for_base, list_records_for_table +- Knowledge base content (Notion, GitHub, files, notes) → automatically searched +- Real-time public web data → call web_search +- Reading a specific webpage → call scrape_webpage + diff --git a/surfsense_backend/app/agents/new_chat/prompts/composer.py b/surfsense_backend/app/agents/new_chat/prompts/composer.py new file mode 100644 index 000000000..44060f75f --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/composer.py @@ -0,0 +1,359 @@ +""" +Prompt composer for the SurfSense ``new_chat`` agent. + +This module assembles the agent's system prompt from the markdown fragments +under :mod:`app.agents.new_chat.prompts`. It replaces the monolithic +``system_prompt.py`` with a clean, fragment-based composition: + +:: + + prompts/ + base/ # agent identity, KB policy, tool routing, … + providers/ # provider-specific tweaks (anthropic, gpt5, …) + tools/ # one ``.md`` per tool + examples/ # one ``.md`` per tool with call examples + routing/ # connector-specific routing notes (linear, slack, …) + +Tier 3a in the OpenCode-port plan. + +Backwards compatibility +======================= + +``system_prompt.py`` re-exports :func:`compose_system_prompt` and wraps it +in functions with the same signatures as the legacy +``build_surfsense_system_prompt`` / ``build_configurable_system_prompt`` so +existing call sites do not change. +""" + +from __future__ import annotations + +import re +from collections.abc import Iterable +from datetime import UTC, datetime +from importlib import resources + +from app.db import ChatVisibility + +# ----------------------------------------------------------------------------- +# Provider variant detection +# ----------------------------------------------------------------------------- + +ProviderVariant = str # "anthropic" | "openai_reasoning" | "openai_classic" | "google" | "default" + +_OPENAI_REASONING_RE = re.compile(r"\b(gpt-5|o\d|o-)", re.IGNORECASE) +_OPENAI_CLASSIC_RE = re.compile(r"\bgpt-4", re.IGNORECASE) +_ANTHROPIC_RE = re.compile(r"\bclaude\b", re.IGNORECASE) +_GOOGLE_RE = re.compile(r"\bgemini\b", re.IGNORECASE) + + +def detect_provider_variant(model_name: str | None) -> ProviderVariant: + """Pick a provider-specific prompt variant from a model id string. + + Heuristic match on the model id; returns ``"default"`` when nothing + matches so the composer can fall back to the empty placeholder file. + """ + if not model_name: + return "default" + name = model_name.strip() + if _OPENAI_REASONING_RE.search(name): + return "openai_reasoning" + if _OPENAI_CLASSIC_RE.search(name): + return "openai_classic" + if _ANTHROPIC_RE.search(name): + return "anthropic" + if _GOOGLE_RE.search(name): + return "google" + return "default" + + +# ----------------------------------------------------------------------------- +# Fragment loading +# ----------------------------------------------------------------------------- + + +_PROMPTS_PACKAGE = "app.agents.new_chat.prompts" + + +def _read_fragment(subpath: str) -> str: + """Read a fragment file from the ``prompts/`` resource tree. + + Returns the raw contents stripped of any single trailing newline so + composition can append explicit separators without compounding blank + lines. Missing files return an empty string so optional fragments + (e.g. provider hints) act as no-ops. + """ + parts = subpath.split("/") + try: + ref = resources.files(_PROMPTS_PACKAGE).joinpath(*parts) + if not ref.is_file(): + return "" + text = ref.read_text(encoding="utf-8") + except (FileNotFoundError, ModuleNotFoundError): + return "" + if text.endswith("\n"): + text = text[:-1] + return text + + +# ----------------------------------------------------------------------------- +# Tool ordering + memory variant resolution +# ----------------------------------------------------------------------------- + + +# Ordered for reading flow: fundamentals first, then artifact generators, +# then memory at the end (mirrors the legacy ``_ALL_TOOL_NAMES_ORDERED``). +ALL_TOOL_NAMES_ORDERED: tuple[str, ...] = ( + "search_surfsense_docs", + "web_search", + "generate_podcast", + "generate_video_presentation", + "generate_report", + "generate_resume", + "generate_image", + "scrape_webpage", + "update_memory", +) + + +_MEMORY_VARIANT_TOOLS: frozenset[str] = frozenset({"update_memory"}) + + +def _tool_fragment_path(tool_name: str, variant: str) -> str: + """Resolve a tool's instruction fragment path. + + Tools listed in :data:`_MEMORY_VARIANT_TOOLS` switch on the conversation + visibility and load ``tools/_.md``; everything else + falls back to ``tools/.md``. + """ + if tool_name in _MEMORY_VARIANT_TOOLS: + return f"tools/{tool_name}_{variant}.md" + return f"tools/{tool_name}.md" + + +def _example_fragment_path(tool_name: str, variant: str) -> str: + if tool_name in _MEMORY_VARIANT_TOOLS: + return f"examples/{tool_name}_{variant}.md" + return f"examples/{tool_name}.md" + + +def _format_tool_label(tool_name: str) -> str: + return tool_name.replace("_", " ").title() + + +# ----------------------------------------------------------------------------- +# Section builders +# ----------------------------------------------------------------------------- + + +def _build_system_instructions( + *, + visibility: ChatVisibility, + resolved_today: str, +) -> str: + """Reconstruct the legacy ```` block from fragments.""" + variant = "team" if visibility == ChatVisibility.SEARCH_SPACE else "private" + + sections = [ + _read_fragment(f"base/agent_{variant}.md"), + _read_fragment(f"base/kb_only_policy_{variant}.md"), + _read_fragment(f"base/tool_routing_{variant}.md"), + _read_fragment("base/parameter_resolution.md"), + _read_fragment(f"base/memory_protocol_{variant}.md"), + ] + body = "\n\n".join(s for s in sections if s) + block = f"\n\n{body}\n\n\n" + return block.format(resolved_today=resolved_today) + + +def _build_mcp_routing_block( + mcp_connector_tools: dict[str, list[str]] | None, +) -> str: + """Emit the ```` block when at least one MCP server is wired.""" + if not mcp_connector_tools: + return "" + lines: list[str] = [ + "\n", + "You also have direct tools from these user-connected MCP servers.", + "Their data is NEVER in the knowledge base — call their tools directly.", + "", + ] + for server_name, tool_names in mcp_connector_tools.items(): + lines.append(f"- {server_name} → {', '.join(tool_names)}") + lines.append("\n") + return "\n".join(lines) + + +def _build_tools_section( + *, + visibility: ChatVisibility, + enabled_tool_names: set[str] | None, + disabled_tool_names: set[str] | None, +) -> str: + """Reconstruct the ```` block + ```` block.""" + variant = "team" if visibility == ChatVisibility.SEARCH_SPACE else "private" + + parts: list[str] = [] + preamble = _read_fragment("tools/_preamble.md") + if preamble: + parts.append(preamble + "\n") + + examples: list[str] = [] + + for tool_name in ALL_TOOL_NAMES_ORDERED: + if enabled_tool_names is not None and tool_name not in enabled_tool_names: + continue + + instruction = _read_fragment(_tool_fragment_path(tool_name, variant)) + if instruction: + parts.append(instruction + "\n") + + example = _read_fragment(_example_fragment_path(tool_name, variant)) + if example: + examples.append(example + "\n") + + known_disabled = ( + set(disabled_tool_names) & set(ALL_TOOL_NAMES_ORDERED) + if disabled_tool_names + else set() + ) + if known_disabled: + disabled_list = ", ".join( + _format_tool_label(n) + for n in ALL_TOOL_NAMES_ORDERED + if n in known_disabled + ) + parts.append( + "\n" + "DISABLED TOOLS (by user):\n" + f"The following tools are available in SurfSense but have been disabled by the user for this session: {disabled_list}.\n" + "You do NOT have access to these tools and MUST NOT claim you can use them.\n" + "If the user asks about a capability provided by a disabled tool, let them know the relevant tool\n" + "is currently disabled and they can re-enable it.\n" + ) + + parts.append("\n\n") + + if examples: + parts.append("") + parts.extend(examples) + parts.append("\n") + + return "".join(parts) + + +def _build_provider_block(provider_variant: ProviderVariant) -> str: + """Optional provider-tuned hints. Empty for ``"default"``.""" + if not provider_variant or provider_variant == "default": + return "" + text = _read_fragment(f"providers/{provider_variant}.md") + return f"\n{text}\n" if text else "" + + +def _build_routing_block(connector_routing: Iterable[str] | None) -> str: + if not connector_routing: + return "" + fragments: list[str] = [] + for name in connector_routing: + text = _read_fragment(f"routing/{name}.md") + if text: + fragments.append(text) + if not fragments: + return "" + return "\n" + "\n\n".join(fragments) + "\n" + + +def _build_citation_block(citations_enabled: bool) -> str: + fragment = ( + _read_fragment("base/citations_on.md") + if citations_enabled + else _read_fragment("base/citations_off.md") + ) + return f"\n{fragment}\n" if fragment else "" + + +# ----------------------------------------------------------------------------- +# Public API +# ----------------------------------------------------------------------------- + + +def compose_system_prompt( + *, + today: datetime | None = None, + thread_visibility: ChatVisibility | None = None, + enabled_tool_names: set[str] | None = None, + disabled_tool_names: set[str] | None = None, + mcp_connector_tools: dict[str, list[str]] | None = None, + custom_system_instructions: str | None = None, + use_default_system_instructions: bool = True, + citations_enabled: bool = True, + provider_variant: ProviderVariant | None = None, + model_name: str | None = None, + connector_routing: Iterable[str] | None = None, +) -> str: + """Assemble the SurfSense system prompt from disk fragments. + + Args: + today: Optional clock injection for tests. + thread_visibility: Private vs shared (team) — drives memory wording + and a few base block variants. + enabled_tool_names: When provided, only these tools' instructions + are included; ``None`` keeps the legacy "include everything" + behavior. + disabled_tool_names: User-disabled tools (note appended to prompt). + mcp_connector_tools: ``{server_name: [tool_names...]}`` to inject + an explicit MCP routing block. + custom_system_instructions: Free-form instructions that override + the default ```` block (legacy support + for ``NewLLMConfig.system_instructions``). + use_default_system_instructions: When ``custom_system_instructions`` + is empty/None, fall back to defaults (legacy semantics). + citations_enabled: Include ``citations_on.md`` (true) or + ``citations_off.md`` (false). + provider_variant: Explicit provider variant override + (``"anthropic" | "openai_reasoning" | "openai_classic" | "google" | "default"``). + When ``None``, falls back to :func:`detect_provider_variant` + on ``model_name``. + model_name: Used to auto-detect ``provider_variant`` when not + provided explicitly. + connector_routing: Optional list of routing fragment names + (``["linear", "slack", ...]``) to include from + ``prompts/routing/``. + + Returns: + The fully composed system prompt string. + """ + resolved_today = (today or datetime.now(UTC)).astimezone(UTC).date().isoformat() + visibility = thread_visibility or ChatVisibility.PRIVATE + + if custom_system_instructions and custom_system_instructions.strip(): + sys_block = custom_system_instructions.format(resolved_today=resolved_today) + elif use_default_system_instructions: + sys_block = _build_system_instructions( + visibility=visibility, resolved_today=resolved_today + ) + else: + sys_block = "" + + sys_block += _build_mcp_routing_block(mcp_connector_tools) + + if provider_variant is None: + provider_variant = detect_provider_variant(model_name) + sys_block += _build_provider_block(provider_variant) + sys_block += _build_routing_block(connector_routing) + + tools_block = _build_tools_section( + visibility=visibility, + enabled_tool_names=enabled_tool_names, + disabled_tool_names=disabled_tool_names, + ) + citation_block = _build_citation_block(citations_enabled) + + return sys_block + tools_block + citation_block + + +__all__ = [ + "ALL_TOOL_NAMES_ORDERED", + "ProviderVariant", + "compose_system_prompt", + "detect_provider_variant", +] diff --git a/surfsense_backend/app/agents/new_chat/prompts/examples/__init__.py b/surfsense_backend/app/agents/new_chat/prompts/examples/__init__.py new file mode 100644 index 000000000..8b1378917 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/examples/__init__.py @@ -0,0 +1 @@ + diff --git a/surfsense_backend/app/agents/new_chat/prompts/examples/generate_image.md b/surfsense_backend/app/agents/new_chat/prompts/examples/generate_image.md new file mode 100644 index 000000000..216c2926a --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/examples/generate_image.md @@ -0,0 +1,12 @@ + +- User: "Generate an image of a cat" + - Call: `generate_image(prompt="A fluffy orange tabby cat sitting on a windowsill, bathed in warm golden sunlight, soft bokeh background with green houseplants, photorealistic style, cozy atmosphere")` + - The generated image will automatically be displayed in the chat. +- User: "Draw me a logo for a coffee shop called Bean Dream" + - Call: `generate_image(prompt="Minimalist modern logo design for a coffee shop called 'Bean Dream', featuring a stylized coffee bean with dream-like swirls of steam, clean vector style, warm brown and cream color palette, white background, professional branding")` + - The generated image will automatically be displayed in the chat. +- User: "Show me this image: https://example.com/image.png" + - Simply include it in your response using markdown: `![Image](https://example.com/image.png)` +- User uploads an image file and asks: "What is this image about?" + - The user's uploaded image is already visible in the chat. + - Simply analyze the image content and respond directly. diff --git a/surfsense_backend/app/agents/new_chat/prompts/examples/generate_podcast.md b/surfsense_backend/app/agents/new_chat/prompts/examples/generate_podcast.md new file mode 100644 index 000000000..aabf8ce7a --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/examples/generate_podcast.md @@ -0,0 +1,7 @@ + +- User: "Give me a podcast about AI trends based on what we discussed" + - First search for relevant content, then call: `generate_podcast(source_content="Based on our conversation and search results: [detailed summary of chat + search findings]", podcast_title="AI Trends Podcast")` +- User: "Create a podcast summary of this conversation" + - Call: `generate_podcast(source_content="Complete conversation summary:\n\nUser asked about [topic 1]:\n[Your detailed response]\n\nUser then asked about [topic 2]:\n[Your detailed response]\n\n[Continue for all exchanges in the conversation]", podcast_title="Conversation Summary")` +- User: "Make a podcast about quantum computing" + - First explore `/documents/` (ls/glob/grep/read_file), then: `generate_podcast(source_content="Key insights about quantum computing from retrieved files:\n\n[Comprehensive summary of findings]", podcast_title="Quantum Computing Explained")` diff --git a/surfsense_backend/app/agents/new_chat/prompts/examples/generate_report.md b/surfsense_backend/app/agents/new_chat/prompts/examples/generate_report.md new file mode 100644 index 000000000..7e9d0a595 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/examples/generate_report.md @@ -0,0 +1,13 @@ + +- User: "Generate a report about AI trends" + - Call: `generate_report(topic="AI Trends Report", source_strategy="kb_search", search_queries=["AI trends recent developments", "artificial intelligence industry trends", "AI market growth and predictions"], report_style="detailed")` + - WHY: Has creation verb "generate" → call the tool. No prior discussion → use kb_search. +- User: "Write a research report from this conversation" + - Call: `generate_report(topic="Research Report", source_strategy="conversation", source_content="Complete conversation summary:\n\n...", report_style="deep_research")` + - WHY: Has creation verb "write" → call the tool. Conversation has the content → use source_strategy="conversation". +- User: (after a report on Climate Change was generated) "Add a section about carbon capture technologies" + - Call: `generate_report(topic="Climate Crisis: Causes, Impacts, and Solutions", source_strategy="conversation", source_content="[summary of conversation context if any]", parent_report_id=, user_instructions="Add a new section about carbon capture technologies")` + - WHY: Has modification verb "add" + specific deliverable target → call the tool with parent_report_id. +- User: (after a report was generated) "What else could we add to have more depth?" + - Do NOT call generate_report. Answer in chat with suggestions. + - WHY: No creation/modification verb directed at producing a deliverable. diff --git a/surfsense_backend/app/agents/new_chat/prompts/examples/generate_resume.md b/surfsense_backend/app/agents/new_chat/prompts/examples/generate_resume.md new file mode 100644 index 000000000..d8a6c381e --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/examples/generate_resume.md @@ -0,0 +1,19 @@ + +- User: "Build me a resume. I'm John Doe, engineer at Acme Corp..." + - Call: `generate_resume(user_info="John Doe, engineer at Acme Corp...", max_pages=1)` + - WHY: Has creation verb "build" + resume → call the tool. +- User: "Create my CV with this info: [experience, education, skills]" + - Call: `generate_resume(user_info="[experience, education, skills]", max_pages=1)` +- User: "Build me a resume" (and there is a resume/CV document in the conversation context) + - Extract the FULL content from the document in context, then call: + `generate_resume(user_info="Name: John Doe\nEmail: john@example.com\n\nExperience:\n- Senior Engineer at Acme Corp (2020-2024)\n Led team of 5...\n\nEducation:\n- BS Computer Science, MIT (2016-2020)\n\nSkills: Python, TypeScript, AWS...", max_pages=1)` + - WHY: Document content is available in context — extract ALL of it into user_info. Do NOT ignore referenced documents. +- User: (after resume generated) "Change my title to Senior Engineer" + - Call: `generate_resume(user_info="", user_instructions="Change the job title to Senior Engineer", parent_report_id=, max_pages=1)` + - WHY: Modification verb "change" + refers to existing resume → set parent_report_id. +- User: (after resume generated) "Make this 2 pages and expand projects" + - Call: `generate_resume(user_info="", user_instructions="Expand projects and keep this to at most 2 pages", parent_report_id=, max_pages=2)` + - WHY: Explicit page increase request → set max_pages to 2. +- User: "How should I structure my resume?" + - Do NOT call generate_resume. Answer in chat with advice. + - WHY: No creation/modification verb. diff --git a/surfsense_backend/app/agents/new_chat/prompts/examples/generate_video_presentation.md b/surfsense_backend/app/agents/new_chat/prompts/examples/generate_video_presentation.md new file mode 100644 index 000000000..257ec86cf --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/examples/generate_video_presentation.md @@ -0,0 +1,7 @@ + +- User: "Give me a presentation about AI trends based on what we discussed" + - First search for relevant content, then call: `generate_video_presentation(source_content="Based on our conversation and search results: [detailed summary of chat + search findings]", video_title="AI Trends Presentation")` +- User: "Create slides summarizing this conversation" + - Call: `generate_video_presentation(source_content="Complete conversation summary:\n\nUser asked about [topic 1]:\n[Your detailed response]\n\nUser then asked about [topic 2]:\n[Your detailed response]\n\n[Continue for all exchanges in the conversation]", video_title="Conversation Summary")` +- User: "Make a video presentation about quantum computing" + - First explore `/documents/` (ls/glob/grep/read_file), then: `generate_video_presentation(source_content="Key insights about quantum computing from retrieved files:\n\n[Comprehensive summary of findings]", video_title="Quantum Computing Explained")` diff --git a/surfsense_backend/app/agents/new_chat/prompts/examples/scrape_webpage.md b/surfsense_backend/app/agents/new_chat/prompts/examples/scrape_webpage.md new file mode 100644 index 000000000..0f156bf24 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/examples/scrape_webpage.md @@ -0,0 +1,13 @@ + +- User: "Check out https://dev.to/some-article" + - Call: `scrape_webpage(url="https://dev.to/some-article")` + - Respond with a structured analysis — key points, takeaways. +- User: "Read this article and summarize it for me: https://example.com/blog/ai-trends" + - Call: `scrape_webpage(url="https://example.com/blog/ai-trends")` + - Respond with a thorough summary using headings and bullet points. +- User: (after discussing https://example.com/stats) "Can you get the live data from that page?" + - Call: `scrape_webpage(url="https://example.com/stats")` + - IMPORTANT: Always attempt scraping first. Never refuse before trying the tool. +- User: "https://example.com/blog/weekend-recipes" + - Call: `scrape_webpage(url="https://example.com/blog/weekend-recipes")` + - When a user sends just a URL with no instructions, scrape it and provide a concise summary of the content. diff --git a/surfsense_backend/app/agents/new_chat/prompts/examples/search_surfsense_docs.md b/surfsense_backend/app/agents/new_chat/prompts/examples/search_surfsense_docs.md new file mode 100644 index 000000000..b90f2b7a7 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/examples/search_surfsense_docs.md @@ -0,0 +1,9 @@ + +- User: "How do I install SurfSense?" + - Call: `search_surfsense_docs(query="installation setup")` +- User: "What connectors does SurfSense support?" + - Call: `search_surfsense_docs(query="available connectors integrations")` +- User: "How do I set up the Notion connector?" + - Call: `search_surfsense_docs(query="Notion connector setup configuration")` +- User: "How do I use Docker to run SurfSense?" + - Call: `search_surfsense_docs(query="Docker installation setup")` diff --git a/surfsense_backend/app/agents/new_chat/prompts/examples/update_memory_private.md b/surfsense_backend/app/agents/new_chat/prompts/examples/update_memory_private.md new file mode 100644 index 000000000..f83fe40b4 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/examples/update_memory_private.md @@ -0,0 +1,16 @@ + +- Alex, is empty. User: "I'm a space enthusiast, explain astrophage to me" + - The user casually shared a durable fact. Use their first name in the entry, short neutral heading: + update_memory(updated_memory="## Interests & background\n- (2025-03-15) [fact] Alex is a space enthusiast\n") +- User: "Remember that I prefer concise answers over detailed explanations" + - Durable preference. Merge with existing memory, add a new heading: + update_memory(updated_memory="## Interests & background\n- (2025-03-15) [fact] Alex is a space enthusiast\n\n## Response style\n- (2025-03-15) [pref] Alex prefers concise answers over detailed explanations\n") +- User: "I actually moved to Tokyo last month" + - Updated fact, date prefix reflects when recorded: + update_memory(updated_memory="## Interests & background\n...\n\n## Personal context\n- (2025-03-15) [fact] Alex lives in Tokyo (previously London)\n...") +- User: "I'm a freelance photographer working on a nature documentary" + - Durable background info under a fitting heading: + update_memory(updated_memory="...\n\n## Current focus\n- (2025-03-15) [fact] Alex is a freelance photographer\n- (2025-03-15) [fact] Alex is working on a nature documentary\n") +- User: "Always respond in bullet points" + - Standing instruction: + update_memory(updated_memory="...\n\n## Response style\n- (2025-03-15) [instr] Always respond to Alex in bullet points\n") diff --git a/surfsense_backend/app/agents/new_chat/prompts/examples/update_memory_team.md b/surfsense_backend/app/agents/new_chat/prompts/examples/update_memory_team.md new file mode 100644 index 000000000..1c74fdf6e --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/examples/update_memory_team.md @@ -0,0 +1,7 @@ + +- User: "Let's remember that we decided to do weekly standup meetings on Mondays" + - Durable team decision: + update_memory(updated_memory="- (2025-03-15) [fact] Weekly standup meetings on Mondays\n...") +- User: "Our office is in downtown Seattle, 5th floor" + - Durable team fact: + update_memory(updated_memory="- (2025-03-15) [fact] Office location: downtown Seattle, 5th floor\n...") diff --git a/surfsense_backend/app/agents/new_chat/prompts/examples/web_search.md b/surfsense_backend/app/agents/new_chat/prompts/examples/web_search.md new file mode 100644 index 000000000..6b9828ac7 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/examples/web_search.md @@ -0,0 +1,8 @@ + +- User: "What's the current USD to INR exchange rate?" + - Call: `web_search(query="current USD to INR exchange rate")` + - Then answer using the returned web results with citations. +- User: "What's the latest news about AI?" + - Call: `web_search(query="latest AI news today")` +- User: "What's the weather in New York?" + - Call: `web_search(query="weather New York today")` diff --git a/surfsense_backend/app/agents/new_chat/prompts/providers/__init__.py b/surfsense_backend/app/agents/new_chat/prompts/providers/__init__.py new file mode 100644 index 000000000..8b1378917 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/providers/__init__.py @@ -0,0 +1 @@ + diff --git a/surfsense_backend/app/agents/new_chat/prompts/providers/anthropic.md b/surfsense_backend/app/agents/new_chat/prompts/providers/anthropic.md new file mode 100644 index 000000000..6e22ef265 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/providers/anthropic.md @@ -0,0 +1,5 @@ + +You are running on an Anthropic Claude model. Use XML tags liberally to structure +intermediate reasoning when the task is complex. Prefer step-by-step plans inside +`` blocks before producing the final answer. + diff --git a/surfsense_backend/app/agents/new_chat/prompts/providers/default.md b/surfsense_backend/app/agents/new_chat/prompts/providers/default.md new file mode 100644 index 000000000..8b1378917 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/providers/default.md @@ -0,0 +1 @@ + diff --git a/surfsense_backend/app/agents/new_chat/prompts/providers/google.md b/surfsense_backend/app/agents/new_chat/prompts/providers/google.md new file mode 100644 index 000000000..4b31a8388 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/providers/google.md @@ -0,0 +1,4 @@ + +You are running on a Google Gemini model. Prefer concise, structured responses. +When using tools, follow the function-calling protocol and avoid verbose preludes. + diff --git a/surfsense_backend/app/agents/new_chat/prompts/providers/openai_classic.md b/surfsense_backend/app/agents/new_chat/prompts/providers/openai_classic.md new file mode 100644 index 000000000..7ea4366c4 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/providers/openai_classic.md @@ -0,0 +1,5 @@ + +You are running on a classic OpenAI chat model (GPT-4 family). Use direct +function-calling for tools. When editing files, use the standard `edit_file` +or `write_file` tools rather than diff-based patches. + diff --git a/surfsense_backend/app/agents/new_chat/prompts/providers/openai_reasoning.md b/surfsense_backend/app/agents/new_chat/prompts/providers/openai_reasoning.md new file mode 100644 index 000000000..935d3f207 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/providers/openai_reasoning.md @@ -0,0 +1,5 @@ + +You are running on an OpenAI reasoning model (o-series / GPT-5+). Be terse and +direct in your responses. When editing files, prefer the `apply_patch` tool format +where available. Avoid restating the user request before answering. + diff --git a/surfsense_backend/app/agents/new_chat/prompts/routing/__init__.py b/surfsense_backend/app/agents/new_chat/prompts/routing/__init__.py new file mode 100644 index 000000000..8b1378917 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/routing/__init__.py @@ -0,0 +1 @@ + diff --git a/surfsense_backend/app/agents/new_chat/prompts/routing/jira.md b/surfsense_backend/app/agents/new_chat/prompts/routing/jira.md new file mode 100644 index 000000000..8b1378917 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/routing/jira.md @@ -0,0 +1 @@ + diff --git a/surfsense_backend/app/agents/new_chat/prompts/routing/linear.md b/surfsense_backend/app/agents/new_chat/prompts/routing/linear.md new file mode 100644 index 000000000..8b1378917 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/routing/linear.md @@ -0,0 +1 @@ + diff --git a/surfsense_backend/app/agents/new_chat/prompts/routing/slack.md b/surfsense_backend/app/agents/new_chat/prompts/routing/slack.md new file mode 100644 index 000000000..8b1378917 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/routing/slack.md @@ -0,0 +1 @@ + diff --git a/surfsense_backend/app/agents/new_chat/prompts/tools/__init__.py b/surfsense_backend/app/agents/new_chat/prompts/tools/__init__.py new file mode 100644 index 000000000..8b1378917 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/tools/__init__.py @@ -0,0 +1 @@ + diff --git a/surfsense_backend/app/agents/new_chat/prompts/tools/_preamble.md b/surfsense_backend/app/agents/new_chat/prompts/tools/_preamble.md new file mode 100644 index 000000000..2c169e015 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/tools/_preamble.md @@ -0,0 +1,6 @@ + +You have access to the following tools: + +IMPORTANT: You can ONLY use the tools listed below. If a capability is not listed here, you do NOT have it. +Do NOT claim you can do something if the corresponding tool is not listed. + diff --git a/surfsense_backend/app/agents/new_chat/prompts/tools/generate_image.md b/surfsense_backend/app/agents/new_chat/prompts/tools/generate_image.md new file mode 100644 index 000000000..8bde13f22 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/tools/generate_image.md @@ -0,0 +1,11 @@ + +- generate_image: Generate images from text descriptions using AI image models. + - Use this when the user asks you to create, generate, draw, design, or make an image. + - Trigger phrases: "generate an image of", "create a picture of", "draw me", "make an image", "design a logo", "create artwork" + - Args: + - prompt: A detailed text description of the image to generate. Be specific about subject, style, colors, composition, and mood. + - n: Number of images to generate (1-4, default: 1) + - Returns: A dictionary with the generated image metadata. The image will automatically be displayed in the chat. + - IMPORTANT: Write a detailed, descriptive prompt for best results. Don't just pass the user's words verbatim - + expand and improve the prompt with specific details about style, lighting, composition, and mood. + - If the user's request is vague (e.g., "make me an image of a cat"), enhance the prompt with artistic details. diff --git a/surfsense_backend/app/agents/new_chat/prompts/tools/generate_podcast.md b/surfsense_backend/app/agents/new_chat/prompts/tools/generate_podcast.md new file mode 100644 index 000000000..58be143d7 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/tools/generate_podcast.md @@ -0,0 +1,15 @@ + +- generate_podcast: Generate an audio podcast from provided content. + - Use this when the user asks to create, generate, or make a podcast. + - Trigger phrases: "give me a podcast about", "create a podcast", "generate a podcast", "make a podcast", "turn this into a podcast" + - Args: + - source_content: The text content to convert into a podcast. This MUST be comprehensive and include: + * If discussing the current conversation: Include a detailed summary of the FULL chat history (all user questions and your responses) + * If based on knowledge base search: Include the key findings and insights from the search results + * You can combine both: conversation context + search results for richer podcasts + * The more detailed the source_content, the better the podcast quality + - podcast_title: Optional title for the podcast (default: "SurfSense Podcast") + - user_prompt: Optional instructions for podcast style/format (e.g., "Make it casual and fun") + - Returns: A task_id for tracking. The podcast will be generated in the background. + - IMPORTANT: Only one podcast can be generated at a time. If a podcast is already being generated, the tool will return status "already_generating". + - After calling this tool, inform the user that podcast generation has started and they will see the player when it's ready (takes 3-5 minutes). diff --git a/surfsense_backend/app/agents/new_chat/prompts/tools/generate_report.md b/surfsense_backend/app/agents/new_chat/prompts/tools/generate_report.md new file mode 100644 index 000000000..8a285a433 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/tools/generate_report.md @@ -0,0 +1,39 @@ + +- generate_report: Generate or revise a structured Markdown report artifact. + - WHEN TO CALL THIS TOOL — the message must contain a creation or modification VERB directed at producing a deliverable: + * Creation verbs: write, create, generate, draft, produce, summarize into, turn into, make + * Modification verbs: revise, update, expand, add (a section), rewrite, make (it shorter/longer/formal) + * Example triggers: "generate a report about...", "write a document on...", "add a section about budget", "make the report shorter", "rewrite in formal tone" + - WHEN NOT TO CALL THIS TOOL (answer in chat instead): + * Questions or discussion about the report: "What can we add?", "What's missing?", "Is the data accurate?", "How could this be improved?" + * Suggestions or brainstorming: "What other topics could be covered?", "What else could be added?", "What would make this better?" + * Asking for explanations: "Can you explain section 2?", "Why did you include that?", "What does this part mean?" + * Quick follow-ups or critiques: "Is the conclusion strong enough?", "Are there any gaps?", "What about the competitors?" + * THE TEST: Does the message contain a creation/modification VERB (from the list above) directed at producing or changing a deliverable? If NO verb → answer conversationally in chat. Do NOT assume the user wants a revision just because a report exists in the conversation. + - IMPORTANT FORMAT RULE: Reports are ALWAYS generated in Markdown. + - Args: + - topic: Short title for the report (max ~8 words). + - source_content: The text content to base the report on. + * For source_strategy="conversation" or "provided": Include a comprehensive summary of the relevant content. + * For source_strategy="kb_search": Can be empty or minimal — the tool handles searching internally. + * For source_strategy="auto": Include what you have; the tool searches KB if it's not enough. + - source_strategy: Controls how the tool collects source material. One of: + * "conversation" — The conversation already contains enough context (prior Q&A, discussion, pasted text, scraped pages). Pass a thorough summary as source_content. + * "kb_search" — The tool will search the knowledge base internally. Provide search_queries with 1-5 targeted queries. + * "auto" — Use source_content if sufficient, otherwise fall back to internal KB search using search_queries. + * "provided" — Use only what is in source_content (default, backward-compatible). + - search_queries: When source_strategy is "kb_search" or "auto", provide 1-5 specific search queries for the knowledge base. These should be precise, not just the topic name repeated. + - report_style: Controls report depth. Options: "detailed" (DEFAULT), "deep_research", "brief". + Use "brief" ONLY when the user explicitly asks for a short/concise/one-page report (e.g., "one page", "keep it short", "brief report", "500 words"). Default to "detailed" for all other requests. + - user_instructions: Optional specific instructions (e.g., "focus on financial impacts", "include recommendations"). When revising (parent_report_id set), describe WHAT TO CHANGE. If the user mentions a length preference (e.g., "one page", "500 words", "2 pages"), include that VERBATIM here AND set report_style="brief". + - parent_report_id: Set this to the report_id from a previous generate_report result when the user wants to MODIFY an existing report. Do NOT set it for new reports or questions about reports. + - Returns: A dictionary with status "ready" or "failed", report_id, title, and word_count. + - The report is generated immediately in Markdown and displayed inline in the chat. + - Export/download formats (PDF, DOCX, HTML, LaTeX, EPUB, ODT, plain text) are produced from the generated Markdown report. + - SOURCE STRATEGY DECISION (HIGH PRIORITY — follow this exactly): + * If the conversation already has substantive Q&A / discussion on the topic → use source_strategy="conversation" with a comprehensive summary as source_content. + * If the user wants a report on a topic not yet discussed → use source_strategy="kb_search" with targeted search_queries. + * If you have some content but might need more → use source_strategy="auto" with both source_content and search_queries. + * When revising an existing report (parent_report_id set) and the conversation has relevant context → use source_strategy="conversation". The revision will use the previous report content plus your source_content. + * NEVER run a separate KB lookup step and then pass those results to generate_report. The tool handles KB search internally. + - AFTER CALLING THIS TOOL: Do NOT repeat, summarize, or reproduce the report content in the chat. The report is already displayed as an interactive card that the user can open, read, copy, and export. Simply confirm that the report was generated (e.g., "I've generated your report on [topic]. You can view the Markdown report now, and export it in various formats from the card."). NEVER write out the report text in the chat. diff --git a/surfsense_backend/app/agents/new_chat/prompts/tools/generate_resume.md b/surfsense_backend/app/agents/new_chat/prompts/tools/generate_resume.md new file mode 100644 index 000000000..321ea90c9 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/tools/generate_resume.md @@ -0,0 +1,30 @@ + +- generate_resume: Generate or revise a professional resume as a Typst document. + - WHEN TO CALL: The user asks to create, build, generate, write, or draft a resume or CV. + Also when they ask to modify, update, or revise an existing resume from this conversation. + - WHEN NOT TO CALL: General career advice, resume tips, cover letters, or reviewing + a resume without making changes. For cover letters, use generate_report instead. + - The tool produces Typst source code that is compiled to a PDF preview automatically. + - PAGE POLICY: + - Default behavior is ONE PAGE. For new resume creation, set max_pages=1 unless the user explicitly asks for more. + - If the user requests a longer resume (e.g., "make it 2 pages"), set max_pages to that value. + - Args: + - user_info: The user's resume content — work experience, education, skills, contact + info, etc. Can be structured or unstructured text. + CRITICAL: user_info must be COMPREHENSIVE. Do NOT just pass the user's raw message. + You MUST gather and consolidate ALL available information: + * Content from referenced/mentioned documents (e.g., uploaded resumes, CVs, LinkedIn profiles) + that appear in the conversation context — extract and include their FULL content. + * Information the user shared across multiple messages in the conversation. + * Any relevant details from knowledge base search results in the context. + The more complete the user_info, the better the resume. Include names, contact info, + work experience with dates, education, skills, projects, certifications — everything available. + - user_instructions: Optional style or content preferences (e.g. "emphasize leadership", + "keep it to one page"). For revisions, describe what to change. + - parent_report_id: Set this when the user wants to MODIFY an existing resume from + this conversation. Use the report_id from a previous generate_resume result. + - max_pages: Maximum resume length in pages (integer 1-5). Default is 1. + - Returns: Dict with status, report_id, title, and content_type. + - After calling: Give a brief confirmation. Do NOT paste resume content in chat. Do NOT mention report_id or any internal IDs — the resume card is shown automatically. + - VERSIONING: Same rules as generate_report — set parent_report_id for modifications + of an existing resume, leave as None for new resumes. diff --git a/surfsense_backend/app/agents/new_chat/prompts/tools/generate_video_presentation.md b/surfsense_backend/app/agents/new_chat/prompts/tools/generate_video_presentation.md new file mode 100644 index 000000000..c3def88f2 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/tools/generate_video_presentation.md @@ -0,0 +1,9 @@ + +- generate_video_presentation: Generate a video presentation from provided content. + - Use this when the user asks to create a video, presentation, slides, or slide deck. + - Trigger phrases: "give me a presentation", "create slides", "generate a video", "make a slide deck", "turn this into a presentation" + - Args: + - source_content: The text content to turn into a presentation. The more detailed, the better. + - video_title: Optional title (default: "SurfSense Presentation") + - user_prompt: Optional style instructions (e.g., "Make it technical and detailed") + - After calling this tool, inform the user that generation has started and they will see the presentation when it's ready. diff --git a/surfsense_backend/app/agents/new_chat/prompts/tools/scrape_webpage.md b/surfsense_backend/app/agents/new_chat/prompts/tools/scrape_webpage.md new file mode 100644 index 000000000..46e299392 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/tools/scrape_webpage.md @@ -0,0 +1,30 @@ + +- scrape_webpage: Scrape and extract the main content from a webpage. + - Use this when the user wants you to READ and UNDERSTAND the actual content of a webpage. + - CRITICAL — WHEN TO USE (always attempt scraping, never refuse before trying): + * When a user asks to "get", "fetch", "pull", "grab", "scrape", or "read" content from a URL + * When the user wants live/dynamic data from a specific webpage (e.g., tables, scores, stats, prices) + * When a URL was mentioned earlier in the conversation and the user asks for its actual content + * When `/documents/` knowledge-base data is insufficient and the user wants more + - Trigger scenarios: + * "Read this article and summarize it" + * "What does this page say about X?" + * "Summarize this blog post for me" + * "Tell me the key points from this article" + * "What's in this webpage?" + * "Can you analyze this article?" + * "Can you get the live table/data from [URL]?" + * "Scrape it" / "Can you scrape that?" (referring to a previously mentioned URL) + * "Fetch the content from [URL]" + * "Pull the data from that page" + - Args: + - url: The URL of the webpage to scrape (must be HTTP/HTTPS) + - max_length: Maximum content length to return (default: 50000 chars) + - Returns: The page title, description, full content (in markdown), word count, and metadata + - After scraping, provide a comprehensive, well-structured summary with key takeaways using headings or bullet points. + - Reference the source using markdown links [descriptive text](url) — never bare URLs. + - IMAGES: The scraped content may contain image URLs in markdown format like `![alt text](image_url)`. + * When you find relevant/important images in the scraped content, include them in your response using standard markdown image syntax: `![alt text](image_url)`. + * This makes your response more visual and engaging. + * Prioritize showing: diagrams, charts, infographics, key illustrations, or images that help explain the content. + * Don't show every image - just the most relevant 1-3 images that enhance understanding. diff --git a/surfsense_backend/app/agents/new_chat/prompts/tools/search_surfsense_docs.md b/surfsense_backend/app/agents/new_chat/prompts/tools/search_surfsense_docs.md new file mode 100644 index 000000000..133717fec --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/tools/search_surfsense_docs.md @@ -0,0 +1,7 @@ + +- search_surfsense_docs: Search the official SurfSense documentation. + - Use this tool when the user asks anything about SurfSense itself (the application they are using). + - Args: + - query: The search query about SurfSense + - top_k: Number of documentation chunks to retrieve (default: 10) + - Returns: Documentation content with chunk IDs for citations (prefixed with 'doc-', e.g., [citation:doc-123]) diff --git a/surfsense_backend/app/agents/new_chat/prompts/tools/update_memory_private.md b/surfsense_backend/app/agents/new_chat/prompts/tools/update_memory_private.md new file mode 100644 index 000000000..184013804 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/tools/update_memory_private.md @@ -0,0 +1,31 @@ + +- update_memory: Update your personal memory document about the user. + - Your current memory is already in in your context. The `chars` and + `limit` attributes show your current usage and the maximum allowed size. + - This is your curated long-term memory — the distilled essence of what you know about + the user, not raw conversation logs. + - Call update_memory when: + * The user explicitly asks to remember or forget something + * The user shares durable facts or preferences that will matter in future conversations + - The user's first name is provided in . Use it in memory entries + instead of "the user" (e.g. "{name} works at..." not "The user works at..."). + Do not store the name itself as a separate memory entry. + - Do not store short-lived or ephemeral info: one-off questions, greetings, + session logistics, or things that only matter for the current task. + - Args: + - updated_memory: The FULL updated markdown document (not a diff). + Merge new facts with existing ones, update contradictions, remove outdated entries. + Treat every update as a curation pass — consolidate, don't just append. + - Every bullet MUST use this format: - (YYYY-MM-DD) [marker] text + Markers: + [fact] — durable facts (role, background, projects, tools, expertise) + [pref] — preferences (response style, languages, formats, tools) + [instr] — standing instructions (always/never do, response rules) + - Keep it concise and well under the character limit shown in . + - Every entry MUST be under a `##` heading. Keep heading names short (2-3 words) and + natural. Do NOT include the user's name in headings. Organize by context — e.g. + who they are, what they're focused on, how they prefer things. Create, split, or + merge headings freely as the memory grows. + - Each entry MUST be a single bullet point. Be descriptive but concise — include relevant + details and context rather than just a few words. + - During consolidation, prioritize keeping: [instr] > [pref] > [fact]. diff --git a/surfsense_backend/app/agents/new_chat/prompts/tools/update_memory_team.md b/surfsense_backend/app/agents/new_chat/prompts/tools/update_memory_team.md new file mode 100644 index 000000000..7eaca8818 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/tools/update_memory_team.md @@ -0,0 +1,26 @@ + +- update_memory: Update the team's shared memory document for this search space. + - Your current team memory is already in in your context. The `chars` + and `limit` attributes show current usage and the maximum allowed size. + - This is the team's curated long-term memory — decisions, conventions, key facts. + - NEVER store personal memory in team memory (e.g. personal bio, individual + preferences, or user-only standing instructions). + - Call update_memory when: + * A team member explicitly asks to remember or forget something + * The conversation surfaces durable team decisions, conventions, or facts + that will matter in future conversations + - Do not store short-lived or ephemeral info: one-off questions, greetings, + session logistics, or things that only matter for the current task. + - Args: + - updated_memory: The FULL updated markdown document (not a diff). + Merge new facts with existing ones, update contradictions, remove outdated entries. + Treat every update as a curation pass — consolidate, don't just append. + - Every bullet MUST use this format: - (YYYY-MM-DD) [fact] text + Team memory uses ONLY the [fact] marker. Never use [pref] or [instr] in team memory. + - Keep it concise and well under the character limit shown in . + - Every entry MUST be under a `##` heading. Keep heading names short (2-3 words) and + natural. Organize by context — e.g. what the team decided, current architecture, + active processes. Create, split, or merge headings freely as the memory grows. + - Each entry MUST be a single bullet point. Be descriptive but concise — include relevant + details and context rather than just a few words. + - During consolidation, prioritize keeping: decisions/conventions > key facts > current priorities. diff --git a/surfsense_backend/app/agents/new_chat/prompts/tools/web_search.md b/surfsense_backend/app/agents/new_chat/prompts/tools/web_search.md new file mode 100644 index 000000000..7ed7c332d --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/tools/web_search.md @@ -0,0 +1,18 @@ + +- web_search: Search the web for real-time information using all configured search engines. + - Use this for current events, news, prices, weather, public facts, or any question requiring + up-to-date information from the internet. + - This tool dispatches to all configured search engines (SearXNG, Tavily, Linkup, Baidu) in + parallel and merges the results. + - IMPORTANT (REAL-TIME / PUBLIC WEB QUERIES): For questions that require current public web data + (e.g., live exchange rates, stock prices, breaking news, weather, current events), you MUST call + `web_search` instead of answering from memory. + - For these real-time/public web queries, DO NOT answer from memory and DO NOT say you lack internet + access before attempting a web search. + - If the search returns no relevant results, explain that web sources did not return enough + data and ask the user if they want you to retry with a refined query. + - Args: + - query: The search query - use specific, descriptive terms + - top_k: Number of results to retrieve (default: 10, max: 50) + - If search snippets are insufficient for the user's question, use `scrape_webpage` on the most relevant result URL for full content. + - When presenting results, reference sources as markdown links [descriptive text](url) — never bare URLs. diff --git a/surfsense_backend/app/agents/new_chat/skills/__init__.py b/surfsense_backend/app/agents/new_chat/skills/__init__.py new file mode 100644 index 000000000..bb7ac055c --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/skills/__init__.py @@ -0,0 +1,7 @@ +"""SurfSense built-in agent skills (Anthropic Skills format). + +Each subdirectory corresponds to one skill and contains a ``SKILL.md`` file +with YAML frontmatter (name, description, allowed_tools) plus markdown +instructions. The :class:`BuiltinSkillsBackend` exposes them to the +deepagents :class:`SkillsMiddleware`. +""" diff --git a/surfsense_backend/app/agents/new_chat/skills/builtin/__init__.py b/surfsense_backend/app/agents/new_chat/skills/builtin/__init__.py new file mode 100644 index 000000000..8b1378917 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/skills/builtin/__init__.py @@ -0,0 +1 @@ + diff --git a/surfsense_backend/app/agents/new_chat/skills/builtin/email-drafting/SKILL.md b/surfsense_backend/app/agents/new_chat/skills/builtin/email-drafting/SKILL.md new file mode 100644 index 000000000..32e599e98 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/skills/builtin/email-drafting/SKILL.md @@ -0,0 +1,25 @@ +--- +name: email-drafting +description: Draft an email matching the user's voice, with structured intent and CTA +allowed-tools: search_surfsense_docs +--- + +# Email drafting + +## When to use this skill +"Draft an email to ...", "reply to this thread", "write a follow-up to X". Plain "summarize the email" is **not** in scope — that's a comprehension task. + +## Voice +Search the KB for prior emails from the user to similar audiences (same recipient, same topic class). Mirror tone, opening style, sign-off, and length distribution. If there is no precedent, default to: warm, direct, no filler, short paragraphs, one clear ask. + +## Required structure +Every draft includes, in this order: + +1. **Subject line** — concrete, ≤ 8 words, no clickbait, no `Re:` unless replying. +2. **Opening (1 sentence)** — context the recipient already shares; never restate what they wrote unless the thread is long. +3. **Body** — the actual point in one short paragraph. Bullets only if there are >3 discrete items. +4. **Single explicit CTA** — what you want the recipient to do, with a soft deadline if relevant. +5. **Sign-off** — match the user's prior closing style. + +## Always offer alternatives +End your message with: "Want me to make it shorter, more formal, or add a different angle?" — give the user one obvious next step. diff --git a/surfsense_backend/app/agents/new_chat/skills/builtin/kb-research/SKILL.md b/surfsense_backend/app/agents/new_chat/skills/builtin/kb-research/SKILL.md new file mode 100644 index 000000000..c268278ab --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/skills/builtin/kb-research/SKILL.md @@ -0,0 +1,23 @@ +--- +name: kb-research +description: Structured approach to finding and synthesizing information from the user's knowledge base +allowed-tools: search_surfsense_docs, scrape_webpage, read_file, ls_tree, grep, web_search +--- + +# Knowledge-base research + +## When to use this skill +- The user asks "find/look up/research" something specifically inside their knowledge base. +- The user references documents, notes, repos, or connector data they expect to exist already. +- A multi-document synthesis is required (e.g., "summarize what we've discussed about X across all my notes"). + +## Plan +1. Decompose the user's question into 2-4 specific, citation-worthy sub-questions. +2. For each sub-question, run **one** targeted KB search (focused on terms the user would have written, not synonyms). Open the most relevant 2-3 documents fully via `read_file` if their excerpts are too short. +3. Use `grep` to find supporting passages in long files instead of re-reading them end to end. +4. Cite every claim with `[citation:chunk_id]` exactly as the chunk tag specifies. + +## What good output looks like +- Short paragraphs with inline citations. +- Quoted phrases when wording matters. +- An explicit "Not found in your knowledge base" callout when a sub-question has no support — never fabricate. diff --git a/surfsense_backend/app/agents/new_chat/skills/builtin/meeting-prep/SKILL.md b/surfsense_backend/app/agents/new_chat/skills/builtin/meeting-prep/SKILL.md new file mode 100644 index 000000000..9657eb078 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/skills/builtin/meeting-prep/SKILL.md @@ -0,0 +1,22 @@ +--- +name: meeting-prep +description: Pull together briefing materials before a scheduled meeting +allowed-tools: search_surfsense_docs, web_search, scrape_webpage, read_file +--- + +# Meeting preparation + +## When to use this skill +The user mentions an upcoming meeting, call, or interview and asks you to "prep", "brief me", "pull background", or "what do I need to know about X before tomorrow". + +## Output structure +Always produce these sections (omit any with no signal — don't pad): + +1. **Attendees & context** — who's in the room, their roles, what they care about. Pull from KB notes about prior interactions; supplement with public profile facts via `web_search` when names or companies are unfamiliar. +2. **Open threads** — outstanding action items, unresolved decisions, last-mentioned blockers from prior conversation history. +3. **Recent moves** — within the last 30 days: relevant launches, hires, news. Cite KB chunks when present, otherwise external sources. +4. **Suggested questions** — 3-5 questions the user could ask, tailored to the open threads and the attendees' likely priorities. + +## Source ordering +- Always check the user's KB **first** for prior meeting notes, internal docs, or Slack threads about these attendees. +- Only fall back to `web_search` for *publicly verifiable* facts — never to fabricate a participant's preferences or relationships. diff --git a/surfsense_backend/app/agents/new_chat/skills/builtin/report-writing/SKILL.md b/surfsense_backend/app/agents/new_chat/skills/builtin/report-writing/SKILL.md new file mode 100644 index 000000000..17ac2f391 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/skills/builtin/report-writing/SKILL.md @@ -0,0 +1,23 @@ +--- +name: report-writing +description: How to scope, draft, and revise a Markdown report artifact via generate_report +allowed-tools: generate_report, search_surfsense_docs, read_file +--- + +# Report writing + +## When to use this skill +The user explicitly requests a deliverable: "write a report on …", "draft a memo", "produce a brief", "expand the previous report". A creation or modification verb pointed at an artifact is required (see `generate_report`'s when-to-call rules). + +## Decision flow +1. **Source strategy.** Decide which `source_strategy` fits: + - `conversation` — substantive Q&A on the topic already in chat. + - `kb_search` — fresh topic; supply 1–5 precise `search_queries`. + - `auto` — partial conversation context; let the tool fall back. + - `provided` — verbatim source text only. +2. **Style.** Default to `report_style="detailed"` unless the user explicitly asks for "brief", "one page", "500 words". +3. **Revisions.** When modifying an existing report from this conversation, set `parent_report_id` and put the change list in `user_instructions` ("add carbon-capture section", "tighten conclusion"). +4. **Never paste the report back into chat** after `generate_report` returns — confirm and let the artifact card render itself. + +## Hooks for KB-only mode +If `kb_search`/`auto` returns no results, do **not** silently switch to general knowledge. Surface the gap in your confirmation message. diff --git a/surfsense_backend/app/agents/new_chat/skills/builtin/slack-summary/SKILL.md b/surfsense_backend/app/agents/new_chat/skills/builtin/slack-summary/SKILL.md new file mode 100644 index 000000000..33b9e72a2 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/skills/builtin/slack-summary/SKILL.md @@ -0,0 +1,26 @@ +--- +name: slack-summary +description: Distill a Slack channel or thread into actionable summary +allowed-tools: search_surfsense_docs +--- + +# Slack summarization + +## When to use this skill +The user asks to summarize Slack ("what happened in #eng-platform this week", "what did Alice say about the launch", "catch me up on the design channel"). + +## Required inputs +Confirm before searching: +- **Which channel(s) or thread(s)?** Don't guess if ambiguous. +- **What time window?** Default to the last 7 days when not specified, but say so. + +## Output shape +Produce three concise sections: +1. **Key decisions** — explicit choices that were made, with the deciding message cited. +2. **Open questions** — things asked but not answered, with the asking message cited. +3. **Action items** — `@mention` who owes what by when, *only if explicitly stated*. Don't invent assignees. + +## What not to do +- Never produce a chronological play-by-play of every message — distill. +- Never quote private messages without flagging them as such. +- If the channel was empty in the time window, say so — don't fabricate filler. diff --git a/surfsense_backend/app/agents/new_chat/subagents/__init__.py b/surfsense_backend/app/agents/new_chat/subagents/__init__.py new file mode 100644 index 000000000..b9f21a0d2 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/subagents/__init__.py @@ -0,0 +1,26 @@ +"""Specialized user-facing subagents for the SurfSense agent. + +Each subagent is a :class:`deepagents.SubAgent` typed-dict spec passed to +:class:`deepagents.SubAgentMiddleware`, which materializes them as ephemeral +runnables invoked via the ``task`` tool. + +Per-subagent permission rules are injected as a +:class:`PermissionMiddleware` entry inside the subagent's ``middleware`` +field, mirroring opencode ``tool/task.ts`` which seeds child sessions with +deny rules for tools the parent does not want them touching (e.g. +``task``/``todowrite`` recursion, write tools for read-only research roles). +""" + +from .config import ( + build_connector_negotiator_subagent, + build_explore_subagent, + build_report_writer_subagent, + build_specialized_subagents, +) + +__all__ = [ + "build_connector_negotiator_subagent", + "build_explore_subagent", + "build_report_writer_subagent", + "build_specialized_subagents", +] diff --git a/surfsense_backend/app/agents/new_chat/subagents/config.py b/surfsense_backend/app/agents/new_chat/subagents/config.py new file mode 100644 index 000000000..e20bc06bf --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/subagents/config.py @@ -0,0 +1,427 @@ +"""Builders for specialized SurfSense subagents. + +Each subagent is built from three pieces: + +1. A name + description + system prompt (the user-facing contract for + when ``task`` should delegate to this role). +2. A filtered tool list (subset of the parent's bound tools). +3. A :class:`PermissionMiddleware` instance carrying a deny ruleset that + prevents the subagent from acting outside its scope (e.g. an + explore-only role cannot mutate state). + +Skill sources (``/skills/builtin/`` + ``/skills/space/``) are inherited +from the parent unconditionally — every subagent benefits from the same +authored guidance documents. +""" + +from __future__ import annotations + +import logging +from collections.abc import Iterable, Sequence +from typing import TYPE_CHECKING, Any + +from app.agents.new_chat.middleware.skills_backends import default_skills_sources +from app.agents.new_chat.permissions import Rule, Ruleset + +if TYPE_CHECKING: + from deepagents import SubAgent + from langchain_core.language_models import BaseChatModel + from langchain_core.tools import BaseTool + +logger = logging.getLogger(__name__) + + +# --------------------------------------------------------------------------- +# Tool name constants +# --------------------------------------------------------------------------- + +# Read-only tools that ``explore`` is permitted to use. Names match the +# tools provided by the deepagents ``FilesystemMiddleware`` (``ls``, ``read_file``, +# ``glob``, ``grep``) plus the SurfSense-side read tools. +EXPLORE_READ_TOOLS: frozenset[str] = frozenset( + { + "search_surfsense_docs", + "web_search", + "scrape_webpage", + "read_file", + "ls", + "glob", + "grep", + } +) + +# Tools ``report_writer`` may call. The set is intentionally narrow so the +# subagent doesn't drift into tangential research; if richer source-gathering +# is needed, the parent should hand off to ``explore`` first. +REPORT_WRITER_TOOLS: frozenset[str] = frozenset( + { + "search_surfsense_docs", + "read_file", + "generate_report", + } +) + +# Wildcard patterns that match write tools we deny by default in read-only +# subagents. Anchored at start AND end via :func:`Rule` semantics. We use +# substring-style ``*verb*`` patterns because connector tool names typically +# put the verb in the middle (``linear_create_issue``, ``slack_send_message``, +# ``notion_update_page``); strict suffix patterns (``*_create``) miss those. +# +# A handful of canonical exact-match names is appended so that bare verbs +# (``edit``, ``write``) are also blocked even when a connector dropped the +# usual prefix. +WRITE_TOOL_DENY_PATTERNS: tuple[str, ...] = ( + "*create*", + "*update*", + "*delete*", + "*send*", + "*write*", + "*edit*", + "*move*", + "*mkdir*", + "*upload*", + "edit_file", + "write_file", + "move_file", + "mkdir", + "update_memory", + "update_memory_team", + "update_memory_private", +) + + +# --------------------------------------------------------------------------- +# Helpers +# --------------------------------------------------------------------------- + + +# Tool names that are NOT in the registry's ``tools`` list because they +# are provided dynamically by middleware at compile time. We don't pass +# them through ``_filter_tools`` (the actual ``BaseTool`` instances live +# inside the middleware), but we do exempt them from the "missing" warning +# below — operators were seeing spurious noise like +# ``missing: ['glob', 'grep', 'ls', 'read_file']`` even though those +# tools are reachable via :class:`SurfSenseFilesystemMiddleware` once the +# subagent is compiled. +_MIDDLEWARE_PROVIDED_TOOL_NAMES: frozenset[str] = frozenset( + { + "ls", + "read_file", + "write_file", + "edit_file", + "glob", + "grep", + "execute", + "write_todos", + "task", + } +) + + +def _filter_tools( + tools: Sequence[BaseTool], + allowed_names: Iterable[str], +) -> list[BaseTool]: + """Return only tools whose ``name`` appears in ``allowed_names``. + + Tools are looked up by exact name. Names matching + :data:`_MIDDLEWARE_PROVIDED_TOOL_NAMES` are intentionally absent from + ``tools`` (they're injected by middleware at compile time) and are + silently excluded from the "missing" warning so operators don't see + false positives every build. + """ + allowed = set(allowed_names) + selected = [t for t in tools if t.name in allowed] + missing = sorted( + (allowed - {t.name for t in selected}) - _MIDDLEWARE_PROVIDED_TOOL_NAMES + ) + if missing: + logger.info( + "Subagent build: %d/%d registry tools available; missing: %s", + len(selected), + len(allowed - _MIDDLEWARE_PROVIDED_TOOL_NAMES), + missing, + ) + return selected + + +def _read_only_deny_rules() -> list[Rule]: + """Synthesize a list of deny rules covering common write-tool patterns.""" + return [ + Rule(permission=pattern, pattern="*", action="deny") + for pattern in WRITE_TOOL_DENY_PATTERNS + ] + + +def _build_permission_middleware(deny_rules: list[Rule], origin: str): + """Construct a :class:`PermissionMiddleware` seeded with ``deny_rules``. + + Imported lazily because the middleware module pulls in interrupt/HITL + machinery we don't want at import time of this config file. + """ + from app.agents.new_chat.middleware.permission import PermissionMiddleware + + return PermissionMiddleware( + rulesets=[Ruleset(rules=deny_rules, origin=origin)], + ) + + +def _wrap_with_subagent_essentials( + custom_middleware: list, + *, + agent_tools: Sequence[BaseTool], + extra_middleware: Sequence[Any] | None = None, +): + """Compose the final middleware list for a specialized subagent. + + Order, outer to inner: + + 1. ``extra_middleware`` — provided by the caller (typically the parent + agent's ``SurfSenseFilesystemMiddleware`` and ``TodoListMiddleware``) + so the subagent inherits the parent's filesystem/todo view. These + run **before** the subagent-local middleware so their tools are + wired up before permissioning kicks in. + 2. ``custom_middleware`` — subagent-local rules (e.g. permission deny + lists). + 3. :class:`PatchToolCallsMiddleware` — normalizes tool-call shapes. + 4. :class:`DedupHITLToolCallsMiddleware` — collapses duplicate HITL + calls using metadata declared at registry time. + + Without ``extra_middleware`` the subagent will only have the registry + tools listed in its ``tools`` field — meaning ``read_file``, ``ls``, + ``grep``, etc. won't exist. Always pass ``extra_middleware`` from the + parent unless you specifically want a sandboxed subagent. + """ + from deepagents.middleware.patch_tool_calls import PatchToolCallsMiddleware + + from app.agents.new_chat.middleware import DedupHITLToolCallsMiddleware + + return [ + *(extra_middleware or []), + *custom_middleware, + PatchToolCallsMiddleware(), + DedupHITLToolCallsMiddleware(agent_tools=list(agent_tools)), + ] + + +# --------------------------------------------------------------------------- +# System prompts +# --------------------------------------------------------------------------- + +EXPLORE_SYSTEM_PROMPT = """You are the **explore** subagent for SurfSense. + +## Your job +Conduct read-only research across the user's knowledge base, the web, and any documents the parent agent has surfaced. Return a synthesized answer with explicit citations — never speculate beyond the sources you have actually inspected. + +## Tools available +- `search_surfsense_docs` — fast hybrid search over the user's knowledge base. +- `web_search` — only when the user's KB clearly does not contain the answer. +- `scrape_webpage` — to read a URL the user or the search results provided. +- `read_file`, `ls`, `glob`, `grep` — to inspect specific documents or trees the parent has flagged. + +## Rules +- Read-only. You cannot create, edit, delete, send, or move anything. +- Cite every claim. Use `[citation:chunk_id]` exactly as the chunk tag specifies. +- If a sub-question has no support in the inspected sources, say so explicitly. Do not fabricate. +- Return the most useful synthesis in your single final message. The parent agent will not be able to follow up. +""" + + +REPORT_WRITER_SYSTEM_PROMPT = """You are the **report_writer** subagent for SurfSense. + +## Your job +Produce a single high-quality report deliverable using `generate_report`. The parent has already gathered (or knows where to gather) the underlying sources. + +## Workflow +1. **Outline first.** Before calling `generate_report`, write a one-paragraph outline of the sections you plan to produce. Confirm the outline reflects the parent's instructions. +2. **Source resolution.** Decide whether to call `search_surfsense_docs` and `read_file` for any final-checks, or whether the parent's earlier tool calls already cover the source set. +3. **One report.** Call `generate_report` exactly once with `source_strategy` chosen per the topic and chat history (see the `report-writing` skill). +4. **Confirm.** End with a one-sentence summary in your final message — never paste the report back into chat; the artifact card renders itself. +""" + + +CONNECTOR_NEGOTIATOR_SYSTEM_PROMPT = """You are the **connector_negotiator** subagent for SurfSense. + +## Your job +Coordinate cross-connector workflows: chains where the result of one service's tool feeds into another's. Common shapes include "find Linear issues mentioned in last week's Slack messages", "draft a Gmail reply citing a Notion doc", or "list Linear tickets opened by the same person who filed Jira FOO-123". + +## Workflow +1. **Plan.** Identify the connector hops needed and the order they should run in. Write a short plan in your first message. +2. **Verify access.** Use `get_connected_accounts` to confirm the relevant connectors are actually wired up before issuing tool calls. If a connector is missing, stop and report — do not fabricate. +3. **Execute.** Run each hop, citing IDs (issue keys, message ts, page IDs) in your scratch notes so the parent can audit. +4. **Hand back.** Return a structured summary with the final answer plus the chain of evidence (issue → message → page, etc.). + +## Caveats +- If a hop fails, do not retry blindly — return the partial result and explain. +- Mutating tools (create, update, delete, send) require parent permission; you are NOT cleared to call them on your own. +""" + + +# --------------------------------------------------------------------------- +# Subagent builders +# --------------------------------------------------------------------------- + + +def build_explore_subagent( + *, + tools: Sequence[BaseTool], + model: BaseChatModel | None = None, + extra_middleware: Sequence[Any] | None = None, +) -> SubAgent: + """Build the read-only ``explore`` subagent spec. + + Pass ``extra_middleware`` (typically the parent's filesystem + todo + middleware) so the subagent can actually use ``read_file``, ``ls``, + ``grep``, ``glob`` — which its system prompt promises but which only + exist when their middleware is mounted. + """ + from deepagents import SubAgent # noqa: F401 (TypedDict for type clarity) + + selected_tools = _filter_tools(tools, EXPLORE_READ_TOOLS) + deny_rules = _read_only_deny_rules() + permission_mw = _build_permission_middleware( + deny_rules, origin="subagent_explore" + ) + + spec: dict = { + "name": "explore", + "description": ( + "Read-only research across the user's knowledge base and the web. " + "Use when the parent needs deeply-cited synthesis without " + "modifying anything." + ), + "system_prompt": EXPLORE_SYSTEM_PROMPT, + "tools": selected_tools, + "middleware": _wrap_with_subagent_essentials( + [permission_mw], + agent_tools=selected_tools, + extra_middleware=extra_middleware, + ), + "skills": default_skills_sources(), + } + if model is not None: + spec["model"] = model + return spec # type: ignore[return-value] + + +def build_report_writer_subagent( + *, + tools: Sequence[BaseTool], + model: BaseChatModel | None = None, + extra_middleware: Sequence[Any] | None = None, +) -> SubAgent: + """Build the ``report_writer`` subagent spec. + + Read-only deny ruleset still applies — the subagent should call + ``generate_report`` and nothing else mutating. ``generate_report`` + creates a report artifact via a backend service and is intentionally + **not** denied. + + Pass ``extra_middleware`` (typically the parent's filesystem + todo + middleware) so the subagent can run ``read_file`` for source-checks + before calling ``generate_report``. + """ + selected_tools = _filter_tools(tools, REPORT_WRITER_TOOLS) + deny_rules = _read_only_deny_rules() + permission_mw = _build_permission_middleware( + deny_rules, origin="subagent_report_writer" + ) + + spec: dict = { + "name": "report_writer", + "description": ( + "Produce a single Markdown report artifact via generate_report, " + "using the outline-then-fill protocol. Use when the parent has " + "decided a deliverable is needed." + ), + "system_prompt": REPORT_WRITER_SYSTEM_PROMPT, + "tools": selected_tools, + "middleware": _wrap_with_subagent_essentials( + [permission_mw], + agent_tools=selected_tools, + extra_middleware=extra_middleware, + ), + "skills": default_skills_sources(), + } + if model is not None: + spec["model"] = model + return spec # type: ignore[return-value] + + +def build_connector_negotiator_subagent( + *, + tools: Sequence[BaseTool], + model: BaseChatModel | None = None, + extra_middleware: Sequence[Any] | None = None, +) -> SubAgent: + """Build the ``connector_negotiator`` subagent spec. + + Inherits all MCP / connector tools the parent has plus + ``get_connected_accounts``. Read-only by default; permission rules deny + write/mutation patterns. The parent agent re-asks for permission if a + connector mutation is genuinely needed. + + Pass ``extra_middleware`` (typically the parent's filesystem + todo + middleware) so this subagent shares the parent's filesystem view when + citing evidence across hops. + """ + parent_tool_names = {t.name for t in tools} + allowed: set[str] = set() + if "get_connected_accounts" in parent_tool_names: + allowed.add("get_connected_accounts") + # Inherit anything that smells connector- or MCP-related but is not a + # bulk-write API. Heuristic: keep all parent tools; rely on the deny + # ruleset to block mutation patterns. This mirrors the plan: "all + # MCP/connector tools the parent has". + for name in parent_tool_names: + allowed.add(name) + selected_tools = _filter_tools(tools, allowed) + + deny_rules = _read_only_deny_rules() + permission_mw = _build_permission_middleware( + deny_rules, origin="subagent_connector_negotiator" + ) + + spec: dict = { + "name": "connector_negotiator", + "description": ( + "Coordinate read-only chains across connectors (Slack → Linear, " + "Notion → Gmail, etc.). Returns a structured summary with the " + "evidence chain. Cannot mutate connector state." + ), + "system_prompt": CONNECTOR_NEGOTIATOR_SYSTEM_PROMPT, + "tools": selected_tools, + "middleware": _wrap_with_subagent_essentials( + [permission_mw], + agent_tools=selected_tools, + extra_middleware=extra_middleware, + ), + "skills": default_skills_sources(), + } + if model is not None: + spec["model"] = model + return spec # type: ignore[return-value] + + +def build_specialized_subagents( + *, + tools: Sequence[BaseTool], + model: BaseChatModel | None = None, + extra_middleware: Sequence[Any] | None = None, +) -> list[SubAgent]: + """Return the canonical list of specialized subagents to register. + + Order matters only for the order they appear in the ``task`` tool + description — most useful first. + """ + return [ + build_explore_subagent( + tools=tools, model=model, extra_middleware=extra_middleware + ), + build_report_writer_subagent( + tools=tools, model=model, extra_middleware=extra_middleware + ), + build_connector_negotiator_subagent( + tools=tools, model=model, extra_middleware=extra_middleware + ), + ] diff --git a/surfsense_backend/app/agents/new_chat/system_prompt.py b/surfsense_backend/app/agents/new_chat/system_prompt.py index 0c9426892..3919527d9 100644 --- a/surfsense_backend/app/agents/new_chat/system_prompt.py +++ b/surfsense_backend/app/agents/new_chat/system_prompt.py @@ -1,842 +1,43 @@ """ -System prompt building for SurfSense agents. +Thin compatibility wrapper around :mod:`app.agents.new_chat.prompts.composer`. -This module provides functions and constants for building the SurfSense system prompt -with configurable user instructions and citation support. +Tier 3a of the OpenCode-port plan replaced the monolithic prompt strings +in this module with a fragment tree under ``prompts/`` and a composer +function. This module preserves the public function surface +(``build_surfsense_system_prompt`` / ``build_configurable_system_prompt`` / +``get_default_system_instructions`` / ``SURFSENSE_SYSTEM_PROMPT``) so that +existing call sites — `chat_deepagent.py`, anonymous chat routes, and the +configurable-prompt admin path — keep working without churn. -The prompt is composed of three parts: -1. System Instructions (configurable via NewLLMConfig) -2. Tools Instructions (always included, not configurable) -3. Citation Instructions (toggleable via NewLLMConfig.citations_enabled) +For new call sites prefer importing ``compose_system_prompt`` directly +from :mod:`app.agents.new_chat.prompts.composer`. """ +from __future__ import annotations + from datetime import UTC, datetime from app.db import ChatVisibility -# Default system instructions - can be overridden via NewLLMConfig.system_instructions -SURFSENSE_SYSTEM_INSTRUCTIONS = """ - -You are SurfSense, a reasoning and acting AI agent designed to answer user questions using the user's personal knowledge base. - -Today's date (UTC): {resolved_today} - -When writing mathematical formulas or equations, ALWAYS use LaTeX notation. NEVER use backtick code spans or Unicode symbols for math. - -NEVER expose internal tool parameter names, backend IDs, or implementation details to the user. Always use natural, user-friendly language instead. - - -CRITICAL RULE — KNOWLEDGE BASE FIRST, NEVER DEFAULT TO GENERAL KNOWLEDGE: -- You MUST answer questions ONLY using information retrieved from the user's knowledge base, web search results, scraped webpages, or other tool outputs. -- You MUST NOT answer factual or informational questions from your own training data or general knowledge unless the user explicitly grants permission. -- If the knowledge base search returns no relevant results AND no other tool provides the answer, you MUST: - 1. Inform the user that you could not find relevant information in their knowledge base. - 2. Ask the user: "Would you like me to answer from my general knowledge instead?" - 3. ONLY provide a general-knowledge answer AFTER the user explicitly says yes. -- This policy does NOT apply to: - * Casual conversation, greetings, or meta-questions about SurfSense itself (e.g., "what can you do?") - * Formatting, summarization, or analysis of content already present in the conversation - * Following user instructions that are clearly task-oriented (e.g., "rewrite this in bullet points") - * Tool-usage actions like generating reports, podcasts, images, or scraping webpages - * Queries about services that have direct tools (Linear, ClickUp, Jira, Slack, Airtable) — see below - - - -CRITICAL — You have direct tools for these services: Linear, ClickUp, Jira, Slack, Airtable. -Their data is NEVER in the knowledge base. You MUST call their tools immediately — never -say "I don't see it in the knowledge base" or ask the user if they want you to check. -Ignore any knowledge base results for these services. - -When to use which tool: -- Linear (issues) → list_issues, get_issue, save_issue (create/update) -- ClickUp (tasks) → clickup_search, clickup_get_task -- Jira (issues) → getAccessibleAtlassianResources (cloudId discovery), getVisibleJiraProjects (project discovery), getJiraProjectIssueTypesMetadata (issue type discovery), searchJiraIssuesUsingJql, createJiraIssue, editJiraIssue -- Slack (messages, channels) → slack_search_channels, slack_read_channel, slack_read_thread -- Airtable (bases, tables, records) → list_bases, list_tables_for_base, list_records_for_table -- Knowledge base content (Notion, GitHub, files, notes) → automatically searched -- Real-time public web data → call web_search -- Reading a specific webpage → call scrape_webpage - - - -Some service tools require identifiers or context you do not have (account IDs, -workspace names, channel IDs, project keys, etc.). NEVER ask the user for raw -IDs or technical identifiers — they cannot memorise them. - -Instead, follow this discovery pattern: -1. Call a listing/discovery tool to find available options. -2. ONE result → use it silently, no question to the user. -3. MULTIPLE results → present the options by their display names and let the - user choose. Never show raw UUIDs — always use friendly names. - -Discovery tools by level: -- Which account/workspace? → get_connected_accounts("") -- Which Jira site (cloudId)? → getAccessibleAtlassianResources -- Which Jira project? → getVisibleJiraProjects (after resolving cloudId) -- Which Jira issue type? → getJiraProjectIssueTypesMetadata (after resolving project) -- Which channel? → slack_search_channels -- Which base? → list_bases -- Which table? → list_tables_for_base (after resolving baseId) -- Which task? → clickup_search -- Which issue? → list_issues (Linear) or searchJiraIssuesUsingJql (Jira) - -For Jira specifically: ALWAYS call getAccessibleAtlassianResources first to -obtain the cloudId, then pass it to other Jira tools. When creating an issue, -chain: getAccessibleAtlassianResources → getVisibleJiraProjects → createJiraIssue. -If there is only one option at each step, use it silently. If multiple, present -friendly names. - -Chain discovery when needed — e.g. for Airtable records: list_bases → pick -base → list_tables_for_base → pick table → list_records_for_table. - -MULTI-ACCOUNT TOOL NAMING: When the user has multiple accounts connected for -the same service, tool names are prefixed to avoid collisions — e.g. -linear_25_list_issues and linear_30_list_issues instead of two list_issues. -Each prefixed tool's description starts with [Account: ] so you -know which account it targets. Use get_connected_accounts("") to see -the full list of accounts with their connector IDs and display names. -When only one account is connected, tools have their normal unprefixed names. - - - -IMPORTANT — After understanding each user message, ALWAYS check: does this message -reveal durable facts about the user (role, interests, preferences, projects, -background, or standing instructions)? If yes, you MUST call update_memory -alongside your normal response — do not defer this to a later turn. - - - -""" - -# Default system instructions for shared (team) threads: team context + message format for attribution -_SYSTEM_INSTRUCTIONS_SHARED = """ - -You are SurfSense, a reasoning and acting AI agent designed to answer questions in this team space using the team's shared knowledge base. - -In this team thread, each message is prefixed with **[DisplayName of the author]**. Use this to attribute and reference the author of anything in the discussion (who asked a question, made a suggestion, or contributed an idea) and to cite who said what in your answers. - -Today's date (UTC): {resolved_today} - -When writing mathematical formulas or equations, ALWAYS use LaTeX notation. NEVER use backtick code spans or Unicode symbols for math. - -NEVER expose internal tool parameter names, backend IDs, or implementation details to the user. Always use natural, user-friendly language instead. - - -CRITICAL RULE — KNOWLEDGE BASE FIRST, NEVER DEFAULT TO GENERAL KNOWLEDGE: -- You MUST answer questions ONLY using information retrieved from the team's shared knowledge base, web search results, scraped webpages, or other tool outputs. -- You MUST NOT answer factual or informational questions from your own training data or general knowledge unless a team member explicitly grants permission. -- If the knowledge base search returns no relevant results AND no other tool provides the answer, you MUST: - 1. Inform the team that you could not find relevant information in the shared knowledge base. - 2. Ask: "Would you like me to answer from my general knowledge instead?" - 3. ONLY provide a general-knowledge answer AFTER a team member explicitly says yes. -- This policy does NOT apply to: - * Casual conversation, greetings, or meta-questions about SurfSense itself (e.g., "what can you do?") - * Formatting, summarization, or analysis of content already present in the conversation - * Following user instructions that are clearly task-oriented (e.g., "rewrite this in bullet points") - * Tool-usage actions like generating reports, podcasts, images, or scraping webpages - * Queries about services that have direct tools (Linear, ClickUp, Jira, Slack, Airtable) — see below - - - -CRITICAL — You have direct tools for these services: Linear, ClickUp, Jira, Slack, Airtable. -Their data is NEVER in the knowledge base. You MUST call their tools immediately — never -say "I don't see it in the knowledge base" or ask if they want you to check. -Ignore any knowledge base results for these services. - -When to use which tool: -- Linear (issues) → list_issues, get_issue, save_issue (create/update) -- ClickUp (tasks) → clickup_search, clickup_get_task -- Jira (issues) → getAccessibleAtlassianResources (cloudId discovery), getVisibleJiraProjects (project discovery), getJiraProjectIssueTypesMetadata (issue type discovery), searchJiraIssuesUsingJql, createJiraIssue, editJiraIssue -- Slack (messages, channels) → slack_search_channels, slack_read_channel, slack_read_thread -- Airtable (bases, tables, records) → list_bases, list_tables_for_base, list_records_for_table -- Knowledge base content (Notion, GitHub, files, notes) → automatically searched -- Real-time public web data → call web_search -- Reading a specific webpage → call scrape_webpage - - - -Some service tools require identifiers or context you do not have (account IDs, -workspace names, channel IDs, project keys, etc.). NEVER ask the user for raw -IDs or technical identifiers — they cannot memorise them. - -Instead, follow this discovery pattern: -1. Call a listing/discovery tool to find available options. -2. ONE result → use it silently, no question to the user. -3. MULTIPLE results → present the options by their display names and let the - user choose. Never show raw UUIDs — always use friendly names. - -Discovery tools by level: -- Which account/workspace? → get_connected_accounts("") -- Which Jira site (cloudId)? → getAccessibleAtlassianResources -- Which Jira project? → getVisibleJiraProjects (after resolving cloudId) -- Which Jira issue type? → getJiraProjectIssueTypesMetadata (after resolving project) -- Which channel? → slack_search_channels -- Which base? → list_bases -- Which table? → list_tables_for_base (after resolving baseId) -- Which task? → clickup_search -- Which issue? → list_issues (Linear) or searchJiraIssuesUsingJql (Jira) - -For Jira specifically: ALWAYS call getAccessibleAtlassianResources first to -obtain the cloudId, then pass it to other Jira tools. When creating an issue, -chain: getAccessibleAtlassianResources → getVisibleJiraProjects → createJiraIssue. -If there is only one option at each step, use it silently. If multiple, present -friendly names. - -Chain discovery when needed — e.g. for Airtable records: list_bases → pick -base → list_tables_for_base → pick table → list_records_for_table. - -MULTI-ACCOUNT TOOL NAMING: When the user has multiple accounts connected for -the same service, tool names are prefixed to avoid collisions — e.g. -linear_25_list_issues and linear_30_list_issues instead of two list_issues. -Each prefixed tool's description starts with [Account: ] so you -know which account it targets. Use get_connected_accounts("") to see -the full list of accounts with their connector IDs and display names. -When only one account is connected, tools have their normal unprefixed names. - - - -IMPORTANT — After understanding each user message, ALWAYS check: does this message -reveal durable facts about the team (decisions, conventions, architecture, processes, -or key facts)? If yes, you MUST call update_memory alongside your normal response — -do not defer this to a later turn. - - - -""" - - -def _get_system_instructions( - thread_visibility: ChatVisibility | None = None, today: datetime | None = None -) -> str: - """Build system instructions based on thread visibility (private vs shared).""" - - resolved_today = (today or datetime.now(UTC)).astimezone(UTC).date().isoformat() - visibility = thread_visibility or ChatVisibility.PRIVATE - if visibility == ChatVisibility.SEARCH_SPACE: - return _SYSTEM_INSTRUCTIONS_SHARED.format(resolved_today=resolved_today) - else: - return SURFSENSE_SYSTEM_INSTRUCTIONS.format(resolved_today=resolved_today) - - -# ============================================================================= -# Per-tool prompt instructions keyed by registry tool name. -# Only tools present in the enabled set will be included in the system prompt. -# ============================================================================= - -_TOOLS_PREAMBLE = """ - -You have access to the following tools: - -IMPORTANT: You can ONLY use the tools listed below. If a capability is not listed here, you do NOT have it. -Do NOT claim you can do something if the corresponding tool is not listed. - -""" - -_TOOL_INSTRUCTIONS: dict[str, str] = {} - -_TOOL_INSTRUCTIONS["search_surfsense_docs"] = """ -- search_surfsense_docs: Search the official SurfSense documentation. - - Use this tool when the user asks anything about SurfSense itself (the application they are using). - - Args: - - query: The search query about SurfSense - - top_k: Number of documentation chunks to retrieve (default: 10) - - Returns: Documentation content with chunk IDs for citations (prefixed with 'doc-', e.g., [citation:doc-123]) -""" - -_TOOL_INSTRUCTIONS["generate_podcast"] = """ -- generate_podcast: Generate an audio podcast from provided content. - - Use this when the user asks to create, generate, or make a podcast. - - Trigger phrases: "give me a podcast about", "create a podcast", "generate a podcast", "make a podcast", "turn this into a podcast" - - Args: - - source_content: The text content to convert into a podcast. This MUST be comprehensive and include: - * If discussing the current conversation: Include a detailed summary of the FULL chat history (all user questions and your responses) - * If based on knowledge base search: Include the key findings and insights from the search results - * You can combine both: conversation context + search results for richer podcasts - * The more detailed the source_content, the better the podcast quality - - podcast_title: Optional title for the podcast (default: "SurfSense Podcast") - - user_prompt: Optional instructions for podcast style/format (e.g., "Make it casual and fun") - - Returns: A task_id for tracking. The podcast will be generated in the background. - - IMPORTANT: Only one podcast can be generated at a time. If a podcast is already being generated, the tool will return status "already_generating". - - After calling this tool, inform the user that podcast generation has started and they will see the player when it's ready (takes 3-5 minutes). -""" - -_TOOL_INSTRUCTIONS["generate_video_presentation"] = """ -- generate_video_presentation: Generate a video presentation from provided content. - - Use this when the user asks to create a video, presentation, slides, or slide deck. - - Trigger phrases: "give me a presentation", "create slides", "generate a video", "make a slide deck", "turn this into a presentation" - - Args: - - source_content: The text content to turn into a presentation. The more detailed, the better. - - video_title: Optional title (default: "SurfSense Presentation") - - user_prompt: Optional style instructions (e.g., "Make it technical and detailed") - - After calling this tool, inform the user that generation has started and they will see the presentation when it's ready. -""" - -_TOOL_INSTRUCTIONS["generate_report"] = """ -- generate_report: Generate or revise a structured Markdown report artifact. - - WHEN TO CALL THIS TOOL — the message must contain a creation or modification VERB directed at producing a deliverable: - * Creation verbs: write, create, generate, draft, produce, summarize into, turn into, make - * Modification verbs: revise, update, expand, add (a section), rewrite, make (it shorter/longer/formal) - * Example triggers: "generate a report about...", "write a document on...", "add a section about budget", "make the report shorter", "rewrite in formal tone" - - WHEN NOT TO CALL THIS TOOL (answer in chat instead): - * Questions or discussion about the report: "What can we add?", "What's missing?", "Is the data accurate?", "How could this be improved?" - * Suggestions or brainstorming: "What other topics could be covered?", "What else could be added?", "What would make this better?" - * Asking for explanations: "Can you explain section 2?", "Why did you include that?", "What does this part mean?" - * Quick follow-ups or critiques: "Is the conclusion strong enough?", "Are there any gaps?", "What about the competitors?" - * THE TEST: Does the message contain a creation/modification VERB (from the list above) directed at producing or changing a deliverable? If NO verb → answer conversationally in chat. Do NOT assume the user wants a revision just because a report exists in the conversation. - - IMPORTANT FORMAT RULE: Reports are ALWAYS generated in Markdown. - - Args: - - topic: Short title for the report (max ~8 words). - - source_content: The text content to base the report on. - * For source_strategy="conversation" or "provided": Include a comprehensive summary of the relevant content. - * For source_strategy="kb_search": Can be empty or minimal — the tool handles searching internally. - * For source_strategy="auto": Include what you have; the tool searches KB if it's not enough. - - source_strategy: Controls how the tool collects source material. One of: - * "conversation" — The conversation already contains enough context (prior Q&A, discussion, pasted text, scraped pages). Pass a thorough summary as source_content. - * "kb_search" — The tool will search the knowledge base internally. Provide search_queries with 1-5 targeted queries. - * "auto" — Use source_content if sufficient, otherwise fall back to internal KB search using search_queries. - * "provided" — Use only what is in source_content (default, backward-compatible). - - search_queries: When source_strategy is "kb_search" or "auto", provide 1-5 specific search queries for the knowledge base. These should be precise, not just the topic name repeated. - - report_style: Controls report depth. Options: "detailed" (DEFAULT), "deep_research", "brief". - Use "brief" ONLY when the user explicitly asks for a short/concise/one-page report (e.g., "one page", "keep it short", "brief report", "500 words"). Default to "detailed" for all other requests. - - user_instructions: Optional specific instructions (e.g., "focus on financial impacts", "include recommendations"). When revising (parent_report_id set), describe WHAT TO CHANGE. If the user mentions a length preference (e.g., "one page", "500 words", "2 pages"), include that VERBATIM here AND set report_style="brief". - - parent_report_id: Set this to the report_id from a previous generate_report result when the user wants to MODIFY an existing report. Do NOT set it for new reports or questions about reports. - - Returns: A dictionary with status "ready" or "failed", report_id, title, and word_count. - - The report is generated immediately in Markdown and displayed inline in the chat. - - Export/download formats (PDF, DOCX, HTML, LaTeX, EPUB, ODT, plain text) are produced from the generated Markdown report. - - SOURCE STRATEGY DECISION (HIGH PRIORITY — follow this exactly): - * If the conversation already has substantive Q&A / discussion on the topic → use source_strategy="conversation" with a comprehensive summary as source_content. - * If the user wants a report on a topic not yet discussed → use source_strategy="kb_search" with targeted search_queries. - * If you have some content but might need more → use source_strategy="auto" with both source_content and search_queries. - * When revising an existing report (parent_report_id set) and the conversation has relevant context → use source_strategy="conversation". The revision will use the previous report content plus your source_content. - * NEVER run a separate KB lookup step and then pass those results to generate_report. The tool handles KB search internally. - - AFTER CALLING THIS TOOL: Do NOT repeat, summarize, or reproduce the report content in the chat. The report is already displayed as an interactive card that the user can open, read, copy, and export. Simply confirm that the report was generated (e.g., "I've generated your report on [topic]. You can view the Markdown report now, and export it in various formats from the card."). NEVER write out the report text in the chat. -""" - -_TOOL_INSTRUCTIONS["generate_image"] = """ -- generate_image: Generate images from text descriptions using AI image models. - - Use this when the user asks you to create, generate, draw, design, or make an image. - - Trigger phrases: "generate an image of", "create a picture of", "draw me", "make an image", "design a logo", "create artwork" - - Args: - - prompt: A detailed text description of the image to generate. Be specific about subject, style, colors, composition, and mood. - - n: Number of images to generate (1-4, default: 1) - - Returns: A dictionary with the generated image metadata. The image will automatically be displayed in the chat. - - IMPORTANT: Write a detailed, descriptive prompt for best results. Don't just pass the user's words verbatim - - expand and improve the prompt with specific details about style, lighting, composition, and mood. - - If the user's request is vague (e.g., "make me an image of a cat"), enhance the prompt with artistic details. -""" - -_TOOL_INSTRUCTIONS["scrape_webpage"] = """ -- scrape_webpage: Scrape and extract the main content from a webpage. - - Use this when the user wants you to READ and UNDERSTAND the actual content of a webpage. - - CRITICAL — WHEN TO USE (always attempt scraping, never refuse before trying): - * When a user asks to "get", "fetch", "pull", "grab", "scrape", or "read" content from a URL - * When the user wants live/dynamic data from a specific webpage (e.g., tables, scores, stats, prices) - * When a URL was mentioned earlier in the conversation and the user asks for its actual content - * When `/documents/` knowledge-base data is insufficient and the user wants more - - Trigger scenarios: - * "Read this article and summarize it" - * "What does this page say about X?" - * "Summarize this blog post for me" - * "Tell me the key points from this article" - * "What's in this webpage?" - * "Can you analyze this article?" - * "Can you get the live table/data from [URL]?" - * "Scrape it" / "Can you scrape that?" (referring to a previously mentioned URL) - * "Fetch the content from [URL]" - * "Pull the data from that page" - - Args: - - url: The URL of the webpage to scrape (must be HTTP/HTTPS) - - max_length: Maximum content length to return (default: 50000 chars) - - Returns: The page title, description, full content (in markdown), word count, and metadata - - After scraping, provide a comprehensive, well-structured summary with key takeaways using headings or bullet points. - - Reference the source using markdown links [descriptive text](url) — never bare URLs. - - IMAGES: The scraped content may contain image URLs in markdown format like `![alt text](image_url)`. - * When you find relevant/important images in the scraped content, include them in your response using standard markdown image syntax: `![alt text](image_url)`. - * This makes your response more visual and engaging. - * Prioritize showing: diagrams, charts, infographics, key illustrations, or images that help explain the content. - * Don't show every image - just the most relevant 1-3 images that enhance understanding. -""" - -_TOOL_INSTRUCTIONS["web_search"] = """ -- web_search: Search the web for real-time information using all configured search engines. - - Use this for current events, news, prices, weather, public facts, or any question requiring - up-to-date information from the internet. - - This tool dispatches to all configured search engines (SearXNG, Tavily, Linkup, Baidu) in - parallel and merges the results. - - IMPORTANT (REAL-TIME / PUBLIC WEB QUERIES): For questions that require current public web data - (e.g., live exchange rates, stock prices, breaking news, weather, current events), you MUST call - `web_search` instead of answering from memory. - - For these real-time/public web queries, DO NOT answer from memory and DO NOT say you lack internet - access before attempting a web search. - - If the search returns no relevant results, explain that web sources did not return enough - data and ask the user if they want you to retry with a refined query. - - Args: - - query: The search query - use specific, descriptive terms - - top_k: Number of results to retrieve (default: 10, max: 50) - - If search snippets are insufficient for the user's question, use `scrape_webpage` on the most relevant result URL for full content. - - When presenting results, reference sources as markdown links [descriptive text](url) — never bare URLs. -""" - -# Memory tool instructions have private and shared variants. -# We store them keyed as "update_memory" with sub-keys. -_MEMORY_TOOL_INSTRUCTIONS: dict[str, dict[str, str]] = { - "update_memory": { - "private": """ -- update_memory: Update your personal memory document about the user. - - Your current memory is already in in your context. The `chars` and - `limit` attributes show your current usage and the maximum allowed size. - - This is your curated long-term memory — the distilled essence of what you know about - the user, not raw conversation logs. - - Call update_memory when: - * The user explicitly asks to remember or forget something - * The user shares durable facts or preferences that will matter in future conversations - - The user's first name is provided in . Use it in memory entries - instead of "the user" (e.g. "{name} works at..." not "The user works at..."). - Do not store the name itself as a separate memory entry. - - Do not store short-lived or ephemeral info: one-off questions, greetings, - session logistics, or things that only matter for the current task. - - Args: - - updated_memory: The FULL updated markdown document (not a diff). - Merge new facts with existing ones, update contradictions, remove outdated entries. - Treat every update as a curation pass — consolidate, don't just append. - - Every bullet MUST use this format: - (YYYY-MM-DD) [marker] text - Markers: - [fact] — durable facts (role, background, projects, tools, expertise) - [pref] — preferences (response style, languages, formats, tools) - [instr] — standing instructions (always/never do, response rules) - - Keep it concise and well under the character limit shown in . - - Every entry MUST be under a `##` heading. Keep heading names short (2-3 words) and - natural. Do NOT include the user's name in headings. Organize by context — e.g. - who they are, what they're focused on, how they prefer things. Create, split, or - merge headings freely as the memory grows. - - Each entry MUST be a single bullet point. Be descriptive but concise — include relevant - details and context rather than just a few words. - - During consolidation, prioritize keeping: [instr] > [pref] > [fact]. -""", - "shared": """ -- update_memory: Update the team's shared memory document for this search space. - - Your current team memory is already in in your context. The `chars` - and `limit` attributes show current usage and the maximum allowed size. - - This is the team's curated long-term memory — decisions, conventions, key facts. - - NEVER store personal memory in team memory (e.g. personal bio, individual - preferences, or user-only standing instructions). - - Call update_memory when: - * A team member explicitly asks to remember or forget something - * The conversation surfaces durable team decisions, conventions, or facts - that will matter in future conversations - - Do not store short-lived or ephemeral info: one-off questions, greetings, - session logistics, or things that only matter for the current task. - - Args: - - updated_memory: The FULL updated markdown document (not a diff). - Merge new facts with existing ones, update contradictions, remove outdated entries. - Treat every update as a curation pass — consolidate, don't just append. - - Every bullet MUST use this format: - (YYYY-MM-DD) [fact] text - Team memory uses ONLY the [fact] marker. Never use [pref] or [instr] in team memory. - - Keep it concise and well under the character limit shown in . - - Every entry MUST be under a `##` heading. Keep heading names short (2-3 words) and - natural. Organize by context — e.g. what the team decided, current architecture, - active processes. Create, split, or merge headings freely as the memory grows. - - Each entry MUST be a single bullet point. Be descriptive but concise — include relevant - details and context rather than just a few words. - - During consolidation, prioritize keeping: decisions/conventions > key facts > current priorities. -""", - }, -} - -_MEMORY_TOOL_EXAMPLES: dict[str, dict[str, str]] = { - "update_memory": { - "private": """ -- Alex, is empty. User: "I'm a space enthusiast, explain astrophage to me" - - The user casually shared a durable fact. Use their first name in the entry, short neutral heading: - update_memory(updated_memory="## Interests & background\\n- (2025-03-15) [fact] Alex is a space enthusiast\\n") -- User: "Remember that I prefer concise answers over detailed explanations" - - Durable preference. Merge with existing memory, add a new heading: - update_memory(updated_memory="## Interests & background\\n- (2025-03-15) [fact] Alex is a space enthusiast\\n\\n## Response style\\n- (2025-03-15) [pref] Alex prefers concise answers over detailed explanations\\n") -- User: "I actually moved to Tokyo last month" - - Updated fact, date prefix reflects when recorded: - update_memory(updated_memory="## Interests & background\\n...\\n\\n## Personal context\\n- (2025-03-15) [fact] Alex lives in Tokyo (previously London)\\n...") -- User: "I'm a freelance photographer working on a nature documentary" - - Durable background info under a fitting heading: - update_memory(updated_memory="...\\n\\n## Current focus\\n- (2025-03-15) [fact] Alex is a freelance photographer\\n- (2025-03-15) [fact] Alex is working on a nature documentary\\n") -- User: "Always respond in bullet points" - - Standing instruction: - update_memory(updated_memory="...\\n\\n## Response style\\n- (2025-03-15) [instr] Always respond to Alex in bullet points\\n") -""", - "shared": """ -- User: "Let's remember that we decided to do weekly standup meetings on Mondays" - - Durable team decision: - update_memory(updated_memory="- (2025-03-15) [fact] Weekly standup meetings on Mondays\\n...") -- User: "Our office is in downtown Seattle, 5th floor" - - Durable team fact: - update_memory(updated_memory="- (2025-03-15) [fact] Office location: downtown Seattle, 5th floor\\n...") -""", - }, -} - -# Per-tool examples keyed by tool name. Only examples for enabled tools are included. -_TOOL_EXAMPLES: dict[str, str] = {} - -_TOOL_EXAMPLES["search_surfsense_docs"] = """ -- User: "How do I install SurfSense?" - - Call: `search_surfsense_docs(query="installation setup")` -- User: "What connectors does SurfSense support?" - - Call: `search_surfsense_docs(query="available connectors integrations")` -- User: "How do I set up the Notion connector?" - - Call: `search_surfsense_docs(query="Notion connector setup configuration")` -- User: "How do I use Docker to run SurfSense?" - - Call: `search_surfsense_docs(query="Docker installation setup")` -""" - -_TOOL_EXAMPLES["generate_podcast"] = """ -- User: "Give me a podcast about AI trends based on what we discussed" - - First search for relevant content, then call: `generate_podcast(source_content="Based on our conversation and search results: [detailed summary of chat + search findings]", podcast_title="AI Trends Podcast")` -- User: "Create a podcast summary of this conversation" - - Call: `generate_podcast(source_content="Complete conversation summary:\\n\\nUser asked about [topic 1]:\\n[Your detailed response]\\n\\nUser then asked about [topic 2]:\\n[Your detailed response]\\n\\n[Continue for all exchanges in the conversation]", podcast_title="Conversation Summary")` -- User: "Make a podcast about quantum computing" - - First explore `/documents/` (ls/glob/grep/read_file), then: `generate_podcast(source_content="Key insights about quantum computing from retrieved files:\\n\\n[Comprehensive summary of findings]", podcast_title="Quantum Computing Explained")` -""" - -_TOOL_EXAMPLES["generate_video_presentation"] = """ -- User: "Give me a presentation about AI trends based on what we discussed" - - First search for relevant content, then call: `generate_video_presentation(source_content="Based on our conversation and search results: [detailed summary of chat + search findings]", video_title="AI Trends Presentation")` -- User: "Create slides summarizing this conversation" - - Call: `generate_video_presentation(source_content="Complete conversation summary:\\n\\nUser asked about [topic 1]:\\n[Your detailed response]\\n\\nUser then asked about [topic 2]:\\n[Your detailed response]\\n\\n[Continue for all exchanges in the conversation]", video_title="Conversation Summary")` -- User: "Make a video presentation about quantum computing" - - First explore `/documents/` (ls/glob/grep/read_file), then: `generate_video_presentation(source_content="Key insights about quantum computing from retrieved files:\\n\\n[Comprehensive summary of findings]", video_title="Quantum Computing Explained")` -""" - -_TOOL_EXAMPLES["generate_report"] = """ -- User: "Generate a report about AI trends" - - Call: `generate_report(topic="AI Trends Report", source_strategy="kb_search", search_queries=["AI trends recent developments", "artificial intelligence industry trends", "AI market growth and predictions"], report_style="detailed")` - - WHY: Has creation verb "generate" → call the tool. No prior discussion → use kb_search. -- User: "Write a research report from this conversation" - - Call: `generate_report(topic="Research Report", source_strategy="conversation", source_content="Complete conversation summary:\\n\\n...", report_style="deep_research")` - - WHY: Has creation verb "write" → call the tool. Conversation has the content → use source_strategy="conversation". -- User: (after a report on Climate Change was generated) "Add a section about carbon capture technologies" - - Call: `generate_report(topic="Climate Crisis: Causes, Impacts, and Solutions", source_strategy="conversation", source_content="[summary of conversation context if any]", parent_report_id=, user_instructions="Add a new section about carbon capture technologies")` - - WHY: Has modification verb "add" + specific deliverable target → call the tool with parent_report_id. -- User: (after a report was generated) "What else could we add to have more depth?" - - Do NOT call generate_report. Answer in chat with suggestions. - - WHY: No creation/modification verb directed at producing a deliverable. -""" - -_TOOL_EXAMPLES["scrape_webpage"] = """ -- User: "Check out https://dev.to/some-article" - - Call: `scrape_webpage(url="https://dev.to/some-article")` - - Respond with a structured analysis — key points, takeaways. -- User: "Read this article and summarize it for me: https://example.com/blog/ai-trends" - - Call: `scrape_webpage(url="https://example.com/blog/ai-trends")` - - Respond with a thorough summary using headings and bullet points. -- User: (after discussing https://example.com/stats) "Can you get the live data from that page?" - - Call: `scrape_webpage(url="https://example.com/stats")` - - IMPORTANT: Always attempt scraping first. Never refuse before trying the tool. -- User: "https://example.com/blog/weekend-recipes" - - Call: `scrape_webpage(url="https://example.com/blog/weekend-recipes")` - - When a user sends just a URL with no instructions, scrape it and provide a concise summary of the content. -""" - -_TOOL_EXAMPLES["generate_image"] = """ -- User: "Generate an image of a cat" - - Call: `generate_image(prompt="A fluffy orange tabby cat sitting on a windowsill, bathed in warm golden sunlight, soft bokeh background with green houseplants, photorealistic style, cozy atmosphere")` - - The generated image will automatically be displayed in the chat. -- User: "Draw me a logo for a coffee shop called Bean Dream" - - Call: `generate_image(prompt="Minimalist modern logo design for a coffee shop called 'Bean Dream', featuring a stylized coffee bean with dream-like swirls of steam, clean vector style, warm brown and cream color palette, white background, professional branding")` - - The generated image will automatically be displayed in the chat. -- User: "Show me this image: https://example.com/image.png" - - Simply include it in your response using markdown: `![Image](https://example.com/image.png)` -- User uploads an image file and asks: "What is this image about?" - - The user's uploaded image is already visible in the chat. - - Simply analyze the image content and respond directly. -""" - -_TOOL_EXAMPLES["web_search"] = """ -- User: "What's the current USD to INR exchange rate?" - - Call: `web_search(query="current USD to INR exchange rate")` - - Then answer using the returned web results with citations. -- User: "What's the latest news about AI?" - - Call: `web_search(query="latest AI news today")` -- User: "What's the weather in New York?" - - Call: `web_search(query="weather New York today")` -""" - -_TOOL_INSTRUCTIONS["generate_resume"] = """ -- generate_resume: Generate or revise a professional resume as a Typst document. - - WHEN TO CALL: The user asks to create, build, generate, write, or draft a resume or CV. - Also when they ask to modify, update, or revise an existing resume from this conversation. - - WHEN NOT TO CALL: General career advice, resume tips, cover letters, or reviewing - a resume without making changes. For cover letters, use generate_report instead. - - The tool produces Typst source code that is compiled to a PDF preview automatically. - - PAGE POLICY: - - Default behavior is ONE PAGE. For new resume creation, set max_pages=1 unless the user explicitly asks for more. - - If the user requests a longer resume (e.g., "make it 2 pages"), set max_pages to that value. - - Args: - - user_info: The user's resume content — work experience, education, skills, contact - info, etc. Can be structured or unstructured text. - CRITICAL: user_info must be COMPREHENSIVE. Do NOT just pass the user's raw message. - You MUST gather and consolidate ALL available information: - * Content from referenced/mentioned documents (e.g., uploaded resumes, CVs, LinkedIn profiles) - that appear in the conversation context — extract and include their FULL content. - * Information the user shared across multiple messages in the conversation. - * Any relevant details from knowledge base search results in the context. - The more complete the user_info, the better the resume. Include names, contact info, - work experience with dates, education, skills, projects, certifications — everything available. - - user_instructions: Optional style or content preferences (e.g. "emphasize leadership", - "keep it to one page"). For revisions, describe what to change. - - parent_report_id: Set this when the user wants to MODIFY an existing resume from - this conversation. Use the report_id from a previous generate_resume result. - - max_pages: Maximum resume length in pages (integer 1-5). Default is 1. - - Returns: Dict with status, report_id, title, and content_type. - - After calling: Give a brief confirmation. Do NOT paste resume content in chat. Do NOT mention report_id or any internal IDs — the resume card is shown automatically. - - VERSIONING: Same rules as generate_report — set parent_report_id for modifications - of an existing resume, leave as None for new resumes. -""" - -_TOOL_EXAMPLES["generate_resume"] = """ -- User: "Build me a resume. I'm John Doe, engineer at Acme Corp..." - - Call: `generate_resume(user_info="John Doe, engineer at Acme Corp...", max_pages=1)` - - WHY: Has creation verb "build" + resume → call the tool. -- User: "Create my CV with this info: [experience, education, skills]" - - Call: `generate_resume(user_info="[experience, education, skills]", max_pages=1)` -- User: "Build me a resume" (and there is a resume/CV document in the conversation context) - - Extract the FULL content from the document in context, then call: - `generate_resume(user_info="Name: John Doe\\nEmail: john@example.com\\n\\nExperience:\\n- Senior Engineer at Acme Corp (2020-2024)\\n Led team of 5...\\n\\nEducation:\\n- BS Computer Science, MIT (2016-2020)\\n\\nSkills: Python, TypeScript, AWS...", max_pages=1)` - - WHY: Document content is available in context — extract ALL of it into user_info. Do NOT ignore referenced documents. -- User: (after resume generated) "Change my title to Senior Engineer" - - Call: `generate_resume(user_info="", user_instructions="Change the job title to Senior Engineer", parent_report_id=, max_pages=1)` - - WHY: Modification verb "change" + refers to existing resume → set parent_report_id. -- User: (after resume generated) "Make this 2 pages and expand projects" - - Call: `generate_resume(user_info="", user_instructions="Expand projects and keep this to at most 2 pages", parent_report_id=, max_pages=2)` - - WHY: Explicit page increase request → set max_pages to 2. -- User: "How should I structure my resume?" - - Do NOT call generate_resume. Answer in chat with advice. - - WHY: No creation/modification verb. -""" - -# All tool names that have prompt instructions (order matters for prompt readability) -_ALL_TOOL_NAMES_ORDERED = [ - "search_surfsense_docs", - "web_search", - "generate_podcast", - "generate_video_presentation", - "generate_report", - "generate_resume", - "generate_image", - "scrape_webpage", - "update_memory", -] - - -def _format_tool_name(name: str) -> str: - """Convert snake_case tool name to a human-readable label.""" - return name.replace("_", " ").title() - - -def _get_tools_instructions( - thread_visibility: ChatVisibility | None = None, - enabled_tool_names: set[str] | None = None, - disabled_tool_names: set[str] | None = None, -) -> str: - """Build tools instructions containing only the enabled tools. - - Args: - thread_visibility: Private vs shared — affects memory tool wording. - enabled_tool_names: Set of tool names that are actually bound to the agent. - When None, all tools are included (backward-compatible default). - disabled_tool_names: Set of tool names that the user explicitly disabled. - When provided, a note is appended telling the model about these tools - so it can inform the user they can re-enable them. - """ - visibility = thread_visibility or ChatVisibility.PRIVATE - memory_variant = ( - "shared" if visibility == ChatVisibility.SEARCH_SPACE else "private" - ) - - parts: list[str] = [_TOOLS_PREAMBLE] - examples: list[str] = [] - - for tool_name in _ALL_TOOL_NAMES_ORDERED: - if enabled_tool_names is not None and tool_name not in enabled_tool_names: - continue - - if tool_name in _TOOL_INSTRUCTIONS: - parts.append(_TOOL_INSTRUCTIONS[tool_name]) - elif tool_name in _MEMORY_TOOL_INSTRUCTIONS: - parts.append(_MEMORY_TOOL_INSTRUCTIONS[tool_name][memory_variant]) - - if tool_name in _TOOL_EXAMPLES: - examples.append(_TOOL_EXAMPLES[tool_name]) - elif tool_name in _MEMORY_TOOL_EXAMPLES: - examples.append(_MEMORY_TOOL_EXAMPLES[tool_name][memory_variant]) - - # Append a note about user-disabled tools so the model can inform the user - known_disabled = ( - disabled_tool_names & set(_ALL_TOOL_NAMES_ORDERED) - if disabled_tool_names - else set() - ) - if known_disabled: - disabled_list = ", ".join( - _format_tool_name(n) for n in _ALL_TOOL_NAMES_ORDERED if n in known_disabled - ) - parts.append(f""" -DISABLED TOOLS (by user): -The following tools are available in SurfSense but have been disabled by the user for this session: {disabled_list}. -You do NOT have access to these tools and MUST NOT claim you can use them. -If the user asks about a capability provided by a disabled tool, let them know the relevant tool -is currently disabled and they can re-enable it. -""") - - parts.append("\n\n") - - if examples: - parts.append("") - parts.extend(examples) - parts.append("\n") - - return "".join(parts) - - -# Backward-compatible constant: all tools included (private memory variant) -SURFSENSE_TOOLS_INSTRUCTIONS = _get_tools_instructions() - - -SURFSENSE_CITATION_INSTRUCTIONS = """ - -CRITICAL CITATION REQUIREMENTS: - -1. For EVERY piece of information you include from the documents, add a citation in the format [citation:chunk_id] where chunk_id is the exact value from the `` tag inside ``. -2. Make sure ALL factual statements from the documents have proper citations. -3. If multiple chunks support the same point, include all relevant citations [citation:chunk_id1], [citation:chunk_id2]. -4. You MUST use the exact chunk_id values from the `` attributes. Do not create your own citation numbers. -5. Every citation MUST be in the format [citation:chunk_id] where chunk_id is the exact chunk id value. -6. Never modify or change the chunk_id - always use the original values exactly as provided in the chunk tags. -7. Do not return citations as clickable links. -8. Never format citations as markdown links like "([citation:5](https://example.com))". Always use plain square brackets only. -9. Citations must ONLY appear as [citation:chunk_id] or [citation:chunk_id1], [citation:chunk_id2] format - never with parentheses, hyperlinks, or other formatting. -10. Never make up chunk IDs. Only use chunk_id values that are explicitly provided in the `` tags. -11. If you are unsure about a chunk_id, do not include a citation rather than guessing or making one up. - - -The documents you receive are structured like this: - -**Knowledge base documents (numeric chunk IDs):** - - - 42 - GITHUB_CONNECTOR - <![CDATA[Some repo / file / issue title]]> - - - - - - - - - - -**Web search results (URL chunk IDs):** - - - WEB_SEARCH - <![CDATA[Some web search result]]> - - - - - - - - -IMPORTANT: You MUST cite using the EXACT chunk ids from the `` tags. -- For knowledge base documents, chunk ids are numeric (e.g. 123, 124) or prefixed (e.g. doc-45). -- For live web search results, chunk ids are URLs (e.g. https://example.com/article). -Do NOT cite document_id. Always use the chunk id. - - - -- Every fact from the documents must have a citation in the format [citation:chunk_id] where chunk_id is the EXACT id value from a `` tag -- Citations should appear at the end of the sentence containing the information they support -- Multiple citations should be separated by commas: [citation:chunk_id1], [citation:chunk_id2], [citation:chunk_id3] -- No need to return references section. Just citations in answer. -- NEVER create your own citation format - use the exact chunk_id values from the documents in the [citation:chunk_id] format -- NEVER format citations as clickable links or as markdown links like "([citation:5](https://example.com))". Always use plain square brackets only -- NEVER make up chunk IDs if you are unsure about the chunk_id. It is better to omit the citation than to guess -- Copy the EXACT chunk id from the XML - if it says ``, use [citation:doc-123] -- If the chunk id is a URL like ``, use [citation:https://example.com/page] - - - -CORRECT citation formats: -- [citation:5] (numeric chunk ID from knowledge base) -- [citation:doc-123] (for Surfsense documentation chunks) -- [citation:https://example.com/article] (URL chunk ID from web search results) -- [citation:chunk_id1], [citation:chunk_id2], [citation:chunk_id3] (multiple citations) - -INCORRECT citation formats (DO NOT use): -- Using parentheses and markdown links: ([citation:5](https://github.com/MODSetter/SurfSense)) -- Using parentheses around brackets: ([citation:5]) -- Using hyperlinked text: [link to source 5](https://example.com) -- Using footnote style: ... library¹ -- Making up source IDs when source_id is unknown -- Using old IEEE format: [1], [2], [3] -- Using source types instead of IDs: [citation:GITHUB_CONNECTOR] instead of [citation:5] - - - -Based on your GitHub repositories and video content, Python's asyncio library provides tools for writing concurrent code using the async/await syntax [citation:5]. It's particularly useful for I/O-bound and high-level structured network code [citation:5]. - -According to web search results, the key advantage of asyncio is that it can improve performance by allowing other code to run while waiting for I/O operations to complete [citation:https://docs.python.org/3/library/asyncio.html]. This makes it excellent for scenarios like web scraping, API calls, database operations, or any situation where your program spends time waiting for external resources. - -However, from your video learning, it's important to note that asyncio is not suitable for CPU-bound tasks as it runs on a single thread [citation:12]. For computationally intensive work, you'd want to use multiprocessing instead. - - -""" - -# Anti-citation prompt - used when citations are disabled -# This explicitly tells the model NOT to include citations -SURFSENSE_NO_CITATION_INSTRUCTIONS = """ - -IMPORTANT: Citations are DISABLED for this configuration. - -DO NOT include any citations in your responses. Specifically: -1. Do NOT use the [citation:chunk_id] format anywhere in your response. -2. Do NOT reference document IDs, chunk IDs, or source IDs. -3. Simply provide the information naturally without any citation markers. -4. Write your response as if you're having a normal conversation, incorporating the information from your knowledge seamlessly. - -When answering questions based on documents from the knowledge base: -- Present the information directly and confidently -- Do not mention that information comes from specific documents or chunks -- Integrate facts naturally into your response without attribution markers - -Your goal is to provide helpful, informative answers in a clean, readable format without any citation notation. - -""" - - -def _build_mcp_routing_block( - mcp_connector_tools: dict[str, list[str]] | None, -) -> str: - """Build an additional tool routing block for generic MCP connectors. - - When users add MCP servers (e.g. GitLab, GitHub), the LLM needs to know - those tools exist and should be called directly — not searched in the - knowledge base. - """ - if not mcp_connector_tools: - return "" - - lines = [ - "\n", - "You also have direct tools from these user-connected MCP servers.", - "Their data is NEVER in the knowledge base — call their tools directly.", - "", - ] - for server_name, tool_names in mcp_connector_tools.items(): - lines.append(f"- {server_name} → {', '.join(tool_names)}") - lines.append("\n") - return "\n".join(lines) +from .prompts.composer import ( + _read_fragment, + compose_system_prompt, + detect_provider_variant, +) + +# Public re-exports for backwards compatibility (some legacy code reads the +# raw default-instructions text directly). +SURFSENSE_SYSTEM_INSTRUCTIONS_TEMPLATE = ( + "\nDefault SurfSense agent system instructions are now\n" + "composed from prompts/base/*.md. See compose_system_prompt() for details.\n" + "" +) + +# Citation block re-exposed for legacy importers that referenced this constant +# directly. The composer is the canonical source; this is a frozen snapshot +# loaded at module-init time. +SURFSENSE_CITATION_INSTRUCTIONS = _read_fragment("base/citations_on.md") +SURFSENSE_NO_CITATION_INSTRUCTIONS = _read_fragment("base/citations_off.md") def build_surfsense_system_prompt( @@ -845,36 +46,23 @@ def build_surfsense_system_prompt( enabled_tool_names: set[str] | None = None, disabled_tool_names: set[str] | None = None, mcp_connector_tools: dict[str, list[str]] | None = None, + *, + model_name: str | None = None, ) -> str: + """Build the default SurfSense system prompt (citations on, defaults). + + See :func:`app.agents.new_chat.prompts.composer.compose_system_prompt` + for full parameter docs. """ - Build the SurfSense system prompt with default settings. - - This is a convenience function that builds the prompt with: - - Default system instructions - - Tools instructions (only for enabled tools) - - Citation instructions enabled - - Args: - today: Optional datetime for today's date (defaults to current UTC date) - thread_visibility: Optional; when provided, used for conditional prompt (e.g. private vs shared memory wording). Defaults to private behavior when None. - enabled_tool_names: Set of tool names actually bound to the agent. When None all tools are included. - disabled_tool_names: Set of tool names the user explicitly disabled. Included as a note so the model can inform the user. - mcp_connector_tools: Mapping of MCP server display name → list of tool names - for generic MCP connectors. Injected into the system prompt so the LLM - knows to call these tools directly. - - Returns: - Complete system prompt string - """ - - visibility = thread_visibility or ChatVisibility.PRIVATE - system_instructions = _get_system_instructions(visibility, today) - system_instructions += _build_mcp_routing_block(mcp_connector_tools) - tools_instructions = _get_tools_instructions( - visibility, enabled_tool_names, disabled_tool_names + return compose_system_prompt( + today=today, + thread_visibility=thread_visibility, + enabled_tool_names=enabled_tool_names, + disabled_tool_names=disabled_tool_names, + mcp_connector_tools=mcp_connector_tools, + citations_enabled=True, + model_name=model_name, ) - citation_instructions = SURFSENSE_CITATION_INSTRUCTIONS - return system_instructions + tools_instructions + citation_instructions def build_configurable_system_prompt( @@ -886,75 +74,54 @@ def build_configurable_system_prompt( enabled_tool_names: set[str] | None = None, disabled_tool_names: set[str] | None = None, mcp_connector_tools: dict[str, list[str]] | None = None, + *, + model_name: str | None = None, ) -> str: + """Build a configurable SurfSense system prompt (NewLLMConfig path). + + See :func:`app.agents.new_chat.prompts.composer.compose_system_prompt` + for full parameter docs. """ - Build a configurable SurfSense system prompt based on NewLLMConfig settings. - - The prompt is composed of three parts: - 1. System Instructions - either custom or default SURFSENSE_SYSTEM_INSTRUCTIONS - 2. Tools Instructions - only for enabled tools, with a note about disabled ones - 3. Citation Instructions - either SURFSENSE_CITATION_INSTRUCTIONS or SURFSENSE_NO_CITATION_INSTRUCTIONS - - Args: - custom_system_instructions: Custom system instructions to use. If empty/None and - use_default_system_instructions is True, defaults to - SURFSENSE_SYSTEM_INSTRUCTIONS. - use_default_system_instructions: Whether to use default instructions when - custom_system_instructions is empty/None. - citations_enabled: Whether to include citation instructions (True) or - anti-citation instructions (False). - today: Optional datetime for today's date (defaults to current UTC date) - thread_visibility: Optional; when provided, used for conditional prompt (e.g. private vs shared memory wording). Defaults to private behavior when None. - enabled_tool_names: Set of tool names actually bound to the agent. When None all tools are included. - disabled_tool_names: Set of tool names the user explicitly disabled. Included as a note so the model can inform the user. - mcp_connector_tools: Mapping of MCP server display name → list of tool names - for generic MCP connectors. Injected into the system prompt so the LLM - knows to call these tools directly. - - Returns: - Complete system prompt string - """ - resolved_today = (today or datetime.now(UTC)).astimezone(UTC).date().isoformat() - - # Determine system instructions - if custom_system_instructions and custom_system_instructions.strip(): - system_instructions = custom_system_instructions.format( - resolved_today=resolved_today - ) - elif use_default_system_instructions: - visibility = thread_visibility or ChatVisibility.PRIVATE - system_instructions = _get_system_instructions(visibility, today) - else: - system_instructions = "" - - system_instructions += _build_mcp_routing_block(mcp_connector_tools) - - # Tools instructions: only include enabled tools, note disabled ones - tools_instructions = _get_tools_instructions( - thread_visibility, enabled_tool_names, disabled_tool_names + return compose_system_prompt( + today=today, + thread_visibility=thread_visibility, + enabled_tool_names=enabled_tool_names, + disabled_tool_names=disabled_tool_names, + mcp_connector_tools=mcp_connector_tools, + custom_system_instructions=custom_system_instructions, + use_default_system_instructions=use_default_system_instructions, + citations_enabled=citations_enabled, + model_name=model_name, ) - # Citation instructions based on toggle - citation_instructions = ( - SURFSENSE_CITATION_INSTRUCTIONS - if citations_enabled - else SURFSENSE_NO_CITATION_INSTRUCTIONS - ) - - return system_instructions + tools_instructions + citation_instructions - def get_default_system_instructions() -> str: + """Return the default ```` block (no tools / citations). + + Useful for populating the UI when seeding ``NewLLMConfig.system_instructions``. + The output reflects the current fragment tree, not a baked-in constant. """ - Get the default system instructions template. + resolved_today = datetime.now(UTC).date().isoformat() + from .prompts.composer import _build_system_instructions # local import - This is useful for populating the UI with the default value when - creating a new NewLLMConfig. - - Returns: - Default system instructions string (with {resolved_today} placeholder) - """ - return SURFSENSE_SYSTEM_INSTRUCTIONS.strip() + return _build_system_instructions( + visibility=ChatVisibility.PRIVATE, + resolved_today=resolved_today, + ).strip() +# Backwards compatibility — some modules import the constant directly. SURFSENSE_SYSTEM_PROMPT = build_surfsense_system_prompt() + + +__all__ = [ + "SURFSENSE_CITATION_INSTRUCTIONS", + "SURFSENSE_NO_CITATION_INSTRUCTIONS", + "SURFSENSE_SYSTEM_INSTRUCTIONS_TEMPLATE", + "SURFSENSE_SYSTEM_PROMPT", + "build_configurable_system_prompt", + "build_surfsense_system_prompt", + "compose_system_prompt", + "detect_provider_variant", + "get_default_system_instructions", +] diff --git a/surfsense_backend/app/agents/new_chat/tools/invalid_tool.py b/surfsense_backend/app/agents/new_chat/tools/invalid_tool.py new file mode 100644 index 000000000..df10fcbe3 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/tools/invalid_tool.py @@ -0,0 +1,52 @@ +""" +The ``invalid`` fallback tool. + +When the model emits a tool call whose name doesn't match any registered +tool, :class:`ToolCallNameRepairMiddleware` rewrites the call to ``invalid`` +with the original name and a parser/validation error string. This tool's +execution then returns that error to the model so it can self-correct. + +Mirrors ``opencode/packages/opencode/src/tool/invalid.ts``. Tier 1.6 in +the OpenCode-port plan. + +Critically, the :class:`ToolDefinition` for this tool is **excluded** from +the system-prompt tool list and from ``LLMToolSelectorMiddleware`` selection +(see ``ToolDefinition.always_include`` filtering in the registry) — the +model never advertises ``invalid`` as a callable. It only ever shows up +in the tool registry so LangGraph can dispatch the rewritten call. +""" + +from __future__ import annotations + +from langchain_core.tools import tool + +INVALID_TOOL_NAME = "invalid" +INVALID_TOOL_DESCRIPTION = "Do not use" + + +def _format_invalid_message(tool: str | None, error: str | None) -> str: + """Return the user-visible error string. Mirrors ``invalid.ts``.""" + name = tool or "" + detail = error or "(no error message provided)" + return ( + f"The arguments provided to the tool `{name}` are invalid: {detail}\n" + f"Read the tool's docstring carefully and try again with valid arguments." + ) + + +@tool(name_or_callable=INVALID_TOOL_NAME, description=INVALID_TOOL_DESCRIPTION) +def invalid_tool(tool: str | None = None, error: str | None = None) -> str: + """Return a human-readable explanation of a tool-call validation failure. + + Activated only when :class:`ToolCallNameRepairMiddleware` rewrites a + failed tool call to ``invalid`` with the original tool name and the + error message produced during validation. + """ + return _format_invalid_message(tool, error) + + +__all__ = [ + "INVALID_TOOL_DESCRIPTION", + "INVALID_TOOL_NAME", + "invalid_tool", +] diff --git a/surfsense_backend/app/agents/new_chat/tools/registry.py b/surfsense_backend/app/agents/new_chat/tools/registry.py index 3ac8677b9..f5ee1a61d 100644 --- a/surfsense_backend/app/agents/new_chat/tools/registry.py +++ b/surfsense_backend/app/agents/new_chat/tools/registry.py @@ -43,6 +43,9 @@ from typing import Any from langchain_core.tools import BaseTool +from app.agents.new_chat.middleware.dedup_tool_calls import ( + wrap_dedup_key_by_arg_name, +) from app.db import ChatVisibility from .confluence import ( @@ -125,6 +128,14 @@ class ToolDefinition: enabled_by_default: Whether the tool is enabled when no explicit config is provided required_connector: Searchable type string (e.g. ``"LINEAR_CONNECTOR"``) that must be in ``available_connectors`` for the tool to be enabled. + dedup_key: Optional callable that maps a tool's ``args`` dict to a + string signature used by :class:`DedupHITLToolCallsMiddleware` + to drop duplicate calls. Replaces the legacy hardcoded + ``_NATIVE_HITL_TOOL_DEDUP_KEYS`` map (Tier 2.3 in the + OpenCode-port plan). + reverse: Optional callable that, given the tool's ``(args, result)``, + returns a ``ReverseDescriptor`` describing the inverse tool + invocation. Consumed by the snapshot/revert pipeline (Tier 5). """ @@ -135,6 +146,8 @@ class ToolDefinition: enabled_by_default: bool = True hidden: bool = False required_connector: str | None = None + dedup_key: Callable[[dict[str, Any]], str] | None = None + reverse: Callable[[dict[str, Any], Any], dict[str, Any]] | None = None # ============================================================================= @@ -288,6 +301,7 @@ BUILTIN_TOOLS: list[ToolDefinition] = [ ), requires=["db_session", "search_space_id", "user_id"], required_connector="NOTION_CONNECTOR", + dedup_key=wrap_dedup_key_by_arg_name("title"), ), ToolDefinition( name="update_notion_page", @@ -299,6 +313,7 @@ BUILTIN_TOOLS: list[ToolDefinition] = [ ), requires=["db_session", "search_space_id", "user_id"], required_connector="NOTION_CONNECTOR", + dedup_key=wrap_dedup_key_by_arg_name("page_title"), ), ToolDefinition( name="delete_notion_page", @@ -310,6 +325,7 @@ BUILTIN_TOOLS: list[ToolDefinition] = [ ), requires=["db_session", "search_space_id", "user_id"], required_connector="NOTION_CONNECTOR", + dedup_key=wrap_dedup_key_by_arg_name("page_title"), ), # ========================================================================= # GOOGLE DRIVE TOOLS - create files, delete files @@ -325,6 +341,7 @@ BUILTIN_TOOLS: list[ToolDefinition] = [ ), requires=["db_session", "search_space_id", "user_id"], required_connector="GOOGLE_DRIVE_FILE", + dedup_key=wrap_dedup_key_by_arg_name("file_name"), ), ToolDefinition( name="delete_google_drive_file", @@ -336,6 +353,7 @@ BUILTIN_TOOLS: list[ToolDefinition] = [ ), requires=["db_session", "search_space_id", "user_id"], required_connector="GOOGLE_DRIVE_FILE", + dedup_key=wrap_dedup_key_by_arg_name("file_name"), ), # ========================================================================= # DROPBOX TOOLS - create and trash files @@ -351,6 +369,7 @@ BUILTIN_TOOLS: list[ToolDefinition] = [ ), requires=["db_session", "search_space_id", "user_id"], required_connector="DROPBOX_FILE", + dedup_key=wrap_dedup_key_by_arg_name("file_name"), ), ToolDefinition( name="delete_dropbox_file", @@ -362,6 +381,7 @@ BUILTIN_TOOLS: list[ToolDefinition] = [ ), requires=["db_session", "search_space_id", "user_id"], required_connector="DROPBOX_FILE", + dedup_key=wrap_dedup_key_by_arg_name("file_name"), ), # ========================================================================= # ONEDRIVE TOOLS - create and trash files @@ -377,6 +397,7 @@ BUILTIN_TOOLS: list[ToolDefinition] = [ ), requires=["db_session", "search_space_id", "user_id"], required_connector="ONEDRIVE_FILE", + dedup_key=wrap_dedup_key_by_arg_name("file_name"), ), ToolDefinition( name="delete_onedrive_file", @@ -388,6 +409,7 @@ BUILTIN_TOOLS: list[ToolDefinition] = [ ), requires=["db_session", "search_space_id", "user_id"], required_connector="ONEDRIVE_FILE", + dedup_key=wrap_dedup_key_by_arg_name("file_name"), ), # ========================================================================= # GOOGLE CALENDAR TOOLS - search, create, update, delete events @@ -414,6 +436,7 @@ BUILTIN_TOOLS: list[ToolDefinition] = [ ), requires=["db_session", "search_space_id", "user_id"], required_connector="GOOGLE_CALENDAR_CONNECTOR", + dedup_key=wrap_dedup_key_by_arg_name("title"), ), ToolDefinition( name="update_calendar_event", @@ -425,6 +448,7 @@ BUILTIN_TOOLS: list[ToolDefinition] = [ ), requires=["db_session", "search_space_id", "user_id"], required_connector="GOOGLE_CALENDAR_CONNECTOR", + dedup_key=wrap_dedup_key_by_arg_name("event_title_or_id"), ), ToolDefinition( name="delete_calendar_event", @@ -436,6 +460,7 @@ BUILTIN_TOOLS: list[ToolDefinition] = [ ), requires=["db_session", "search_space_id", "user_id"], required_connector="GOOGLE_CALENDAR_CONNECTOR", + dedup_key=wrap_dedup_key_by_arg_name("event_title_or_id"), ), # ========================================================================= # GMAIL TOOLS - search, read, create drafts, update drafts, send, trash @@ -473,6 +498,7 @@ BUILTIN_TOOLS: list[ToolDefinition] = [ ), requires=["db_session", "search_space_id", "user_id"], required_connector="GOOGLE_GMAIL_CONNECTOR", + dedup_key=wrap_dedup_key_by_arg_name("subject"), ), ToolDefinition( name="send_gmail_email", @@ -484,6 +510,7 @@ BUILTIN_TOOLS: list[ToolDefinition] = [ ), requires=["db_session", "search_space_id", "user_id"], required_connector="GOOGLE_GMAIL_CONNECTOR", + dedup_key=wrap_dedup_key_by_arg_name("subject"), ), ToolDefinition( name="trash_gmail_email", @@ -495,6 +522,7 @@ BUILTIN_TOOLS: list[ToolDefinition] = [ ), requires=["db_session", "search_space_id", "user_id"], required_connector="GOOGLE_GMAIL_CONNECTOR", + dedup_key=wrap_dedup_key_by_arg_name("email_subject_or_id"), ), ToolDefinition( name="update_gmail_draft", @@ -506,6 +534,7 @@ BUILTIN_TOOLS: list[ToolDefinition] = [ ), requires=["db_session", "search_space_id", "user_id"], required_connector="GOOGLE_GMAIL_CONNECTOR", + dedup_key=wrap_dedup_key_by_arg_name("draft_subject_or_id"), ), # ========================================================================= # CONFLUENCE TOOLS - create, update, delete pages @@ -521,6 +550,7 @@ BUILTIN_TOOLS: list[ToolDefinition] = [ ), requires=["db_session", "search_space_id", "user_id"], required_connector="CONFLUENCE_CONNECTOR", + dedup_key=wrap_dedup_key_by_arg_name("title"), ), ToolDefinition( name="update_confluence_page", @@ -532,6 +562,7 @@ BUILTIN_TOOLS: list[ToolDefinition] = [ ), requires=["db_session", "search_space_id", "user_id"], required_connector="CONFLUENCE_CONNECTOR", + dedup_key=wrap_dedup_key_by_arg_name("page_title_or_id"), ), ToolDefinition( name="delete_confluence_page", @@ -543,6 +574,7 @@ BUILTIN_TOOLS: list[ToolDefinition] = [ ), requires=["db_session", "search_space_id", "user_id"], required_connector="CONFLUENCE_CONNECTOR", + dedup_key=wrap_dedup_key_by_arg_name("page_title_or_id"), ), # ========================================================================= # DISCORD TOOLS - list channels, read messages, send messages @@ -755,6 +787,24 @@ def build_tools( # Create the tool tool = tool_def.factory(dependencies) + # Propagate the registry-level metadata so middleware (e.g. + # ``DedupHITLToolCallsMiddleware``) and the action-log/revert + # pipeline can pick the resolvers up via ``tool.metadata`` without + # re-importing :data:`BUILTIN_TOOLS`. + if tool_def.dedup_key is not None or tool_def.reverse is not None: + existing_meta = getattr(tool, "metadata", None) or {} + merged_meta = dict(existing_meta) + if tool_def.dedup_key is not None: + merged_meta.setdefault("dedup_key", tool_def.dedup_key) + if tool_def.reverse is not None: + merged_meta.setdefault("reverse", tool_def.reverse) + try: + tool.metadata = merged_meta + except Exception: + logger.debug( + "Tool %s rejected metadata mutation; relying on registry lookup", + tool_def.name, + ) tools.append(tool) # Add any additional custom tools diff --git a/surfsense_backend/app/db.py b/surfsense_backend/app/db.py index e16590afc..fcd342d29 100644 --- a/surfsense_backend/app/db.py +++ b/surfsense_backend/app/db.py @@ -2250,6 +2250,202 @@ else: ) +class AgentActionLog(BaseModel): + """Append-only audit trail of every tool call dispatched by the agent. + + One row per ``ToolMessage`` produced; written by ``ActionLogMiddleware`` + in its ``aafter_tool`` hook. Rows are referenced by the + ``/api/threads/{thread_id}/revert/{action_id}`` route to look up an + action's stored ``reverse_descriptor`` and replay it. + + The table is intentionally narrow: large tool outputs are NOT stored + here. Result text lives in the langgraph checkpoint; this row only + keeps a short ``result_id`` (the LangChain ``ToolMessage.id`` or a + spilled-content path) for correlation. + """ + + __tablename__ = "agent_action_log" + + thread_id = Column( + Integer, + ForeignKey("new_chat_threads.id", ondelete="CASCADE"), + nullable=False, + index=True, + ) + user_id = Column( + UUID(as_uuid=True), + ForeignKey("user.id", ondelete="SET NULL"), + nullable=True, + index=True, + ) + search_space_id = Column( + Integer, + ForeignKey("searchspaces.id", ondelete="CASCADE"), + nullable=False, + index=True, + ) + turn_id = Column(String(64), nullable=True, index=True) + message_id = Column(String(128), nullable=True, index=True) + tool_name = Column(String(255), nullable=False, index=True) + args = Column(JSONB, nullable=True) + result_id = Column(String(255), nullable=True) + reversible = Column( + Boolean, nullable=False, default=False, server_default=text("false") + ) + reverse_descriptor = Column(JSONB, nullable=True) + error = Column(JSONB, nullable=True) + reverse_of = Column( + Integer, + ForeignKey("agent_action_log.id", ondelete="SET NULL"), + nullable=True, + index=True, + ) + created_at = Column( + TIMESTAMP(timezone=True), + nullable=False, + default=lambda: datetime.now(UTC), + server_default=text("(now() AT TIME ZONE 'utc')"), + index=True, + ) + + __table_args__ = ( + Index("ix_agent_action_log_thread_created", "thread_id", "created_at"), + ) + + +class DocumentRevision(BaseModel): + """Snapshot of a :class:`Document` row taken before a mutating tool call. + + Written by :class:`KnowledgeBasePersistenceMiddleware` (or its safety-net + `commit_staged_filesystem_state`) ahead of any NOTE / FILE / EXTENSION + document write. The row is referenced by ``/revert/{action_id}`` to + restore the original content in place. + """ + + __tablename__ = "document_revisions" + + document_id = Column( + Integer, + ForeignKey("documents.id", ondelete="CASCADE"), + nullable=False, + index=True, + ) + search_space_id = Column( + Integer, + ForeignKey("searchspaces.id", ondelete="CASCADE"), + nullable=False, + index=True, + ) + content_before = Column(Text, nullable=True) + title_before = Column(String, nullable=True) + folder_id_before = Column(Integer, nullable=True) + chunks_before = Column(JSONB, nullable=True) + metadata_before = Column("metadata_before", JSONB, nullable=True) + created_by_turn_id = Column(String(64), nullable=True, index=True) + agent_action_id = Column( + Integer, + ForeignKey("agent_action_log.id", ondelete="SET NULL"), + nullable=True, + index=True, + ) + created_at = Column( + TIMESTAMP(timezone=True), + nullable=False, + default=lambda: datetime.now(UTC), + server_default=text("(now() AT TIME ZONE 'utc')"), + index=True, + ) + + +class FolderRevision(BaseModel): + """Snapshot of a :class:`Folder` row taken before a mkdir / move.""" + + __tablename__ = "folder_revisions" + + folder_id = Column( + Integer, + ForeignKey("folders.id", ondelete="CASCADE"), + nullable=False, + index=True, + ) + search_space_id = Column( + Integer, + ForeignKey("searchspaces.id", ondelete="CASCADE"), + nullable=False, + index=True, + ) + name_before = Column(String(255), nullable=True) + parent_id_before = Column(Integer, nullable=True) + position_before = Column(String(50), nullable=True) + created_by_turn_id = Column(String(64), nullable=True, index=True) + agent_action_id = Column( + Integer, + ForeignKey("agent_action_log.id", ondelete="SET NULL"), + nullable=True, + index=True, + ) + created_at = Column( + TIMESTAMP(timezone=True), + nullable=False, + default=lambda: datetime.now(UTC), + server_default=text("(now() AT TIME ZONE 'utc')"), + index=True, + ) + + +class AgentPermissionRule(BaseModel): + """Persistent permission rule consumed by :class:`PermissionMiddleware`. + + Scoped at one of: search-space-wide (``user_id`` and ``thread_id`` NULL), + user-wide (``user_id`` set, ``thread_id`` NULL), or per-thread + (``thread_id`` set). Loaded at agent build time and converted to + :class:`Rule` instances inside the agent factory. + """ + + __tablename__ = "agent_permission_rules" + + search_space_id = Column( + Integer, + ForeignKey("searchspaces.id", ondelete="CASCADE"), + nullable=False, + index=True, + ) + user_id = Column( + UUID(as_uuid=True), + ForeignKey("user.id", ondelete="CASCADE"), + nullable=True, + index=True, + ) + thread_id = Column( + Integer, + ForeignKey("new_chat_threads.id", ondelete="CASCADE"), + nullable=True, + index=True, + ) + permission = Column(String(255), nullable=False) + pattern = Column(String(255), nullable=False, default="*", server_default="*") + action = Column(String(16), nullable=False) # allow / deny / ask + created_at = Column( + TIMESTAMP(timezone=True), + nullable=False, + default=lambda: datetime.now(UTC), + server_default=text("(now() AT TIME ZONE 'utc')"), + index=True, + ) + + __table_args__ = ( + UniqueConstraint( + "search_space_id", + "user_id", + "thread_id", + "permission", + "pattern", + "action", + name="uq_agent_permission_rules_scope", + ), + ) + + class RefreshToken(Base, TimestampMixin): """ Stores refresh tokens for user session management. diff --git a/surfsense_backend/app/observability/__init__.py b/surfsense_backend/app/observability/__init__.py new file mode 100644 index 000000000..dbf082561 --- /dev/null +++ b/surfsense_backend/app/observability/__init__.py @@ -0,0 +1,7 @@ +"""SurfSense observability surface. + +The single user-visible API right now is :mod:`otel`, which exposes a +small wrapper around the optional ``opentelemetry`` instrumentation. The +wrapper is a no-op when OTEL is not configured, so importing it from +performance-critical paths is safe. +""" diff --git a/surfsense_backend/app/observability/otel.py b/surfsense_backend/app/observability/otel.py new file mode 100644 index 000000000..0229524f2 --- /dev/null +++ b/surfsense_backend/app/observability/otel.py @@ -0,0 +1,319 @@ +""" +OpenTelemetry instrumentation helpers for the SurfSense agent stack. + +Tier 3b in the OpenCode-port plan. + +Goals +===== + +- Provide one tiny, ergonomic API for the spans listed in the plan + (``tool.call``, ``model.call``, ``kb.search``, ``kb.persist``, + ``compaction.run``, ``interrupt.raised``, ``permission.asked``). +- Keep span **names** low-cardinality (``tool.call`` rather than + ``tool.call.``); tool name lives in the ``tool.name`` attribute + so dashboards aggregate cleanly. +- Default to **no-op** behavior unless ``OTEL_EXPORTER_OTLP_ENDPOINT`` is + set, OR an external SDK has installed a real ``TracerProvider`` already + (e.g. via the ``opentelemetry-instrument`` agent). +- Coexist with LangSmith: we never disable LangSmith tracing; we add OTel + alongside. +- Gracefully degrade if the ``opentelemetry-api`` package is missing. +""" + +from __future__ import annotations + +import logging +import os +from collections.abc import Iterator +from contextlib import contextmanager +from typing import Any + +logger = logging.getLogger(__name__) + +# ----------------------------------------------------------------------------- +# Lazy/optional OpenTelemetry import +# ----------------------------------------------------------------------------- + +try: + from opentelemetry import trace as _ot_trace + from opentelemetry.trace import ( + Span as _OtSpan, + Status as _OtStatus, + StatusCode as _OtStatusCode, + ) + + _OTEL_AVAILABLE = True +except ImportError: # pragma: no cover — optional dep + _ot_trace = None # type: ignore[assignment] + _OtSpan = Any # type: ignore[assignment, misc] + _OtStatus = Any # type: ignore[assignment, misc] + _OtStatusCode = Any # type: ignore[assignment, misc] + _OTEL_AVAILABLE = False + + +_INSTRUMENTATION_NAME = "surfsense.new_chat" +_INSTRUMENTATION_VERSION = "0.1.0" + + +# ----------------------------------------------------------------------------- +# Configuration +# ----------------------------------------------------------------------------- + + +def _resolve_enabled() -> bool: + """Return True if OTel spans should actually be emitted.""" + if not _OTEL_AVAILABLE: + return False + # Honor an explicit kill-switch first. + if os.environ.get("SURFSENSE_DISABLE_OTEL", "").lower() in {"1", "true", "yes"}: + return False + # Treat a configured endpoint as the canonical "OTel is wired up" signal. + if os.environ.get("OTEL_EXPORTER_OTLP_ENDPOINT"): + return True + # Or honor an external SDK that already installed a non-default TracerProvider. + if _ot_trace is not None: + try: + provider = _ot_trace.get_tracer_provider() + # The default proxy provider has no real exporter wired up. + type_name = type(provider).__name__ + if type_name not in {"ProxyTracerProvider", "NoOpTracerProvider"}: + return True + except Exception: # pragma: no cover — defensive + return False + return False + + +_ENABLED: bool = _resolve_enabled() + + +def is_enabled() -> bool: + """Return True if instrumentation is actively emitting spans.""" + return _ENABLED + + +def _get_tracer(): + if not _OTEL_AVAILABLE: + return None + try: + return _ot_trace.get_tracer(_INSTRUMENTATION_NAME, _INSTRUMENTATION_VERSION) + except Exception: # pragma: no cover — defensive + return None + + +# ----------------------------------------------------------------------------- +# No-op span used when OTel is disabled (avoids a None check at every call site) +# ----------------------------------------------------------------------------- + + +class _NoopSpan: + """A lightweight stand-in that mimics the subset of ``Span`` we use.""" + + def set_attribute(self, key: str, value: Any) -> None: + return None + + def set_attributes(self, attributes: dict[str, Any]) -> None: + return None + + def add_event(self, name: str, attributes: dict[str, Any] | None = None) -> None: + return None + + def record_exception(self, exception: BaseException) -> None: + return None + + def set_status(self, status: Any) -> None: + return None + + +# ----------------------------------------------------------------------------- +# Public span helpers +# ----------------------------------------------------------------------------- + + +@contextmanager +def span( + name: str, + *, + attributes: dict[str, Any] | None = None, +) -> Iterator[Any]: + """Generic span context manager. + + Yields the underlying span (or a :class:`_NoopSpan` when disabled) + so callers can attach attributes/events incrementally. + + On exception, the span records the error via :meth:`record_exception` + and sets ``StatusCode.ERROR``; the exception is then re-raised. + """ + if not _ENABLED: + yield _NoopSpan() + return + + tracer = _get_tracer() + if tracer is None: # pragma: no cover — defensive + yield _NoopSpan() + return + + with tracer.start_as_current_span(name) as sp: + if attributes: + try: + sp.set_attributes(attributes) + except Exception: # pragma: no cover — defensive + pass + try: + yield sp + except BaseException as exc: + try: + sp.record_exception(exc) + sp.set_status(_OtStatus(_OtStatusCode.ERROR, str(exc))) + except Exception: # pragma: no cover — defensive + pass + raise + + +# ----------------------------------------------------------------------------- +# Domain-specific shortcuts (mirror the plan's enumerated span list) +# ----------------------------------------------------------------------------- + + +def tool_call_span( + tool_name: str, + *, + input_size: int | None = None, + extra: dict[str, Any] | None = None, +): + """Span for an individual tool execution. + + Span name is the constant ``tool.call`` (low-cardinality); the tool + identifier lives in the ``tool.name`` attribute. + """ + attrs: dict[str, Any] = {"tool.name": tool_name} + if input_size is not None: + attrs["tool.input.size"] = int(input_size) + if extra: + attrs.update(extra) + return span("tool.call", attributes=attrs) + + +def model_call_span( + *, + model_id: str | None = None, + provider: str | None = None, + extra: dict[str, Any] | None = None, +): + """Span around a single ``astream`` / ``ainvoke`` call to the LLM.""" + attrs: dict[str, Any] = {} + if model_id: + attrs["model.id"] = model_id + if provider: + attrs["model.provider"] = provider + if extra: + attrs.update(extra) + return span("model.call", attributes=attrs) + + +def kb_search_span( + *, + search_space_id: int | None = None, + query_chars: int | None = None, + extra: dict[str, Any] | None = None, +): + """Span around knowledge-base search routines.""" + attrs: dict[str, Any] = {} + if search_space_id is not None: + attrs["search_space.id"] = int(search_space_id) + if query_chars is not None: + attrs["query.chars"] = int(query_chars) + if extra: + attrs.update(extra) + return span("kb.search", attributes=attrs) + + +def kb_persist_span( + *, + document_type: str | None = None, + document_id: int | None = None, + extra: dict[str, Any] | None = None, +): + """Span around knowledge-base persistence operations (NOTE/EXTENSION/FILE).""" + attrs: dict[str, Any] = {} + if document_type: + attrs["document.type"] = document_type + if document_id is not None: + attrs["document.id"] = int(document_id) + if extra: + attrs.update(extra) + return span("kb.persist", attributes=attrs) + + +def compaction_span( + *, + reason: str | None = None, + messages_in: int | None = None, + extra: dict[str, Any] | None = None, +): + """Span around the compaction (summarization) middleware run.""" + attrs: dict[str, Any] = {} + if reason: + attrs["compaction.reason"] = reason + if messages_in is not None: + attrs["compaction.messages.in"] = int(messages_in) + if extra: + attrs.update(extra) + return span("compaction.run", attributes=attrs) + + +def interrupt_span( + *, + interrupt_type: str, + extra: dict[str, Any] | None = None, +): + """Span recording an interrupt being raised (HITL or permission_ask).""" + attrs: dict[str, Any] = {"interrupt.type": interrupt_type} + if extra: + attrs.update(extra) + return span("interrupt.raised", attributes=attrs) + + +def permission_asked_span( + *, + permission: str, + pattern: str | None = None, + extra: dict[str, Any] | None = None, +): + """Span recording a permission ask (PermissionMiddleware).""" + attrs: dict[str, Any] = {"permission.permission": permission} + if pattern: + attrs["permission.pattern"] = pattern + if extra: + attrs.update(extra) + return span("permission.asked", attributes=attrs) + + +# ----------------------------------------------------------------------------- +# Test/utility hooks +# ----------------------------------------------------------------------------- + + +def reload_for_tests() -> bool: + """Re-evaluate :data:`_ENABLED` from the current environment. + + Tests that toggle ``OTEL_EXPORTER_OTLP_ENDPOINT`` or + ``SURFSENSE_DISABLE_OTEL`` can call this to reset cached state. + Returns the new value of :func:`is_enabled`. + """ + global _ENABLED + _ENABLED = _resolve_enabled() + return _ENABLED + + +__all__ = [ + "compaction_span", + "interrupt_span", + "is_enabled", + "kb_persist_span", + "kb_search_span", + "model_call_span", + "permission_asked_span", + "reload_for_tests", + "span", + "tool_call_span", +] diff --git a/surfsense_backend/app/routes/__init__.py b/surfsense_backend/app/routes/__init__.py index de4e05423..a6a95ad30 100644 --- a/surfsense_backend/app/routes/__init__.py +++ b/surfsense_backend/app/routes/__init__.py @@ -1,5 +1,9 @@ from fastapi import APIRouter +from .agent_action_log_route import router as agent_action_log_router +from .agent_flags_route import router as agent_flags_router +from .agent_permissions_route import router as agent_permissions_router +from .agent_revert_route import router as agent_revert_router from .airtable_add_connector_route import ( router as airtable_add_connector_router, ) @@ -66,6 +70,12 @@ router.include_router(documents_router) router.include_router(folders_router) router.include_router(notes_router) router.include_router(new_chat_router) # Chat with assistant-ui persistence +router.include_router(agent_revert_router) # POST /threads/{id}/revert/{action_id} +router.include_router(agent_action_log_router) # GET /threads/{id}/actions +router.include_router( + agent_permissions_router +) # CRUD for /searchspaces/{id}/agent/permissions/rules +router.include_router(agent_flags_router) # GET /agent/flags router.include_router(sandbox_router) # Sandbox file downloads (Daytona) router.include_router(chat_comments_router) router.include_router(podcasts_router) # Podcast task status and audio diff --git a/surfsense_backend/app/routes/agent_action_log_route.py b/surfsense_backend/app/routes/agent_action_log_route.py new file mode 100644 index 000000000..458635761 --- /dev/null +++ b/surfsense_backend/app/routes/agent_action_log_route.py @@ -0,0 +1,186 @@ +"""``GET /api/threads/{thread_id}/actions``: list agent action-log entries. + +Pairs with ``POST /api/threads/{thread_id}/revert/{action_id}`` (see +``agent_revert_route.py``). The action log is the read-side surface for +the audit/undo UI: it returns a paginated list of every tool call +recorded by :class:`ActionLogMiddleware` against the thread, plus +metadata about whether the action is reversible and whether it has +already been reverted. + +The route is gated by the same ``SURFSENSE_ENABLE_ACTION_LOG`` flag that +controls the middleware. When the flag is off the endpoint returns 503 +so the UI can detect "this deployment doesn't have the action log +enabled" without 404-ing on a missing route. + +The list is ordered DESC by ``created_at`` (newest first) so the +revert UI can render a familiar reverse-chronological feed without an +additional client-side sort. +""" + +from __future__ import annotations + +import logging +from datetime import datetime +from typing import Any + +from fastapi import APIRouter, Depends, HTTPException, Query +from pydantic import BaseModel +from sqlalchemy import func, select +from sqlalchemy.ext.asyncio import AsyncSession + +from app.agents.new_chat.feature_flags import get_flags +from app.db import ( + AgentActionLog, + NewChatThread, + Permission, + User, + get_async_session, +) +from app.users import current_active_user +from app.utils.rbac import check_permission + +logger = logging.getLogger(__name__) + +router = APIRouter() + + +# --------------------------------------------------------------------------- +# Response schemas +# --------------------------------------------------------------------------- + + +class AgentActionRead(BaseModel): + """One row of the action log surfaced to the client.""" + + id: int + thread_id: int + user_id: str | None + search_space_id: int + tool_name: str + args: dict[str, Any] | None + result_id: str | None + reversible: bool + reverse_descriptor: dict[str, Any] | None + error: dict[str, Any] | None + reverse_of: int | None + reverted_by_action_id: int | None + is_revert_action: bool + created_at: datetime + + +class AgentActionListResponse(BaseModel): + """Paginated list response for the action log.""" + + items: list[AgentActionRead] + total: int + page: int + page_size: int + has_more: bool + + +# --------------------------------------------------------------------------- +# Routes +# --------------------------------------------------------------------------- + + +def _flag_guard() -> None: + flags = get_flags() + if flags.disable_new_agent_stack or not flags.enable_action_log: + raise HTTPException( + status_code=503, + detail=( + "Action log is not available on this deployment. Flip " + "SURFSENSE_ENABLE_ACTION_LOG to enable it." + ), + ) + + +@router.get( + "/threads/{thread_id}/actions", + response_model=AgentActionListResponse, +) +async def list_thread_actions( + thread_id: int, + page: int = Query(0, ge=0), + page_size: int = Query(50, ge=1, le=200), + session: AsyncSession = Depends(get_async_session), + user: User = Depends(current_active_user), +) -> AgentActionListResponse: + """List agent actions for a thread, newest first. + + Authorization: + * Caller must be a member of the thread's search space with + ``CHATS_READ`` permission. + + Pagination: + * ``page`` is 0-indexed. + * ``page_size`` defaults to 50, max 200. + """ + + _flag_guard() + + thread = await session.get(NewChatThread, thread_id) + if thread is None: + raise HTTPException(status_code=404, detail="Thread not found.") + + await check_permission( + session, + user, + thread.search_space_id, + Permission.CHATS_READ.value, + "You don't have permission to view this thread's action log.", + ) + + total_stmt = select(func.count(AgentActionLog.id)).where( + AgentActionLog.thread_id == thread_id + ) + total = (await session.execute(total_stmt)).scalar_one() + + rows_stmt = ( + select(AgentActionLog) + .where(AgentActionLog.thread_id == thread_id) + .order_by(AgentActionLog.created_at.desc(), AgentActionLog.id.desc()) + .offset(page * page_size) + .limit(page_size) + ) + rows = (await session.execute(rows_stmt)).scalars().all() + + # Build a reverse_of -> revert_action_id map so the UI can render + # "Reverted" badges on actions that have already been undone. + if rows: + original_ids = [r.id for r in rows] + reverts_stmt = select(AgentActionLog.id, AgentActionLog.reverse_of).where( + AgentActionLog.reverse_of.in_(original_ids) + ) + reverts = (await session.execute(reverts_stmt)).all() + revert_map: dict[int, int] = {orig: rev for rev, orig in reverts} + else: + revert_map = {} + + items = [ + AgentActionRead( + id=row.id, + thread_id=row.thread_id, + user_id=str(row.user_id) if row.user_id is not None else None, + search_space_id=row.search_space_id, + tool_name=row.tool_name, + args=row.args, + result_id=row.result_id, + reversible=bool(row.reversible), + reverse_descriptor=row.reverse_descriptor, + error=row.error, + reverse_of=row.reverse_of, + reverted_by_action_id=revert_map.get(row.id), + is_revert_action=row.reverse_of is not None, + created_at=row.created_at, + ) + for row in rows + ] + + return AgentActionListResponse( + items=items, + total=int(total), + page=page, + page_size=page_size, + has_more=(page + 1) * page_size < int(total), + ) diff --git a/surfsense_backend/app/routes/agent_flags_route.py b/surfsense_backend/app/routes/agent_flags_route.py new file mode 100644 index 000000000..d3c90a58d --- /dev/null +++ b/surfsense_backend/app/routes/agent_flags_route.py @@ -0,0 +1,71 @@ +"""``GET /api/agent/flags``: read-only feature-flag status. + +Surfaces :class:`AgentFeatureFlags` to the frontend so the UI can: + +* Render conditional surfaces (e.g. show the action-log button only when + ``enable_action_log`` is on). +* Display an admin diagnostics card so operators can verify which + middleware tier is active without shelling into the box. + +The endpoint is *read-only*. Flipping flags requires an env-var change +plus a process restart — by design, since the values are baked into the +agent factory at build time. The route does not require any special +permission (any authenticated user can see them) since the flag values +do not leak data, and the UI surfaces are conditionally rendered based +on them anyway. +""" + +from __future__ import annotations + +from dataclasses import asdict + +from fastapi import APIRouter, Depends +from pydantic import BaseModel + +from app.agents.new_chat.feature_flags import AgentFeatureFlags, get_flags +from app.db import User +from app.users import current_active_user + +router = APIRouter() + + +class AgentFeatureFlagsRead(BaseModel): + """Mirror of :class:`AgentFeatureFlags`. Updated together with it.""" + + disable_new_agent_stack: bool + + enable_context_editing: bool + enable_compaction_v2: bool + enable_retry_after: bool + enable_model_fallback: bool + enable_model_call_limit: bool + enable_tool_call_limit: bool + enable_tool_call_repair: bool + enable_doom_loop: bool + + enable_permission: bool + enable_busy_mutex: bool + enable_llm_tool_selector: bool + + enable_skills: bool + enable_specialized_subagents: bool + enable_kb_planner_runnable: bool + + enable_action_log: bool + enable_revert_route: bool + + enable_plugin_loader: bool + + enable_otel: bool + + @classmethod + def from_flags(cls, flags: AgentFeatureFlags) -> "AgentFeatureFlagsRead": + # asdict() avoids missing-field bugs when AgentFeatureFlags grows. + return cls(**asdict(flags)) + + +@router.get("/agent/flags", response_model=AgentFeatureFlagsRead) +async def get_agent_flags( + _user: User = Depends(current_active_user), +) -> AgentFeatureFlagsRead: + return AgentFeatureFlagsRead.from_flags(get_flags()) diff --git a/surfsense_backend/app/routes/agent_permissions_route.py b/surfsense_backend/app/routes/agent_permissions_route.py new file mode 100644 index 000000000..e87af29c7 --- /dev/null +++ b/surfsense_backend/app/routes/agent_permissions_route.py @@ -0,0 +1,280 @@ +"""CRUD for :class:`app.db.AgentPermissionRule`. + +Surfaces the permission rules consumed by +:class:`PermissionMiddleware`. Rules are scoped at one of three levels: + +* **Search-space wide** — both ``user_id`` and ``thread_id`` are NULL. +* **Per-user** — ``user_id`` set, ``thread_id`` NULL. +* **Per-thread** — ``thread_id`` set (``user_id`` typically NULL). + +The middleware reads these rows at agent build time (see +``chat_deepagent.py``). UI lets a search-space owner curate them so +the agent can ask for approval / auto-deny / auto-allow specific +tool patterns. + +The route group is gated by ``SURFSENSE_ENABLE_PERMISSION``: when off +all endpoints return 503 so the UI can render a "feature not enabled" +empty state without breaking on a missing route. +""" + +from __future__ import annotations + +import logging +import re +from datetime import datetime +from typing import Literal + +from fastapi import APIRouter, Depends, HTTPException +from pydantic import BaseModel, Field +from sqlalchemy import select +from sqlalchemy.exc import IntegrityError +from sqlalchemy.ext.asyncio import AsyncSession + +from app.agents.new_chat.feature_flags import get_flags +from app.db import ( + AgentPermissionRule, + NewChatThread, + Permission, + SearchSpace, + User, + get_async_session, +) +from app.users import current_active_user +from app.utils.rbac import check_permission + +logger = logging.getLogger(__name__) + +router = APIRouter() + + +# --------------------------------------------------------------------------- +# Schemas +# --------------------------------------------------------------------------- + + +_ACTION_VALUES: tuple[str, ...] = ("allow", "deny", "ask") +_PERMISSION_PATTERN = re.compile(r"^[a-zA-Z0-9_:.\-*]+$") + + +class AgentPermissionRuleRead(BaseModel): + id: int + search_space_id: int + user_id: str | None + thread_id: int | None + permission: str + pattern: str + action: Literal["allow", "deny", "ask"] + created_at: datetime + + +class AgentPermissionRuleCreate(BaseModel): + permission: str = Field( + ..., + min_length=1, + max_length=255, + description="Tool / capability the rule targets, e.g. 'tool:create_linear_issue'.", + ) + pattern: str = Field( + "*", + min_length=1, + max_length=255, + description="Wildcard pattern (e.g. '*' or 'production-*') applied to the matched tool argument.", + ) + action: Literal["allow", "deny", "ask"] + user_id: str | None = None + thread_id: int | None = None + + +class AgentPermissionRuleUpdate(BaseModel): + pattern: str | None = Field(default=None, min_length=1, max_length=255) + action: Literal["allow", "deny", "ask"] | None = None + + +# --------------------------------------------------------------------------- +# Helpers +# --------------------------------------------------------------------------- + + +def _flag_guard() -> None: + flags = get_flags() + if flags.disable_new_agent_stack or not flags.enable_permission: + raise HTTPException( + status_code=503, + detail=( + "Agent permission rules are not enabled on this deployment. " + "Flip SURFSENSE_ENABLE_PERMISSION to enable them." + ), + ) + + +def _validate_permission_string(value: str) -> str: + if not _PERMISSION_PATTERN.match(value): + raise HTTPException( + status_code=400, + detail=( + "permission must contain only alphanumerics, '.', '_', ':', '-', " + "or '*' wildcards." + ), + ) + return value + + +def _to_read(row: AgentPermissionRule) -> AgentPermissionRuleRead: + return AgentPermissionRuleRead( + id=row.id, + search_space_id=row.search_space_id, + user_id=str(row.user_id) if row.user_id is not None else None, + thread_id=row.thread_id, + permission=row.permission, + pattern=row.pattern, + action=row.action, # type: ignore[arg-type] + created_at=row.created_at, + ) + + +async def _ensure_search_space_membership_admin( + session: AsyncSession, user: User, search_space_id: int +) -> None: + """Curating agent rules == "settings" administration on the space.""" + space = await session.get(SearchSpace, search_space_id) + if space is None: + raise HTTPException(status_code=404, detail="Search space not found.") + await check_permission( + session, + user, + search_space_id, + Permission.SETTINGS_UPDATE.value, + "You don't have permission to manage agent permission rules in this space.", + ) + + +# --------------------------------------------------------------------------- +# Routes +# --------------------------------------------------------------------------- + + +@router.get( + "/searchspaces/{search_space_id}/agent/permissions/rules", + response_model=list[AgentPermissionRuleRead], +) +async def list_rules( + search_space_id: int, + session: AsyncSession = Depends(get_async_session), + user: User = Depends(current_active_user), +) -> list[AgentPermissionRuleRead]: + _flag_guard() + await _ensure_search_space_membership_admin(session, user, search_space_id) + + stmt = ( + select(AgentPermissionRule) + .where(AgentPermissionRule.search_space_id == search_space_id) + .order_by(AgentPermissionRule.created_at.desc(), AgentPermissionRule.id.desc()) + ) + rows = (await session.execute(stmt)).scalars().all() + return [_to_read(r) for r in rows] + + +@router.post( + "/searchspaces/{search_space_id}/agent/permissions/rules", + response_model=AgentPermissionRuleRead, + status_code=201, +) +async def create_rule( + search_space_id: int, + payload: AgentPermissionRuleCreate, + session: AsyncSession = Depends(get_async_session), + user: User = Depends(current_active_user), +) -> AgentPermissionRuleRead: + _flag_guard() + await _ensure_search_space_membership_admin(session, user, search_space_id) + + permission = _validate_permission_string(payload.permission.strip()) + pattern = payload.pattern.strip() or "*" + + if payload.thread_id is not None: + thread = await session.get(NewChatThread, payload.thread_id) + if thread is None or thread.search_space_id != search_space_id: + raise HTTPException( + status_code=404, + detail="Thread not found in this search space.", + ) + + row = AgentPermissionRule( + search_space_id=search_space_id, + user_id=payload.user_id, + thread_id=payload.thread_id, + permission=permission, + pattern=pattern, + action=payload.action, + ) + session.add(row) + try: + await session.commit() + except IntegrityError: + await session.rollback() + raise HTTPException( + status_code=409, + detail=( + "An identical rule already exists for this scope. Update the " + "existing rule instead." + ), + ) + await session.refresh(row) + return _to_read(row) + + +@router.patch( + "/searchspaces/{search_space_id}/agent/permissions/rules/{rule_id}", + response_model=AgentPermissionRuleRead, +) +async def update_rule( + search_space_id: int, + rule_id: int, + payload: AgentPermissionRuleUpdate, + session: AsyncSession = Depends(get_async_session), + user: User = Depends(current_active_user), +) -> AgentPermissionRuleRead: + _flag_guard() + await _ensure_search_space_membership_admin(session, user, search_space_id) + + row = await session.get(AgentPermissionRule, rule_id) + if row is None or row.search_space_id != search_space_id: + raise HTTPException(status_code=404, detail="Rule not found.") + + if payload.pattern is not None: + row.pattern = payload.pattern.strip() or "*" + if payload.action is not None: + row.action = payload.action + + try: + await session.commit() + except IntegrityError: + await session.rollback() + raise HTTPException( + status_code=409, + detail="Update would create a duplicate rule for this scope.", + ) + await session.refresh(row) + return _to_read(row) + + +@router.delete( + "/searchspaces/{search_space_id}/agent/permissions/rules/{rule_id}", + status_code=204, +) +async def delete_rule( + search_space_id: int, + rule_id: int, + session: AsyncSession = Depends(get_async_session), + user: User = Depends(current_active_user), +) -> None: + _flag_guard() + await _ensure_search_space_membership_admin(session, user, search_space_id) + + row = await session.get(AgentPermissionRule, rule_id) + if row is None or row.search_space_id != search_space_id: + raise HTTPException(status_code=404, detail="Rule not found.") + + await session.delete(row) + await session.commit() + return None diff --git a/surfsense_backend/app/routes/agent_revert_route.py b/surfsense_backend/app/routes/agent_revert_route.py new file mode 100644 index 000000000..2f6fe6a32 --- /dev/null +++ b/surfsense_backend/app/routes/agent_revert_route.py @@ -0,0 +1,122 @@ +"""POST ``/api/threads/{thread_id}/revert/{action_id}``: undo an agent action. + +Per the Tier 5 plan, the route ships **before** the UI lights up the per-message +"Undo from here" affordance. To prevent accidental usage during the gap we +return ``503 Service Unavailable`` until the +``SURFSENSE_ENABLE_REVERT_ROUTE`` flag flips. Once enabled, the route runs: + +1. Authentication via :func:`current_active_user`. +2. Action lookup; 404 if the action does not belong to the thread. +3. Authorization via :func:`app.services.revert_service.can_revert`. +4. Revert dispatch via :func:`app.services.revert_service.revert_action`. +5. Idempotent on retries: if the same action is reverted twice the second + call returns 409 ``"already reverted"``. +""" + +from __future__ import annotations + +import logging + +from fastapi import APIRouter, Depends, HTTPException +from sqlalchemy import select +from sqlalchemy.ext.asyncio import AsyncSession + +from app.agents.new_chat.feature_flags import get_flags +from app.db import ( + AgentActionLog, + User, + get_async_session, +) +from app.services.revert_service import ( + RevertOutcome, + can_revert, + load_action, + load_thread, + revert_action, +) +from app.users import current_active_user + +logger = logging.getLogger(__name__) + +router = APIRouter() + + +@router.post("/threads/{thread_id}/revert/{action_id}") +async def revert_agent_action( + thread_id: int, + action_id: int, + session: AsyncSession = Depends(get_async_session), + user: User = Depends(current_active_user), +) -> dict: + flags = get_flags() + if flags.disable_new_agent_stack or not flags.enable_revert_route: + raise HTTPException( + status_code=503, + detail=( + "Revert is not available on this deployment yet. The route " + "ships before the UI; flip SURFSENSE_ENABLE_REVERT_ROUTE to " + "enable it." + ), + ) + + thread = await load_thread(session, thread_id=thread_id) + if thread is None: + raise HTTPException(status_code=404, detail="Thread not found.") + + action = await load_action(session, action_id=action_id, thread_id=thread_id) + if action is None: + raise HTTPException( + status_code=404, + detail="Action not found or does not belong to this thread.", + ) + + # Idempotency: if a successful revert already exists, return 409. + existing_revert = await session.execute( + select(AgentActionLog).where(AgentActionLog.reverse_of == action.id) + ) + if existing_revert.scalars().first() is not None: + raise HTTPException( + status_code=409, + detail="This action has already been reverted.", + ) + + if not can_revert( + requester_user_id=str(user.id) if user is not None else None, + action=action, + is_admin=False, # role lookup is done by RBAC layer; default conservative + ): + raise HTTPException( + status_code=403, + detail="You are not allowed to revert this action.", + ) + + outcome: RevertOutcome + try: + outcome = await revert_action( + session, + action=action, + requester_user_id=str(user.id) if user is not None else None, + ) + except Exception: + logger.exception("Revert dispatch raised for action_id=%s", action_id) + await session.rollback() + raise HTTPException(status_code=500, detail="Internal error during revert.") + + if outcome.status == "ok": + await session.commit() + return { + "status": "ok", + "message": outcome.message, + "new_action_id": outcome.new_action_id, + } + + await session.rollback() + + if outcome.status == "not_found" or outcome.status == "tool_unavailable": + raise HTTPException(status_code=409, detail=outcome.message) + if outcome.status == "permission_denied": + raise HTTPException(status_code=403, detail=outcome.message) + if outcome.status == "reverse_not_implemented": + raise HTTPException(status_code=501, detail=outcome.message) + # not_reversible + raise HTTPException(status_code=409, detail=outcome.message) diff --git a/surfsense_backend/app/services/revert_service.py b/surfsense_backend/app/services/revert_service.py new file mode 100644 index 000000000..e072f90c6 --- /dev/null +++ b/surfsense_backend/app/services/revert_service.py @@ -0,0 +1,279 @@ +"""Revert service for the SurfSense agent action log. + +Implements the actual revert workflow used by +``POST /api/threads/{thread_id}/revert/{action_id}``. The route handler is a +thin auth + flag wrapper around the functions defined here. + +Operation outcomes mirror the plan: + +* **KB-owned actions** (NOTE / FILE / FOLDER mutations): restore from + :class:`app.db.DocumentRevision` / :class:`app.db.FolderRevision` rows + written before the original mutation. +* **Connector-owned actions with a declared ``reverse_descriptor``**: invoke + the inverse tool through the agent's normal permission stack (NOT + bypassed). Out of scope for this PR — returns ``REVERSE_NOT_IMPLEMENTED``. +* **Anything else** (deprecated tool / no descriptor / schema drift): + returns ``NOT_REVERSIBLE`` and the route surfaces it as 409. + +A successful revert appends a NEW row to ``agent_action_log`` with +``reverse_of=`` and the requesting user's +``user_id``, preserving an auditable chain. +""" + +from __future__ import annotations + +import logging +from dataclasses import dataclass +from datetime import UTC, datetime +from typing import Literal + +from sqlalchemy import select +from sqlalchemy.ext.asyncio import AsyncSession + +from app.db import ( + AgentActionLog, + DocumentRevision, + FolderRevision, + NewChatThread, +) + +logger = logging.getLogger(__name__) + + +RevertOutcomeStatus = Literal[ + "ok", + "not_reversible", + "not_found", + "permission_denied", + "tool_unavailable", + "reverse_not_implemented", +] + + +@dataclass +class RevertOutcome: + """Structured result of :func:`revert_action`.""" + + status: RevertOutcomeStatus + message: str + new_action_id: int | None = None + + +# --------------------------------------------------------------------------- +# Lookup helpers +# --------------------------------------------------------------------------- + + +async def load_action( + session: AsyncSession, + *, + action_id: int, + thread_id: int, +) -> AgentActionLog | None: + """Load the action_log row for ``action_id`` if it belongs to the thread.""" + stmt = select(AgentActionLog).where( + AgentActionLog.id == action_id, + AgentActionLog.thread_id == thread_id, + ) + result = await session.execute(stmt) + return result.scalars().first() + + +async def load_thread( + session: AsyncSession, *, thread_id: int +) -> NewChatThread | None: + stmt = select(NewChatThread).where(NewChatThread.id == thread_id) + result = await session.execute(stmt) + return result.scalars().first() + + +# --------------------------------------------------------------------------- +# Authorization +# --------------------------------------------------------------------------- + + +def can_revert( + *, + requester_user_id: str | None, + action: AgentActionLog, + is_admin: bool, +) -> bool: + """Return True iff the requester is allowed to revert this action. + + The plan's rule: "requester must be the original `user_id` on the + action, or hold the search-space admin role." Anonymous actions + (``action.user_id is None``) can only be reverted by admins. + """ + if is_admin: + return True + if action.user_id is None: + return False + return str(action.user_id) == str(requester_user_id) + + +# --------------------------------------------------------------------------- +# Revert paths +# --------------------------------------------------------------------------- + + +async def _restore_document_revision( + session: AsyncSession, *, action: AgentActionLog +) -> RevertOutcome: + """Restore the most recent :class:`DocumentRevision` for ``action``.""" + stmt = ( + select(DocumentRevision) + .where(DocumentRevision.agent_action_id == action.id) + .order_by(DocumentRevision.created_at.desc()) + .limit(1) + ) + result = await session.execute(stmt) + revision = result.scalars().first() + if revision is None: + return RevertOutcome( + status="not_reversible", + message="No document_revisions row tied to this action.", + ) + + from app.db import Document # late import to avoid cycles at module load + + doc = await session.get(Document, revision.document_id) + if doc is None: + return RevertOutcome( + status="tool_unavailable", + message="Original document has been deleted; revert cannot proceed.", + ) + + if revision.content_before is not None: + doc.content = revision.content_before + if revision.title_before is not None: + doc.title = revision.title_before + if revision.folder_id_before is not None: + doc.folder_id = revision.folder_id_before + doc.updated_at = datetime.now(UTC) + return RevertOutcome(status="ok", message="Document restored from snapshot.") + + +async def _restore_folder_revision( + session: AsyncSession, *, action: AgentActionLog +) -> RevertOutcome: + stmt = ( + select(FolderRevision) + .where(FolderRevision.agent_action_id == action.id) + .order_by(FolderRevision.created_at.desc()) + .limit(1) + ) + result = await session.execute(stmt) + revision = result.scalars().first() + if revision is None: + return RevertOutcome( + status="not_reversible", + message="No folder_revisions row tied to this action.", + ) + + from app.db import Folder + + folder = await session.get(Folder, revision.folder_id) + if folder is None: + return RevertOutcome( + status="tool_unavailable", + message="Original folder has been deleted; revert cannot proceed.", + ) + + if revision.name_before is not None: + folder.name = revision.name_before + if revision.parent_id_before is not None: + folder.parent_id = revision.parent_id_before + if revision.position_before is not None: + folder.position = revision.position_before + folder.updated_at = datetime.now(UTC) + return RevertOutcome(status="ok", message="Folder restored from snapshot.") + + +# Tool-name prefixes that route to KB document / folder revert paths. Kept +# as data so a future PR adding new KB-owned tools doesn't have to touch +# this module's control flow. +_DOC_TOOL_PREFIXES: tuple[str, ...] = ( + "edit_file", + "write_file", + "update_memory", + "create_note", + "update_note", + "delete_note", +) +_FOLDER_TOOL_PREFIXES: tuple[str, ...] = ( + "mkdir", + "move_file", + "rename_folder", + "delete_folder", +) + + +async def revert_action( + session: AsyncSession, + *, + action: AgentActionLog, + requester_user_id: str | None, +) -> RevertOutcome: + """Execute the revert for ``action`` and return a structured outcome. + + The function does **not** commit — the caller is expected to commit on + success or roll back on failure. A new ``agent_action_log`` row is + added to the session on success with ``reverse_of=action.id``. + """ + tool_name = (action.tool_name or "").lower() + + if tool_name.startswith(_DOC_TOOL_PREFIXES): + outcome = await _restore_document_revision(session, action=action) + elif tool_name.startswith(_FOLDER_TOOL_PREFIXES): + outcome = await _restore_folder_revision(session, action=action) + elif action.reverse_descriptor: + # Connector-owned reversibles run through the normal permission + # stack; out of scope for this PR — the route returns 503 anyway + # until UI ships, so 501-style "not implemented" is fine. + return RevertOutcome( + status="reverse_not_implemented", + message=( + "Connector-action revert is not yet implemented. The " + "reverse_descriptor is stored; future work will replay it " + "through PermissionMiddleware." + ), + ) + else: + return RevertOutcome( + status="not_reversible", + message=( + f"Tool {action.tool_name!r} is not reversible: no document " + "revision and no reverse_descriptor." + ), + ) + + if outcome.status != "ok": + return outcome + + new_row = AgentActionLog( + thread_id=action.thread_id, + user_id=requester_user_id, + search_space_id=action.search_space_id, + turn_id=None, + message_id=None, + tool_name=f"_revert:{action.tool_name}", + args={"reverted_action_id": action.id}, + result_id=None, + reversible=False, + reverse_descriptor=None, + error=None, + reverse_of=action.id, + ) + session.add(new_row) + await session.flush() + outcome.new_action_id = new_row.id + return outcome + + +__all__ = [ + "RevertOutcome", + "can_revert", + "load_action", + "load_thread", + "revert_action", +] diff --git a/surfsense_backend/app/utils/async_retry.py b/surfsense_backend/app/utils/async_retry.py index a56f6550a..607b7a156 100644 --- a/surfsense_backend/app/utils/async_retry.py +++ b/surfsense_backend/app/utils/async_retry.py @@ -33,7 +33,7 @@ F = TypeVar("F", bound=Callable) def _is_retryable(exc: BaseException) -> bool: if isinstance(exc, ConnectorError): return exc.retryable - return bool(isinstance(exc, (httpx.TimeoutException, httpx.ConnectError))) + return bool(isinstance(exc, httpx.TimeoutException | httpx.ConnectError)) def build_retry( diff --git a/surfsense_backend/tests/integration/harness/__init__.py b/surfsense_backend/tests/integration/harness/__init__.py new file mode 100644 index 000000000..9a7ec07dc --- /dev/null +++ b/surfsense_backend/tests/integration/harness/__init__.py @@ -0,0 +1,146 @@ +""" +Integration test harness for the SurfSense agent stack. + +The plan calls for an ``LLMToolEmulator``-backed harness for end-to-end +replay of ``stream_new_chat``. The currently-installed langchain version +does not expose ``LLMToolEmulator``, so this harness builds the equivalent +on top of :class:`langchain_core.language_models.fake_chat_models.FakeMessagesListChatModel`. + +The harness lets a test author script a sequence of model responses +(text + optional tool calls) and replay them against the new_chat agent +graph. Tools are stubbed via ``StubToolSpec`` -> ``langchain_core.tools.tool`` +decorator and execute deterministic Python callbacks. + +Used by: +- ``tests/integration/agents/new_chat/test_feature_flag_smoke.py`` to + confirm the kill-switch path produces identical-shape output regardless + of which middleware flags are toggled. +- Future per-tier PRs to record golden transcripts. +""" + +from __future__ import annotations + +import uuid +from collections.abc import Callable, Sequence +from dataclasses import dataclass, field +from typing import Any + +from langchain_core.language_models import LanguageModelInput +from langchain_core.language_models.fake_chat_models import ( + FakeMessagesListChatModel, +) +from langchain_core.messages import AIMessage, BaseMessage +from langchain_core.runnables import Runnable +from langchain_core.tools import BaseTool, tool + + +class _ToolBindingFakeChatModel(FakeMessagesListChatModel): + """Adapter so the harness model can pretend it understands ``bind_tools``. + + The base ``FakeMessagesListChatModel`` raises ``NotImplementedError`` from + ``bind_tools``, but ``langchain.agents.create_agent`` always calls + ``bind_tools`` to attach the tool registry. We don't actually need the + fake to honor the tool schema — it's already scripted to emit the right + tool calls — so we return self. + """ + + def bind_tools( # type: ignore[override] + self, + tools: Sequence[Any], + *, + tool_choice: Any = None, + **kwargs: Any, + ) -> Runnable[LanguageModelInput, AIMessage]: + return self + + +@dataclass +class StubToolSpec: + """A test-mode tool: a name, description, and a deterministic body.""" + + name: str + description: str + handler: Callable[..., Any] + args_schema: dict[str, Any] | None = None + + def build(self) -> BaseTool: + """Realize as a `langchain_core.tools.BaseTool`.""" + + @tool(name_or_callable=self.name, description=self.description) + def _stub_tool(**kwargs: Any) -> Any: + return self.handler(**kwargs) + + return _stub_tool + + +@dataclass +class ScriptedTurn: + """One scripted assistant turn. + + `text` is the assistant text (may be empty if pure tool call). + `tool_calls` is a list of dicts ``{name, args, id}``; if non-empty, the + agent will route to those tools and append a follow-up turn. + """ + + text: str = "" + tool_calls: list[dict[str, Any]] = field(default_factory=list) + + +def build_scripted_messages(turns: list[ScriptedTurn]) -> list[BaseMessage]: + """Convert :class:`ScriptedTurn` records to AIMessage payloads.""" + out: list[BaseMessage] = [] + for turn in turns: + tool_calls: list[dict[str, Any]] = [] + for tc in turn.tool_calls: + tool_calls.append( + { + "name": tc["name"], + "args": tc.get("args", {}), + "id": tc.get("id") or f"call_{uuid.uuid4().hex[:8]}", + } + ) + out.append(AIMessage(content=turn.text, tool_calls=tool_calls or [])) + return out + + +@dataclass +class ScriptedHarness: + """Bundle of (model, tools) ready to plug into ``create_agent``.""" + + model: _ToolBindingFakeChatModel + tools: list[BaseTool] + + +def build_scripted_harness( + *, + turns: list[ScriptedTurn], + tools: list[StubToolSpec] | None = None, + sleep: float | None = None, +) -> ScriptedHarness: + """Construct a deterministic agent harness from a script. + + Example:: + + harness = build_scripted_harness( + turns=[ + ScriptedTurn(tool_calls=[{"name": "echo", "args": {"x": 1}}]), + ScriptedTurn(text="done"), + ], + tools=[ + StubToolSpec(name="echo", description="echo args", handler=lambda **kw: kw), + ], + ) + """ + messages = build_scripted_messages(turns) + model = _ToolBindingFakeChatModel(responses=messages, sleep=sleep) + realized_tools = [t.build() for t in (tools or [])] + return ScriptedHarness(model=model, tools=realized_tools) + + +__all__ = [ + "ScriptedHarness", + "ScriptedTurn", + "StubToolSpec", + "build_scripted_harness", + "build_scripted_messages", +] diff --git a/surfsense_backend/tests/integration/harness/test_scripted_harness.py b/surfsense_backend/tests/integration/harness/test_scripted_harness.py new file mode 100644 index 000000000..6e9f7ab91 --- /dev/null +++ b/surfsense_backend/tests/integration/harness/test_scripted_harness.py @@ -0,0 +1,53 @@ +"""Smoke test: scripted harness drives create_agent end-to-end and produces a tool-call-then-final-text trace.""" + +from __future__ import annotations + +import pytest +from langchain.agents import create_agent + +from tests.integration.harness import ( + ScriptedTurn, + StubToolSpec, + build_scripted_harness, +) + +pytestmark = pytest.mark.integration + + +@pytest.mark.asyncio +async def test_scripted_harness_drives_basic_agent() -> None: + harness = build_scripted_harness( + turns=[ + ScriptedTurn( + tool_calls=[ + {"name": "echo", "args": {"x": 1}, "id": "call_1"}, + ] + ), + ScriptedTurn(text="done"), + ], + tools=[ + StubToolSpec( + name="echo", + description="Echo args back.", + handler=lambda **kwargs: {"echoed": kwargs}, + ), + ], + ) + + agent = create_agent( + harness.model, + system_prompt="You are a test agent.", + tools=harness.tools, + ) + + result = await agent.ainvoke({"messages": [("user", "do the thing")]}) + messages = result["messages"] + final_ai = next( + (m for m in reversed(messages) if m.__class__.__name__ == "AIMessage"), + None, + ) + assert final_ai is not None + assert final_ai.content == "done" + tool_messages = [m for m in messages if m.__class__.__name__ == "ToolMessage"] + assert len(tool_messages) == 1 + assert "echoed" in str(tool_messages[0].content) diff --git a/surfsense_backend/tests/unit/agents/__init__.py b/surfsense_backend/tests/unit/agents/__init__.py new file mode 100644 index 000000000..8b1378917 --- /dev/null +++ b/surfsense_backend/tests/unit/agents/__init__.py @@ -0,0 +1 @@ + diff --git a/surfsense_backend/tests/unit/agents/new_chat/__init__.py b/surfsense_backend/tests/unit/agents/new_chat/__init__.py new file mode 100644 index 000000000..8b1378917 --- /dev/null +++ b/surfsense_backend/tests/unit/agents/new_chat/__init__.py @@ -0,0 +1 @@ + diff --git a/surfsense_backend/tests/unit/agents/new_chat/prompts/__init__.py b/surfsense_backend/tests/unit/agents/new_chat/prompts/__init__.py new file mode 100644 index 000000000..a92d371bd --- /dev/null +++ b/surfsense_backend/tests/unit/agents/new_chat/prompts/__init__.py @@ -0,0 +1 @@ +"""__init__ stub so pytest discovers the prompts test module.""" diff --git a/surfsense_backend/tests/unit/agents/new_chat/prompts/test_composer.py b/surfsense_backend/tests/unit/agents/new_chat/prompts/test_composer.py new file mode 100644 index 000000000..d35b7aa8b --- /dev/null +++ b/surfsense_backend/tests/unit/agents/new_chat/prompts/test_composer.py @@ -0,0 +1,201 @@ +"""Tests for the prompt fragment composer (Tier 3a).""" + +from __future__ import annotations + +from datetime import UTC, datetime + +import pytest + +from app.agents.new_chat.prompts.composer import ( + ALL_TOOL_NAMES_ORDERED, + compose_system_prompt, + detect_provider_variant, +) +from app.db import ChatVisibility + +pytestmark = pytest.mark.unit + + +@pytest.fixture +def fixed_today() -> datetime: + return datetime(2025, 6, 1, 12, 0, tzinfo=UTC) + + +class TestProviderVariantDetection: + @pytest.mark.parametrize( + "model_name,expected", + [ + ("openai:gpt-4o-mini", "openai_classic"), + ("openai:gpt-4-turbo", "openai_classic"), + ("openai:gpt-5", "openai_reasoning"), + ("openai:gpt-5-codex", "openai_reasoning"), + ("openai:o1-preview", "openai_reasoning"), + ("openai:o3-mini", "openai_reasoning"), + ("anthropic:claude-3-5-sonnet", "anthropic"), + ("anthropic/claude-opus-4", "anthropic"), + ("google:gemini-2.0-flash", "google"), + ("vertex:gemini-1.5-pro", "google"), + ("groq:mixtral-8x7b", "default"), + (None, "default"), + ("", "default"), + ], + ) + def test_detection(self, model_name: str | None, expected: str) -> None: + assert detect_provider_variant(model_name) == expected + + +class TestCompose: + def test_default_prompt_has_required_blocks(self, fixed_today: datetime) -> None: + prompt = compose_system_prompt(today=fixed_today) + # System instruction wrapper + assert "" in prompt + assert "" in prompt + # Date interpolated + assert "2025-06-01" in prompt + # Core policy blocks present + assert "" in prompt + assert "" in prompt + assert "" in prompt + assert "" in prompt + # Tools + assert "" in prompt + assert "" in prompt + # Citations on by default + assert "" in prompt + assert "[citation:chunk_id]" in prompt + + def test_team_visibility_uses_team_variants( + self, fixed_today: datetime + ) -> None: + prompt = compose_system_prompt( + today=fixed_today, + thread_visibility=ChatVisibility.SEARCH_SPACE, + ) + # Team-specific phrasing in the agent block + assert "team space" in prompt + # Memory protocol mentions team + assert "team" in prompt + # Should NOT mention the user-only memory phrasing + assert "personal knowledge base" not in prompt + + def test_private_visibility_uses_private_variants( + self, fixed_today: datetime + ) -> None: + prompt = compose_system_prompt( + today=fixed_today, + thread_visibility=ChatVisibility.PRIVATE, + ) + assert "personal knowledge base" in prompt + # Should NOT mention the team-specific phrasing about prefixed authors + assert "[DisplayName of the author]" not in prompt + + def test_citations_disabled_swaps_block(self, fixed_today: datetime) -> None: + prompt_on = compose_system_prompt(today=fixed_today, citations_enabled=True) + prompt_off = compose_system_prompt(today=fixed_today, citations_enabled=False) + assert "Citations are DISABLED" in prompt_off + assert "Citations are DISABLED" not in prompt_on + assert "[citation:chunk_id]" in prompt_on + + def test_enabled_tool_filter_only_includes_listed_tools( + self, fixed_today: datetime + ) -> None: + prompt = compose_system_prompt( + today=fixed_today, + enabled_tool_names={"web_search", "scrape_webpage"}, + ) + assert "web_search:" in prompt or "- web_search:" in prompt + assert "scrape_webpage:" in prompt or "- scrape_webpage:" in prompt + # Excluded tools should NOT appear in tool listing + assert "generate_podcast:" not in prompt + assert "generate_image:" not in prompt + + def test_disabled_tool_note_is_appended(self, fixed_today: datetime) -> None: + prompt = compose_system_prompt( + today=fixed_today, + enabled_tool_names={"web_search"}, + disabled_tool_names={"generate_image", "generate_podcast"}, + ) + assert "DISABLED TOOLS (by user):" in prompt + assert "Generate Image" in prompt + assert "Generate Podcast" in prompt + + def test_mcp_routing_block_emits_when_provided( + self, fixed_today: datetime + ) -> None: + prompt = compose_system_prompt( + today=fixed_today, + mcp_connector_tools={"My GitLab": ["gitlab_search", "gitlab_create_mr"]}, + ) + assert "" in prompt + assert "My GitLab" in prompt + assert "gitlab_search" in prompt + + def test_mcp_routing_block_absent_when_no_servers( + self, fixed_today: datetime + ) -> None: + prompt = compose_system_prompt(today=fixed_today, mcp_connector_tools={}) + assert "" not in prompt + + def test_provider_block_renders_when_anthropic( + self, fixed_today: datetime + ) -> None: + prompt = compose_system_prompt( + today=fixed_today, model_name="anthropic:claude-3-5-sonnet" + ) + assert "" in prompt + assert "Anthropic" in prompt or "Claude" in prompt + + def test_provider_block_absent_for_default(self, fixed_today: datetime) -> None: + prompt = compose_system_prompt(today=fixed_today, model_name="custom:foo") + assert "" not in prompt + + def test_custom_system_instructions_override_default( + self, fixed_today: datetime + ) -> None: + custom = "You are a custom assistant. Today is {resolved_today}." + prompt = compose_system_prompt( + today=fixed_today, custom_system_instructions=custom + ) + assert "You are a custom assistant. Today is 2025-06-01." in prompt + # Default block should NOT be present + assert "" not in prompt + + def test_use_default_false_with_no_custom_yields_no_system_block( + self, fixed_today: datetime + ) -> None: + prompt = compose_system_prompt( + today=fixed_today, + use_default_system_instructions=False, + ) + # No system_instruction wrapper but tools/citations still emitted + assert "" not in prompt + assert "" in prompt + + def test_all_known_tools_have_fragments(self) -> None: + # Soft assertion: verify that every tool in the canonical order + # produces non-empty content for at least one variant. + for tool in ALL_TOOL_NAMES_ORDERED: + prompt = compose_system_prompt( + today=datetime(2025, 1, 1, tzinfo=UTC), + enabled_tool_names={tool}, + ) + assert tool in prompt, f"tool {tool!r} missing from composed prompt" + + +class TestStableOrderingForCacheStability: + """Regression guard: prompt cache hit-rate depends on byte-stable prefix.""" + + def test_composition_is_deterministic_given_same_inputs( + self, fixed_today: datetime + ) -> None: + a = compose_system_prompt( + today=fixed_today, + enabled_tool_names={"web_search", "scrape_webpage"}, + mcp_connector_tools={"X": ["x_a", "x_b"]}, + ) + b = compose_system_prompt( + today=fixed_today, + enabled_tool_names={"scrape_webpage", "web_search"}, # set order shouldn't matter + mcp_connector_tools={"X": ["x_a", "x_b"]}, + ) + assert a == b diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_action_log.py b/surfsense_backend/tests/unit/agents/new_chat/test_action_log.py new file mode 100644 index 000000000..6834b5be7 --- /dev/null +++ b/surfsense_backend/tests/unit/agents/new_chat/test_action_log.py @@ -0,0 +1,311 @@ +"""Unit tests for ActionLogMiddleware (Tier 5.2).""" + +from __future__ import annotations + +from dataclasses import dataclass +from typing import Any +from unittest.mock import AsyncMock, patch + +import pytest +from langchain_core.messages import ToolMessage +from langchain_core.tools import tool + +from app.agents.new_chat.feature_flags import AgentFeatureFlags +from app.agents.new_chat.middleware.action_log import ActionLogMiddleware +from app.agents.new_chat.tools.registry import ToolDefinition + + +@dataclass +class _FakeRequest: + """Minimal stand-in for ToolCallRequest used in unit tests.""" + + tool_call: dict[str, Any] + tool: Any = None + state: Any = None + runtime: Any = None + + +@tool +def make_widget(color: str, size: int) -> str: + """Create a widget.""" + return f"made {color} {size}" + + +def _enabled_flags(**overrides: bool) -> AgentFeatureFlags: + return AgentFeatureFlags( + disable_new_agent_stack=False, + enable_action_log=True, + **overrides, + ) + + +def _disabled_flags() -> AgentFeatureFlags: + return AgentFeatureFlags(disable_new_agent_stack=False, enable_action_log=False) + + +@pytest.fixture +def patch_get_flags(): + def _patch(flags: AgentFeatureFlags): + return patch( + "app.agents.new_chat.middleware.action_log.get_flags", + return_value=flags, + ) + + return _patch + + +@pytest.fixture +def fake_session_factory(): + """Patch ``shielded_async_session`` with a recording fake.""" + captured: dict[str, list] = {"rows": []} + + class _FakeSession: + async def __aenter__(self): + return self + + async def __aexit__(self, exc_type, exc, tb): + return False + + def add(self, row): + captured["rows"].append(row) + + async def commit(self): + captured["committed"] = True + + def _factory(): + return _FakeSession() + + return captured, _factory + + +class TestActionLogMiddlewareDisabled: + @pytest.mark.asyncio + async def test_no_op_when_flag_off(self, patch_get_flags) -> None: + mw = ActionLogMiddleware(thread_id=1, search_space_id=1, user_id=None) + request = _FakeRequest( + tool_call={"name": "make_widget", "args": {"color": "red", "size": 1}, "id": "tc1"} + ) + handler = AsyncMock(return_value=ToolMessage(content="ok", tool_call_id="tc1")) + with patch_get_flags(_disabled_flags()): + result = await mw.awrap_tool_call(request, handler) + handler.assert_awaited_once() + assert isinstance(result, ToolMessage) + + @pytest.mark.asyncio + async def test_no_op_when_thread_id_none(self, patch_get_flags) -> None: + mw = ActionLogMiddleware(thread_id=None, search_space_id=1, user_id=None) + request = _FakeRequest( + tool_call={"name": "make_widget", "args": {}, "id": "tc1"} + ) + handler = AsyncMock(return_value=ToolMessage(content="ok", tool_call_id="tc1")) + with patch_get_flags(_enabled_flags()): + result = await mw.awrap_tool_call(request, handler) + assert isinstance(result, ToolMessage) + + +class TestActionLogMiddlewarePersistence: + @pytest.mark.asyncio + async def test_writes_row_on_success( + self, patch_get_flags, fake_session_factory + ) -> None: + captured, factory = fake_session_factory + mw = ActionLogMiddleware(thread_id=42, search_space_id=7, user_id="u1") + request = _FakeRequest( + tool_call={ + "name": "make_widget", + "args": {"color": "red", "size": 3}, + "id": "tc-abc", + }, + ) + result_msg = ToolMessage( + content="ok", tool_call_id="tc-abc", id="msg-1" + ) + handler = AsyncMock(return_value=result_msg) + + with patch_get_flags(_enabled_flags()), patch( + "app.db.shielded_async_session", side_effect=lambda: factory() + ): + result = await mw.awrap_tool_call(request, handler) + + assert result is result_msg + assert len(captured["rows"]) == 1 + row = captured["rows"][0] + assert row.thread_id == 42 + assert row.search_space_id == 7 + assert row.user_id == "u1" + assert row.tool_name == "make_widget" + assert row.args == {"color": "red", "size": 3} + assert row.result_id == "msg-1" + assert row.error is None + assert row.reverse_descriptor is None + assert row.reversible is False + + @pytest.mark.asyncio + async def test_writes_row_on_failure_and_reraises( + self, patch_get_flags, fake_session_factory + ) -> None: + captured, factory = fake_session_factory + mw = ActionLogMiddleware(thread_id=42, search_space_id=7, user_id="u1") + request = _FakeRequest( + tool_call={"name": "make_widget", "args": {"color": "red"}, "id": "tc1"} + ) + handler = AsyncMock(side_effect=ValueError("boom")) + + with patch_get_flags(_enabled_flags()), patch( + "app.db.shielded_async_session", side_effect=lambda: factory() + ), pytest.raises(ValueError, match="boom"): + await mw.awrap_tool_call(request, handler) + + assert len(captured["rows"]) == 1 + row = captured["rows"][0] + assert row.tool_name == "make_widget" + assert row.error == {"type": "ValueError", "message": "boom"} + assert row.result_id is None + + @pytest.mark.asyncio + async def test_persistence_failure_does_not_break_tool_call( + self, patch_get_flags + ) -> None: + """Even if the DB write blows up, the tool's result must reach the model.""" + mw = ActionLogMiddleware(thread_id=1, search_space_id=1, user_id=None) + request = _FakeRequest( + tool_call={"name": "make_widget", "args": {}, "id": "tc1"} + ) + result_msg = ToolMessage(content="ok", tool_call_id="tc1") + handler = AsyncMock(return_value=result_msg) + + def _exploding_session(): + raise RuntimeError("DB is down") + + with patch_get_flags(_enabled_flags()), patch( + "app.db.shielded_async_session", side_effect=_exploding_session + ): + result = await mw.awrap_tool_call(request, handler) + assert result is result_msg + + +class TestReverseDescriptor: + @pytest.mark.asyncio + async def test_renders_reverse_descriptor_when_tool_declares_one( + self, patch_get_flags, fake_session_factory + ) -> None: + captured, factory = fake_session_factory + + def _reverse(args, result): + return {"tool": "delete_widget", "args": {"id": result["id"]}} + + tool_def = ToolDefinition( + name="make_widget", + description="Create a widget", + factory=lambda deps: make_widget, + reverse=_reverse, + ) + mw = ActionLogMiddleware( + thread_id=1, + search_space_id=1, + user_id="u", + tool_definitions={"make_widget": tool_def}, + ) + request = _FakeRequest( + tool_call={ + "name": "make_widget", + "args": {"color": "blue", "size": 1}, + "id": "tc-xyz", + }, + ) + result_msg = ToolMessage( + content='{"id": "widget-9"}', tool_call_id="tc-xyz", id="msg-9" + ) + handler = AsyncMock(return_value=result_msg) + + with patch_get_flags(_enabled_flags()), patch( + "app.db.shielded_async_session", side_effect=lambda: factory() + ): + await mw.awrap_tool_call(request, handler) + + row = captured["rows"][0] + assert row.reversible is True + assert row.reverse_descriptor == { + "tool": "delete_widget", + "args": {"id": "widget-9"}, + } + + @pytest.mark.asyncio + async def test_swallows_reverse_callable_errors( + self, patch_get_flags, fake_session_factory + ) -> None: + captured, factory = fake_session_factory + + def _bad_reverse(args, result): + raise RuntimeError("reverse blew up") + + tool_def = ToolDefinition( + name="make_widget", + description="Create a widget", + factory=lambda deps: make_widget, + reverse=_bad_reverse, + ) + mw = ActionLogMiddleware( + thread_id=1, + search_space_id=1, + user_id=None, + tool_definitions={"make_widget": tool_def}, + ) + request = _FakeRequest( + tool_call={"name": "make_widget", "args": {}, "id": "tc1"} + ) + result_msg = ToolMessage(content="ok", tool_call_id="tc1") + handler = AsyncMock(return_value=result_msg) + + with patch_get_flags(_enabled_flags()), patch( + "app.db.shielded_async_session", side_effect=lambda: factory() + ): + await mw.awrap_tool_call(request, handler) + + row = captured["rows"][0] + assert row.reversible is False + assert row.reverse_descriptor is None + + @pytest.mark.asyncio + async def test_no_reverse_when_tool_definition_missing( + self, patch_get_flags, fake_session_factory + ) -> None: + captured, factory = fake_session_factory + mw = ActionLogMiddleware(thread_id=1, search_space_id=1, user_id=None) + request = _FakeRequest( + tool_call={"name": "unknown_tool", "args": {}, "id": "tc1"} + ) + handler = AsyncMock( + return_value=ToolMessage(content="ok", tool_call_id="tc1") + ) + with patch_get_flags(_enabled_flags()), patch( + "app.db.shielded_async_session", side_effect=lambda: factory() + ): + await mw.awrap_tool_call(request, handler) + row = captured["rows"][0] + assert row.reversible is False + + +class TestArgsTruncation: + @pytest.mark.asyncio + async def test_huge_args_payload_is_truncated( + self, patch_get_flags, fake_session_factory + ) -> None: + captured, factory = fake_session_factory + mw = ActionLogMiddleware(thread_id=1, search_space_id=1, user_id=None) + # Build a > 32KB string so the persisted payload triggers the truncation path. + huge = "x" * (40 * 1024) + request = _FakeRequest( + tool_call={"name": "make_widget", "args": {"blob": huge}, "id": "tc1"}, + ) + handler = AsyncMock( + return_value=ToolMessage(content="ok", tool_call_id="tc1") + ) + with patch_get_flags(_enabled_flags()), patch( + "app.db.shielded_async_session", side_effect=lambda: factory() + ): + await mw.awrap_tool_call(request, handler) + row = captured["rows"][0] + assert row.args is not None + assert row.args.get("_truncated") is True + assert row.args.get("_size", 0) >= 40 * 1024 diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_busy_mutex.py b/surfsense_backend/tests/unit/agents/new_chat/test_busy_mutex.py new file mode 100644 index 000000000..0c7bf17f6 --- /dev/null +++ b/surfsense_backend/tests/unit/agents/new_chat/test_busy_mutex.py @@ -0,0 +1,90 @@ +"""Tests for BusyMutexMiddleware: per-thread lock + cancel event behavior.""" + +from __future__ import annotations + +import pytest + +from app.agents.new_chat.errors import BusyError +from app.agents.new_chat.middleware.busy_mutex import ( + BusyMutexMiddleware, + get_cancel_event, + manager, + request_cancel, + reset_cancel, +) + +pytestmark = pytest.mark.unit + + +class _Runtime: + def __init__(self, thread_id: str | None) -> None: + self.config = {"configurable": {"thread_id": thread_id}} + + +@pytest.mark.asyncio +async def test_first_acquire_succeeds_and_release_unblocks() -> None: + mw = BusyMutexMiddleware() + runtime = _Runtime("t1") + await mw.abefore_agent({}, runtime) + + # Lock should now be held + lock = manager.lock_for("t1") + assert lock.locked() + + await mw.aafter_agent({}, runtime) + assert not lock.locked() + + +@pytest.mark.asyncio +async def test_second_concurrent_acquire_raises_busy() -> None: + mw_a = BusyMutexMiddleware() + mw_b = BusyMutexMiddleware() + runtime = _Runtime("t-conflict") + await mw_a.abefore_agent({}, runtime) + + with pytest.raises(BusyError) as excinfo: + await mw_b.abefore_agent({}, runtime) + assert excinfo.value.request_id == "t-conflict" + + await mw_a.aafter_agent({}, runtime) + # After release, mw_b can acquire + await mw_b.abefore_agent({}, runtime) + await mw_b.aafter_agent({}, runtime) + + +@pytest.mark.asyncio +async def test_cancel_event_lifecycle() -> None: + mw = BusyMutexMiddleware() + runtime = _Runtime("t-cancel") + + await mw.abefore_agent({}, runtime) + event = get_cancel_event("t-cancel") + assert not event.is_set() + + request_cancel("t-cancel") + assert event.is_set() + + # End of turn should reset + await mw.aafter_agent({}, runtime) + assert not event.is_set() + + +@pytest.mark.asyncio +async def test_no_thread_id_raises_when_required() -> None: + mw = BusyMutexMiddleware(require_thread_id=True) + runtime = _Runtime(None) + with pytest.raises(BusyError): + await mw.abefore_agent({}, runtime) + + +@pytest.mark.asyncio +async def test_no_thread_id_skipped_when_not_required() -> None: + mw = BusyMutexMiddleware(require_thread_id=False) + runtime = _Runtime(None) + await mw.abefore_agent({}, runtime) + await mw.aafter_agent({}, runtime) + + +def test_reset_cancel_idempotent() -> None: + # Should not raise even if event was never created + reset_cancel("never-seen") diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_compaction.py b/surfsense_backend/tests/unit/agents/new_chat/test_compaction.py new file mode 100644 index 000000000..4d8d6805c --- /dev/null +++ b/surfsense_backend/tests/unit/agents/new_chat/test_compaction.py @@ -0,0 +1,107 @@ +"""Tests for SurfSenseCompactionMiddleware: protected SystemMessage handling and content sanitization.""" + +from __future__ import annotations + +import pytest +from langchain_core.messages import ( + AIMessage, + HumanMessage, + SystemMessage, + ToolMessage, +) + +from app.agents.new_chat.middleware.compaction import ( + PROTECTED_SYSTEM_PREFIXES, + _is_protected_system_message, + _sanitize_message_content, +) + +pytestmark = pytest.mark.unit + + +class TestIsProtectedSystemMessage: + @pytest.mark.parametrize("prefix", PROTECTED_SYSTEM_PREFIXES) + def test_each_prefix_protected(self, prefix: str) -> None: + msg = SystemMessage(content=f"{prefix}\nbody\n") + assert _is_protected_system_message(msg) is True + + def test_unprotected_system_message(self) -> None: + assert _is_protected_system_message(SystemMessage(content="random instructions")) is False + + def test_human_message_never_protected(self) -> None: + assert _is_protected_system_message(HumanMessage(content="...")) is False + + def test_tolerates_leading_whitespace(self) -> None: + msg = SystemMessage(content=" \n\n...") + assert _is_protected_system_message(msg) is True + + +class TestSanitizeMessageContent: + def test_returns_same_message_when_content_present(self) -> None: + msg = AIMessage(content="hello") + assert _sanitize_message_content(msg) is msg + + def test_replaces_none_with_empty_string(self) -> None: + # Pydantic blocks ``content=None`` at construction; the real + # crash happens when the streaming layer mutates ``content`` + # after-the-fact. Replicate that by force-setting on a built + # message. + msg = AIMessage( + content="", + tool_calls=[{"name": "x", "args": {}, "id": "1"}], + ) + # Bypass pydantic validation to simulate the LiteLLM/Bedrock case + object.__setattr__(msg, "content", None) + sanitized = _sanitize_message_content(msg) + assert sanitized.content == "" + + +class TestPartitionMessages: + """Verify the partition override surfaces protected SystemMessages + into ``preserved_messages`` regardless of cutoff position. + """ + + def _build_partitioner(self): + # Construct a thin shim — we can't easily instantiate the full + # SurfSenseCompactionMiddleware without a real model, but the + # override path needs ``_lc_helper`` to delegate to. We mock + # that with a simple slicing partitioner equivalent to the real one. + from app.agents.new_chat.middleware.compaction import ( + SurfSenseCompactionMiddleware, + ) + + class _LcHelper: + @staticmethod + def _partition_messages(messages, cutoff): + return messages[:cutoff], messages[cutoff:] + + class _Stub(SurfSenseCompactionMiddleware): + def __init__(self): + self._lc_helper = _LcHelper() + + return _Stub() + + def test_protected_system_message_preserved_even_in_summarize_half(self) -> None: + partitioner = self._build_partitioner() + protected = SystemMessage(content="\n...") + msgs = [ + HumanMessage(content="old human"), + AIMessage(content="old ai"), + protected, + ToolMessage(content="tool 1", tool_call_id="t1"), + HumanMessage(content="new"), + ] + # Cutoff = 4 means everything before index 4 should be summarized + to_summary, preserved = partitioner._partition_messages(msgs, 4) + + assert protected not in to_summary + assert protected in preserved + # The non-protected old messages remain in to_summary + assert any(isinstance(m, HumanMessage) and m.content == "old human" for m in to_summary) + + def test_unprotected_messages_unaffected(self) -> None: + partitioner = self._build_partitioner() + msgs = [HumanMessage(content="a"), HumanMessage(content="b"), HumanMessage(content="c")] + to_summary, preserved = partitioner._partition_messages(msgs, 2) + assert [m.content for m in to_summary] == ["a", "b"] + assert [m.content for m in preserved] == ["c"] diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_context_editing.py b/surfsense_backend/tests/unit/agents/new_chat/test_context_editing.py new file mode 100644 index 000000000..3c31155d4 --- /dev/null +++ b/surfsense_backend/tests/unit/agents/new_chat/test_context_editing.py @@ -0,0 +1,107 @@ +"""Tests for SpillToBackendEdit and SpillingContextEditingMiddleware.""" + +from __future__ import annotations + +from typing import Any + +import pytest +from langchain_core.messages import AIMessage, HumanMessage, ToolMessage + +from app.agents.new_chat.middleware.context_editing import ( + SpillToBackendEdit, + _build_spill_placeholder, +) + +pytestmark = pytest.mark.unit + + +def _build_history(num_pairs: int = 6) -> list[Any]: + """Build a long history of (AIMessage with tool_call, ToolMessage) pairs.""" + msgs: list[Any] = [HumanMessage(content="please do many things")] + for i in range(num_pairs): + msgs.append( + AIMessage( + content="", + tool_calls=[ + {"name": f"tool_{i}", "args": {"i": i}, "id": f"call-{i}"}, + ], + ) + ) + msgs.append( + ToolMessage( + content="x" * 5000, + tool_call_id=f"call-{i}", + name=f"tool_{i}", + id=f"tool-msg-{i}", + ) + ) + return msgs + + +def _approx_count(messages: list[Any]) -> int: + """Trivial token counter: 1 token per 4 chars.""" + total = 0 + for msg in messages: + content = getattr(msg, "content", "") + if isinstance(content, str): + total += len(content) // 4 + return total + + +class TestSpillEdit: + def test_below_trigger_does_nothing(self) -> None: + edit = SpillToBackendEdit(trigger=1_000_000, keep=2) + msgs = _build_history(3) + original_lengths = [len(getattr(m, "content", "")) for m in msgs] + edit.apply(msgs, count_tokens=_approx_count) + new_lengths = [len(getattr(m, "content", "")) for m in msgs] + assert original_lengths == new_lengths + assert edit.pending_spills == [] + + def test_above_trigger_clears_and_records(self) -> None: + edit = SpillToBackendEdit(trigger=100, keep=1, path_prefix="/tool_outputs") + msgs = _build_history(4) + edit.apply(msgs, count_tokens=_approx_count) + + # The most-recent ToolMessage (keep=1) should remain intact + tool_messages = [m for m in msgs if isinstance(m, ToolMessage)] + intact = tool_messages[-1] + assert intact.content.startswith("x") # untouched + + # Earlier ToolMessages should now contain the placeholder text + cleared = [ + m for m in tool_messages + if isinstance(m.content, str) and m.content.startswith("[cleared") + ] + assert len(cleared) >= 1 + # And the spill list should match + assert len(edit.pending_spills) == len(cleared) + + def test_excluded_tools_not_cleared(self) -> None: + edit = SpillToBackendEdit( + trigger=100, + keep=0, + exclude_tools=("tool_0",), + ) + msgs = _build_history(4) + edit.apply(msgs, count_tokens=_approx_count) + + first_tool = next( + m for m in msgs if isinstance(m, ToolMessage) and m.name == "tool_0" + ) + # Excluded — untouched + assert first_tool.content.startswith("x") + + def test_drain_clears_pending(self) -> None: + edit = SpillToBackendEdit(trigger=100, keep=1) + msgs = _build_history(4) + edit.apply(msgs, count_tokens=_approx_count) + first_drain = edit.drain_pending() + assert len(first_drain) > 0 + assert edit.drain_pending() == [] + + def test_placeholder_format(self) -> None: + path = "/tool_outputs/thread-1/tool-msg-0.txt" + text = _build_spill_placeholder(path) + assert path in text + assert "explore" in text # mentions the recovery agent diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_dedup_tool_calls.py b/surfsense_backend/tests/unit/agents/new_chat/test_dedup_tool_calls.py new file mode 100644 index 000000000..95017d744 --- /dev/null +++ b/surfsense_backend/tests/unit/agents/new_chat/test_dedup_tool_calls.py @@ -0,0 +1,132 @@ +"""Tests for declarative dedup_key on ToolDefinition (Tier 2.3 migration).""" + +from __future__ import annotations + +import pytest +from langchain_core.messages import AIMessage +from langchain_core.tools import StructuredTool + +from app.agents.new_chat.middleware.dedup_tool_calls import ( + DedupHITLToolCallsMiddleware, +) + +pytestmark = pytest.mark.unit + + +def _make_tool(name: str, *, dedup_key=None, hitl_dedup_key=None): + metadata = {} + if dedup_key is not None: + metadata["dedup_key"] = dedup_key + if hitl_dedup_key is not None: + metadata["hitl"] = True + metadata["hitl_dedup_key"] = hitl_dedup_key + + def _fn(**kwargs): + return "ok" + + return StructuredTool.from_function( + func=_fn, name=name, description="x", metadata=metadata + ) + + +def _msg(*calls: dict) -> AIMessage: + return AIMessage(content="", tool_calls=list(calls)) + + +class _Runtime: + pass + + +def test_callable_dedup_key_takes_priority() -> None: + tool = _make_tool( + "create_doc", + dedup_key=lambda args: f"{args.get('parent_id')}::{args.get('title')}", + ) + mw = DedupHITLToolCallsMiddleware(agent_tools=[tool]) + state = { + "messages": [ + _msg( + {"name": "create_doc", "args": {"parent_id": "x", "title": "y"}, "id": "1"}, + {"name": "create_doc", "args": {"parent_id": "x", "title": "y"}, "id": "2"}, + {"name": "create_doc", "args": {"parent_id": "x", "title": "z"}, "id": "3"}, + ) + ] + } + out = mw.after_model(state, _Runtime()) + assert out is not None + new_calls = out["messages"][0].tool_calls + assert len(new_calls) == 2 # one duplicate dropped + assert {c["id"] for c in new_calls} == {"1", "3"} + + +def test_string_hitl_dedup_key_still_works() -> None: + tool = _make_tool("send_x", hitl_dedup_key="subject") + mw = DedupHITLToolCallsMiddleware(agent_tools=[tool]) + state = { + "messages": [ + _msg( + {"name": "send_x", "args": {"subject": "Hello"}, "id": "1"}, + {"name": "send_x", "args": {"subject": "hello"}, "id": "2"}, # case + ) + ] + } + out = mw.after_model(state, _Runtime()) + assert out is not None + assert len(out["messages"][0].tool_calls) == 1 + + +def test_no_agent_tools_means_no_dedup() -> None: + """After the cleanup tier removed the legacy ``_NATIVE_HITL_TOOL_DEDUP_KEYS`` + map, dedup is purely declarative — no resolvers means no dedup runs. + + Coverage for the previously hardcoded native HITL tools now lives on + each :class:`ToolDefinition.dedup_key` in + :mod:`app.agents.new_chat.tools.registry`, which is wired through to + ``tool.metadata`` by :func:`build_tools`. + """ + mw = DedupHITLToolCallsMiddleware(agent_tools=None) + state = { + "messages": [ + _msg( + {"name": "create_notion_page", "args": {"title": "X"}, "id": "1"}, + {"name": "create_notion_page", "args": {"title": "x"}, "id": "2"}, + ) + ] + } + out = mw.after_model(state, _Runtime()) + assert out is None + + +def test_registry_propagates_dedup_key_to_tool_metadata() -> None: + """Smoke-check the wiring path that replaced the legacy native map. + + ``ToolDefinition.dedup_key`` set in the registry must be copied onto + the constructed tool's ``metadata`` so :class:`DedupHITLToolCallsMiddleware` + can pick it up at agent build time. + """ + from app.agents.new_chat.tools.registry import ( + BUILTIN_TOOLS, + wrap_dedup_key_by_arg_name, + ) + + notion_tool_defs = [t for t in BUILTIN_TOOLS if t.name == "create_notion_page"] + assert notion_tool_defs, "registry should still expose create_notion_page" + tool_def = notion_tool_defs[0] + assert tool_def.dedup_key is not None + # Same wrapping helper used in the registry — sanity check identity + sample = wrap_dedup_key_by_arg_name("title")({"title": "Plan"}) + assert sample == "plan" + + +def test_unknown_tool_passes_through() -> None: + mw = DedupHITLToolCallsMiddleware(agent_tools=None) + state = { + "messages": [ + _msg( + {"name": "anything_else", "args": {"x": 1}, "id": "1"}, + {"name": "anything_else", "args": {"x": 1}, "id": "2"}, + ) + ] + } + out = mw.after_model(state, _Runtime()) + assert out is None # no dedup configured -> kept diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_default_permissions_layering.py b/surfsense_backend/tests/unit/agents/new_chat/test_default_permissions_layering.py new file mode 100644 index 000000000..d49edbfec --- /dev/null +++ b/surfsense_backend/tests/unit/agents/new_chat/test_default_permissions_layering.py @@ -0,0 +1,128 @@ +"""Lock in the default-allow layering used by ``chat_deepagent``. + +The agent factory wires ``PermissionMiddleware`` with three rulesets, +earliest -> latest: + +1. ``surfsense_defaults`` (single ``allow */*`` rule) +2. ``connector_synthesized`` (deny rules for tools whose required + connector is missing) +3. (future) user-defined rules from the Agent Permissions UI + +Without #1 every read-only built-in (``ls``, ``read_file``, ``grep``, +``glob``, ``web_search`` …) defaulted to ``ask`` because +``permissions.evaluate`` returns ``ask`` when no rule matches. That +caused two production-painful behaviors: + +* Resume payloads with a prior reject decision bled into innocent + read-only tool calls, raising ``RejectedError("ls")``. +* Mutating connector tools got *double* prompted — once via the + middleware ``ask`` and again via the per-tool ``interrupt()`` in + ``app.agents.new_chat.tools.hitl``. + +These tests pin the layering so a refactor that drops the default +ruleset fails loud. +""" + +from __future__ import annotations + +import pytest + +from app.agents.new_chat.permissions import ( + Rule, + Ruleset, + aggregate_action, + evaluate_many, +) + +pytestmark = pytest.mark.unit + + +def _layered_rulesets(connector_denies: list[Rule]) -> list[Ruleset]: + """Replicate ``chat_deepagent`` layering for the test.""" + return [ + Ruleset( + rules=[Rule(permission="*", pattern="*", action="allow")], + origin="surfsense_defaults", + ), + Ruleset(rules=connector_denies, origin="connector_synthesized"), + ] + + +class TestReadOnlyToolsAllowed: + """Read-only built-ins must NOT default to ask.""" + + @pytest.mark.parametrize( + "tool_name", + [ + "ls", + "read_file", + "grep", + "glob", + "web_search", + "scrape_webpage", + "search_surfsense_docs", + "get_connected_accounts", + "write_todos", + "task", + "_noop", + "invalid", + "update_memory", + ], + ) + def test_default_allow_covers_safe_builtin(self, tool_name: str) -> None: + rulesets = _layered_rulesets(connector_denies=[]) + rules = evaluate_many(tool_name, [tool_name], *rulesets) + assert aggregate_action(rules) == "allow" + + +class TestConnectorDenyOverridesDefaultAllow: + """Connector-synthesized denies must beat the default-allow rule.""" + + def test_missing_connector_tool_is_denied(self) -> None: + rulesets = _layered_rulesets( + connector_denies=[ + Rule(permission="linear_create_issue", pattern="*", action="deny") + ] + ) + rules = evaluate_many( + "linear_create_issue", ["linear_create_issue"], *rulesets + ) + assert aggregate_action(rules) == "deny" + + def test_default_allow_still_applies_to_other_tools(self) -> None: + """A deny rule for one tool must not bleed onto unrelated calls.""" + rulesets = _layered_rulesets( + connector_denies=[ + Rule(permission="linear_create_issue", pattern="*", action="deny") + ] + ) + rules = evaluate_many("ls", ["ls"], *rulesets) + assert aggregate_action(rules) == "allow" + + +class TestUserRuleOverridesDefault: + """User rules layered last must override the default-allow rule.""" + + def test_user_ask_overrides_default_allow(self) -> None: + defaults = Ruleset( + rules=[Rule(permission="*", pattern="*", action="allow")], + origin="surfsense_defaults", + ) + user_ruleset = Ruleset( + rules=[Rule(permission="ls", pattern="*", action="ask")], + origin="user", + ) + rules = evaluate_many("ls", ["ls"], defaults, user_ruleset) + assert aggregate_action(rules) == "ask" + + def test_user_deny_overrides_default_allow(self) -> None: + defaults = Ruleset( + rules=[Rule(permission="*", pattern="*", action="allow")], + origin="surfsense_defaults", + ) + user_ruleset = Ruleset( + rules=[Rule(permission="send_*", pattern="*", action="deny")], + origin="user", + ) + rules = evaluate_many("send_gmail_email", ["send_gmail_email"], defaults, user_ruleset) + assert aggregate_action(rules) == "deny" diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_doom_loop.py b/surfsense_backend/tests/unit/agents/new_chat/test_doom_loop.py new file mode 100644 index 000000000..c54163dc3 --- /dev/null +++ b/surfsense_backend/tests/unit/agents/new_chat/test_doom_loop.py @@ -0,0 +1,99 @@ +"""Tests for DoomLoopMiddleware signature equality detection.""" + +from __future__ import annotations + +import pytest +from langchain_core.messages import AIMessage + +from app.agents.new_chat.middleware.doom_loop import DoomLoopMiddleware, _signature + +pytestmark = pytest.mark.unit + + +def test_signature_is_stable_for_identical_args() -> None: + a = _signature("search", {"q": "hello", "n": 10}) + b = _signature("search", {"n": 10, "q": "hello"}) + assert a == b + + +def test_signature_changes_with_args() -> None: + a = _signature("search", {"q": "hello"}) + b = _signature("search", {"q": "world"}) + assert a != b + + +def test_signature_changes_with_name() -> None: + a = _signature("search", {"q": "x"}) + b = _signature("read", {"q": "x"}) + assert a != b + + +class _FakeRuntime: + def __init__(self, thread_id: str | None = "thread-1") -> None: + self.config = {"configurable": {"thread_id": thread_id}} + + +def _msg_calling(name: str, args: dict, call_id: str) -> AIMessage: + return AIMessage( + content="", + tool_calls=[{"name": name, "args": args, "id": call_id}], + ) + + +def test_threshold_triggers_after_n_identical_calls() -> None: + mw = DoomLoopMiddleware(threshold=3) + runtime = _FakeRuntime() + + # First two calls — under threshold + for i in range(2): + out = mw.after_model( + {"messages": [_msg_calling("repeat", {"x": 1}, f"call-{i}")]}, + runtime, + ) + assert out is None + + # Third identical call should trigger ``langgraph.types.interrupt``. + # In a unit-test context (no runnable graph), ``interrupt`` raises + # ``RuntimeError`` because ``get_config`` has nothing to bind to — + # we accept that as proof the interrupt path was taken (the + # alternative would be no exception, which would mean the loop + # detection never fired). + with pytest.raises(Exception) as excinfo: + mw.after_model( + {"messages": [_msg_calling("repeat", {"x": 1}, "call-3")]}, + runtime, + ) + name = type(excinfo.value).__name__.lower() + assert ( + "interrupt" in name + or "runtimeerror" in name + ), f"Expected an interrupt-style exception, got {name}" + + +def test_does_not_trigger_when_args_differ() -> None: + mw = DoomLoopMiddleware(threshold=2) + runtime = _FakeRuntime() + out = mw.after_model( + {"messages": [_msg_calling("repeat", {"x": 1}, "1")]}, runtime + ) + assert out is None + out = mw.after_model( + {"messages": [_msg_calling("repeat", {"x": 2}, "2")]}, runtime + ) + assert out is None + + +def test_separate_threads_have_independent_windows() -> None: + mw = DoomLoopMiddleware(threshold=2) + rt_a = _FakeRuntime(thread_id="A") + rt_b = _FakeRuntime(thread_id="B") + + mw.after_model({"messages": [_msg_calling("foo", {}, "1")]}, rt_a) + # thread B should NOT count thread A's call + out = mw.after_model({"messages": [_msg_calling("foo", {}, "1")]}, rt_b) + assert out is None # not yet at threshold for B + + +def test_invalid_threshold_rejected() -> None: + with pytest.raises(ValueError): + DoomLoopMiddleware(threshold=1) diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_feature_flags.py b/surfsense_backend/tests/unit/agents/new_chat/test_feature_flags.py new file mode 100644 index 000000000..38a70a443 --- /dev/null +++ b/surfsense_backend/tests/unit/agents/new_chat/test_feature_flags.py @@ -0,0 +1,120 @@ +"""Tests for the agent feature-flag system.""" + +from __future__ import annotations + +import pytest + +from app.agents.new_chat.feature_flags import ( + AgentFeatureFlags, + reload_for_tests, +) + +pytestmark = pytest.mark.unit + + +def _clear_all(monkeypatch: pytest.MonkeyPatch) -> None: + for name in [ + "SURFSENSE_DISABLE_NEW_AGENT_STACK", + "SURFSENSE_ENABLE_CONTEXT_EDITING", + "SURFSENSE_ENABLE_COMPACTION_V2", + "SURFSENSE_ENABLE_RETRY_AFTER", + "SURFSENSE_ENABLE_MODEL_FALLBACK", + "SURFSENSE_ENABLE_MODEL_CALL_LIMIT", + "SURFSENSE_ENABLE_TOOL_CALL_LIMIT", + "SURFSENSE_ENABLE_TOOL_CALL_REPAIR", + "SURFSENSE_ENABLE_DOOM_LOOP", + "SURFSENSE_ENABLE_PERMISSION", + "SURFSENSE_ENABLE_BUSY_MUTEX", + "SURFSENSE_ENABLE_LLM_TOOL_SELECTOR", + "SURFSENSE_ENABLE_SKILLS", + "SURFSENSE_ENABLE_SPECIALIZED_SUBAGENTS", + "SURFSENSE_ENABLE_KB_PLANNER_RUNNABLE", + "SURFSENSE_ENABLE_ACTION_LOG", + "SURFSENSE_ENABLE_REVERT_ROUTE", + "SURFSENSE_ENABLE_PLUGIN_LOADER", + "SURFSENSE_ENABLE_OTEL", + ]: + monkeypatch.delenv(name, raising=False) + + +def test_defaults_all_off(monkeypatch: pytest.MonkeyPatch) -> None: + _clear_all(monkeypatch) + flags = reload_for_tests() + assert isinstance(flags, AgentFeatureFlags) + assert flags.disable_new_agent_stack is False + assert flags.any_new_middleware_enabled() is False + + +def test_master_kill_switch_overrides_individual_flags( + monkeypatch: pytest.MonkeyPatch, +) -> None: + _clear_all(monkeypatch) + monkeypatch.setenv("SURFSENSE_DISABLE_NEW_AGENT_STACK", "true") + monkeypatch.setenv("SURFSENSE_ENABLE_CONTEXT_EDITING", "true") + monkeypatch.setenv("SURFSENSE_ENABLE_PERMISSION", "true") + + flags = reload_for_tests() + assert flags.disable_new_agent_stack is True + assert flags.enable_context_editing is False + assert flags.enable_permission is False + assert flags.any_new_middleware_enabled() is False + + +@pytest.mark.parametrize("truthy", ["1", "true", "TRUE", "yes", "on"]) +def test_individual_flags_truthy_values( + monkeypatch: pytest.MonkeyPatch, truthy: str +) -> None: + _clear_all(monkeypatch) + monkeypatch.setenv("SURFSENSE_ENABLE_RETRY_AFTER", truthy) + flags = reload_for_tests() + assert flags.enable_retry_after is True + assert flags.any_new_middleware_enabled() is True + + +@pytest.mark.parametrize("falsy", ["0", "false", "no", "off", "", "garbage"]) +def test_individual_flags_falsy_values( + monkeypatch: pytest.MonkeyPatch, falsy: str +) -> None: + _clear_all(monkeypatch) + monkeypatch.setenv("SURFSENSE_ENABLE_RETRY_AFTER", falsy) + flags = reload_for_tests() + assert flags.enable_retry_after is False + + +def test_each_flag_can_be_set_independently(monkeypatch: pytest.MonkeyPatch) -> None: + _clear_all(monkeypatch) + flag_to_env = { + "enable_context_editing": "SURFSENSE_ENABLE_CONTEXT_EDITING", + "enable_compaction_v2": "SURFSENSE_ENABLE_COMPACTION_V2", + "enable_retry_after": "SURFSENSE_ENABLE_RETRY_AFTER", + "enable_model_fallback": "SURFSENSE_ENABLE_MODEL_FALLBACK", + "enable_model_call_limit": "SURFSENSE_ENABLE_MODEL_CALL_LIMIT", + "enable_tool_call_limit": "SURFSENSE_ENABLE_TOOL_CALL_LIMIT", + "enable_tool_call_repair": "SURFSENSE_ENABLE_TOOL_CALL_REPAIR", + "enable_doom_loop": "SURFSENSE_ENABLE_DOOM_LOOP", + "enable_permission": "SURFSENSE_ENABLE_PERMISSION", + "enable_busy_mutex": "SURFSENSE_ENABLE_BUSY_MUTEX", + "enable_llm_tool_selector": "SURFSENSE_ENABLE_LLM_TOOL_SELECTOR", + "enable_skills": "SURFSENSE_ENABLE_SKILLS", + "enable_specialized_subagents": "SURFSENSE_ENABLE_SPECIALIZED_SUBAGENTS", + "enable_kb_planner_runnable": "SURFSENSE_ENABLE_KB_PLANNER_RUNNABLE", + "enable_action_log": "SURFSENSE_ENABLE_ACTION_LOG", + "enable_revert_route": "SURFSENSE_ENABLE_REVERT_ROUTE", + "enable_plugin_loader": "SURFSENSE_ENABLE_PLUGIN_LOADER", + "enable_otel": "SURFSENSE_ENABLE_OTEL", + } + + # `enable_otel` is intentionally orthogonal — it does NOT count toward + # ``any_new_middleware_enabled`` because OTel is observability-only and + # ships under its own ``OTEL_EXPORTER_OTLP_ENDPOINT`` requirement. + counts_toward_middleware = {k for k in flag_to_env if k != "enable_otel"} + + for attr, env_name in flag_to_env.items(): + _clear_all(monkeypatch) + monkeypatch.setenv(env_name, "true") + flags = reload_for_tests() + assert getattr(flags, attr) is True, f"{attr} did not flip on for {env_name}" + if attr in counts_toward_middleware: + assert flags.any_new_middleware_enabled() is True + else: + assert flags.any_new_middleware_enabled() is False diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_noop_injection.py b/surfsense_backend/tests/unit/agents/new_chat/test_noop_injection.py new file mode 100644 index 000000000..8555eea76 --- /dev/null +++ b/surfsense_backend/tests/unit/agents/new_chat/test_noop_injection.py @@ -0,0 +1,119 @@ +"""Tests for NoopInjectionMiddleware provider-compat logic.""" + +from __future__ import annotations + +import pytest +from langchain_core.messages import AIMessage, HumanMessage + +from app.agents.new_chat.middleware.noop_injection import ( + NOOP_TOOL_NAME, + NoopInjectionMiddleware, + _last_ai_has_tool_calls, + _provider_needs_noop, +) + +pytestmark = pytest.mark.unit + + +class _LiteLLMModel: + def _get_ls_params(self): + return {"ls_provider": "litellm"} + + +class _BedrockModel: + def _get_ls_params(self): + return {"ls_provider": "bedrock"} + + +class _OpenAIModel: + def _get_ls_params(self): + return {"ls_provider": "openai"} + + +class _ChatLiteLLM: # name-only fallback + pass + + +class TestProviderDetection: + def test_litellm(self) -> None: + assert _provider_needs_noop(_LiteLLMModel()) is True + + def test_bedrock(self) -> None: + assert _provider_needs_noop(_BedrockModel()) is True + + def test_openai_does_not_need(self) -> None: + assert _provider_needs_noop(_OpenAIModel()) is False + + def test_class_name_fallback(self) -> None: + assert _provider_needs_noop(_ChatLiteLLM()) is True + + +class TestHistoryDetection: + def test_last_ai_has_tool_calls(self) -> None: + msgs = [ + HumanMessage(content="hi"), + AIMessage(content="", tool_calls=[{"name": "x", "args": {}, "id": "1"}]), + ] + assert _last_ai_has_tool_calls(msgs) is True + + def test_last_ai_no_tool_calls(self) -> None: + msgs = [ + HumanMessage(content="hi"), + AIMessage(content="hello"), + ] + assert _last_ai_has_tool_calls(msgs) is False + + def test_no_ai_in_history(self) -> None: + assert _last_ai_has_tool_calls([HumanMessage(content="hi")]) is False + + +class _FakeRequest: + def __init__(self, *, tools, messages, model) -> None: + self.tools = tools + self.messages = messages + self.model = model + + def override(self, *, tools): + return _FakeRequest(tools=tools, messages=self.messages, model=self.model) + + +class TestShouldInject: + def test_injects_when_all_conditions_met(self) -> None: + mw = NoopInjectionMiddleware() + msgs = [ + HumanMessage(content="hi"), + AIMessage(content="", tool_calls=[{"name": "x", "args": {}, "id": "1"}]), + ] + req = _FakeRequest(tools=[], messages=msgs, model=_LiteLLMModel()) + assert mw._should_inject(req) is True + + def test_skips_when_tools_present(self) -> None: + mw = NoopInjectionMiddleware() + req = _FakeRequest( + tools=[object()], + messages=[AIMessage(content="", tool_calls=[{"name": "x", "args": {}, "id": "1"}])], + model=_LiteLLMModel(), + ) + assert mw._should_inject(req) is False + + def test_skips_when_no_history_tool_calls(self) -> None: + mw = NoopInjectionMiddleware() + req = _FakeRequest( + tools=[], + messages=[HumanMessage(content="hi")], + model=_LiteLLMModel(), + ) + assert mw._should_inject(req) is False + + def test_skips_for_openai(self) -> None: + mw = NoopInjectionMiddleware() + req = _FakeRequest( + tools=[], + messages=[AIMessage(content="", tool_calls=[{"name": "x", "args": {}, "id": "1"}])], + model=_OpenAIModel(), + ) + assert mw._should_inject(req) is False + + +def test_noop_tool_name_is_underscore_noop() -> None: + assert NOOP_TOOL_NAME == "_noop" diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_otel_span.py b/surfsense_backend/tests/unit/agents/new_chat/test_otel_span.py new file mode 100644 index 000000000..e5b171612 --- /dev/null +++ b/surfsense_backend/tests/unit/agents/new_chat/test_otel_span.py @@ -0,0 +1,195 @@ +"""Tests for the OtelSpanMiddleware adapter (Tier 3b).""" + +from __future__ import annotations + +from typing import Any +from unittest.mock import MagicMock + +import pytest +from langchain_core.messages import AIMessage, ToolMessage + +from app.agents.new_chat.middleware.otel_span import ( + OtelSpanMiddleware, + _annotate_model_response, + _annotate_tool_result, + _resolve_input_size, + _resolve_model_attrs, + _resolve_tool_name, +) + +pytestmark = pytest.mark.unit + + +@pytest.fixture(autouse=True) +def _disable_otel(monkeypatch: pytest.MonkeyPatch): + monkeypatch.delenv("OTEL_EXPORTER_OTLP_ENDPOINT", raising=False) + monkeypatch.setenv("SURFSENSE_DISABLE_OTEL", "true") + from app.observability import otel as ot + + ot.reload_for_tests() + yield + ot.reload_for_tests() + + +class TestResolveModelAttrs: + def test_extracts_model_name_and_provider(self) -> None: + request = MagicMock() + request.model = MagicMock(spec=["model_name", "provider"]) + request.model.model_name = "gpt-4o-mini" + request.model.provider = "openai" + assert _resolve_model_attrs(request) == ("gpt-4o-mini", "openai") + + def test_handles_missing_model(self) -> None: + request = MagicMock() + request.model = None + assert _resolve_model_attrs(request) == (None, None) + + def test_falls_back_through_attribute_chain(self) -> None: + request = MagicMock() + request.model = MagicMock(spec=["model_id", "_llm_type"]) + request.model.model_id = "claude-3-5-sonnet" + request.model._llm_type = "anthropic-chat" + model_id, provider = _resolve_model_attrs(request) + assert model_id == "claude-3-5-sonnet" + assert provider == "anthropic-chat" + + +class TestResolveToolName: + def test_prefers_request_tool_name(self) -> None: + request = MagicMock() + request.tool = MagicMock(name="ToolStub") + request.tool.name = "scrape_webpage" + assert _resolve_tool_name(request) == "scrape_webpage" + + def test_falls_back_to_tool_call_name(self) -> None: + request = MagicMock() + request.tool = None + request.tool_call = {"name": "web_search", "args": {}} + assert _resolve_tool_name(request) == "web_search" + + def test_unknown_when_nothing_resolves(self) -> None: + request = MagicMock() + request.tool = None + request.tool_call = {} + assert _resolve_tool_name(request) == "unknown" + + +class TestResolveInputSize: + def test_returns_repr_length_of_args(self) -> None: + request = MagicMock() + request.tool_call = {"args": {"query": "hello world"}} + size = _resolve_input_size(request) + assert isinstance(size, int) + assert size > 0 + + def test_handles_no_tool_call(self) -> None: + request = MagicMock() + request.tool_call = None + assert _resolve_input_size(request) is None + + +class TestAnnotateModelResponse: + def test_attaches_token_counts_when_present(self) -> None: + sp = MagicMock() + msg = AIMessage( + content="hello", + usage_metadata={ + "input_tokens": 100, + "output_tokens": 50, + "total_tokens": 150, + }, + ) + _annotate_model_response(sp, msg) + sp.set_attribute.assert_any_call("tokens.prompt", 100) + sp.set_attribute.assert_any_call("tokens.completion", 50) + sp.set_attribute.assert_any_call("tokens.total", 150) + + def test_handles_response_with_no_metadata(self) -> None: + sp = MagicMock() + msg = AIMessage(content="hello") + # Should not raise even when usage_metadata is missing + _annotate_model_response(sp, msg) + + +class TestAnnotateToolResult: + def test_records_size_and_status(self) -> None: + sp = MagicMock() + result = ToolMessage( + content="result text", + tool_call_id="abc", + status="success", + ) + _annotate_tool_result(sp, result) + sp.set_attribute.assert_any_call("tool.output.size", len("result text")) + sp.set_attribute.assert_any_call("tool.status", "success") + + def test_marks_errors(self) -> None: + sp = MagicMock() + result = ToolMessage( + content="oops", + tool_call_id="abc", + additional_kwargs={"error": {"code": "x"}}, + ) + _annotate_tool_result(sp, result) + sp.set_attribute.assert_any_call("tool.error", True) + + +@pytest.mark.asyncio +class TestMiddlewareIntegration: + async def test_awrap_model_call_passes_through_when_disabled(self) -> None: + mw = OtelSpanMiddleware() + called: dict[str, Any] = {} + + async def handler(req): + called["req"] = req + return AIMessage(content="ok") + + request = MagicMock() + result = await mw.awrap_model_call(request, handler) + assert called["req"] is request + assert isinstance(result, AIMessage) + assert result.content == "ok" + + async def test_awrap_tool_call_passes_through_when_disabled(self) -> None: + mw = OtelSpanMiddleware() + + async def handler(req): + return ToolMessage(content="result", tool_call_id="abc") + + request = MagicMock() + result = await mw.awrap_tool_call(request, handler) + assert isinstance(result, ToolMessage) + assert result.content == "result" + + async def test_awrap_model_call_propagates_exceptions(self) -> None: + mw = OtelSpanMiddleware() + + async def handler(req): + raise ValueError("boom") + + with pytest.raises(ValueError): + await mw.awrap_model_call(MagicMock(), handler) + + async def test_with_otel_enabled_does_not_alter_result( + self, monkeypatch: pytest.MonkeyPatch + ) -> None: + monkeypatch.delenv("SURFSENSE_DISABLE_OTEL", raising=False) + monkeypatch.setenv("OTEL_EXPORTER_OTLP_ENDPOINT", "http://localhost:4317") + from app.observability import otel as ot + + ot.reload_for_tests() + try: + mw = OtelSpanMiddleware() + + async def handler(req): + return AIMessage(content="enabled") + + request = MagicMock() + request.model = MagicMock() + request.model.model_name = "gpt-4o" + request.model.provider = "openai" + result = await mw.awrap_model_call(request, handler) + assert isinstance(result, AIMessage) + assert result.content == "enabled" + finally: + ot.reload_for_tests() diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_permission_middleware.py b/surfsense_backend/tests/unit/agents/new_chat/test_permission_middleware.py new file mode 100644 index 000000000..194a6eb27 --- /dev/null +++ b/surfsense_backend/tests/unit/agents/new_chat/test_permission_middleware.py @@ -0,0 +1,116 @@ +"""Tests for PermissionMiddleware end-to-end behavior.""" + +from __future__ import annotations + +import pytest +from langchain_core.messages import AIMessage, ToolMessage + +from app.agents.new_chat.errors import CorrectedError, RejectedError +from app.agents.new_chat.middleware.permission import PermissionMiddleware +from app.agents.new_chat.permissions import Rule, Ruleset + +pytestmark = pytest.mark.unit + + +class _FakeRuntime: + config: dict = {"configurable": {"thread_id": "test"}} + + +def _msg(*tool_calls: dict) -> AIMessage: + return AIMessage(content="", tool_calls=list(tool_calls)) + + +class TestAllow: + def test_passthrough_when_allow(self) -> None: + rs = Ruleset(rules=[Rule("send_email", "*", "allow")]) + mw = PermissionMiddleware(rulesets=[rs]) + state = {"messages": [_msg({"name": "send_email", "args": {}, "id": "1"})]} + out = mw.after_model(state, _FakeRuntime()) + assert out is None # no change + + +class TestDeny: + def test_replaces_with_deny_tool_message(self) -> None: + rs = Ruleset(rules=[Rule("send_email", "*", "deny")]) + mw = PermissionMiddleware(rulesets=[rs]) + state = {"messages": [_msg({"name": "send_email", "args": {}, "id": "1"})]} + out = mw.after_model(state, _FakeRuntime()) + assert out is not None + msgs = out["messages"] + # Find the deny ToolMessage + deny_msgs = [m for m in msgs if isinstance(m, ToolMessage)] + assert len(deny_msgs) == 1 + assert deny_msgs[0].status == "error" + assert "permission_denied" in str(deny_msgs[0].additional_kwargs) + # AIMessage's tool_calls should now be empty (denied call removed) + ai_msg = next(m for m in msgs if isinstance(m, AIMessage)) + assert ai_msg.tool_calls == [] + + def test_mixed_allow_deny(self) -> None: + rs = Ruleset( + rules=[ + Rule("send_email", "*", "deny"), + Rule("read", "*", "allow"), + ] + ) + mw = PermissionMiddleware(rulesets=[rs]) + state = { + "messages": [ + _msg( + {"name": "send_email", "args": {}, "id": "1"}, + {"name": "read", "args": {}, "id": "2"}, + ) + ] + } + out = mw.after_model(state, _FakeRuntime()) + assert out is not None + ai_msg = next(m for m in out["messages"] if isinstance(m, AIMessage)) + assert len(ai_msg.tool_calls) == 1 + assert ai_msg.tool_calls[0]["name"] == "read" + + +class TestAsk: + def test_reject_without_feedback_raises(self) -> None: + # Default: nothing matches -> ask + rs = Ruleset(rules=[]) + mw = PermissionMiddleware(rulesets=[rs]) + + # Bypass real interrupt — patch the helper + mw._raise_interrupt = lambda **kw: {"decision_type": "reject"} # type: ignore[assignment] + state = {"messages": [_msg({"name": "send_email", "args": {}, "id": "1"})]} + with pytest.raises(RejectedError): + mw.after_model(state, _FakeRuntime()) + + def test_reject_with_feedback_raises_corrected(self) -> None: + rs = Ruleset(rules=[]) + mw = PermissionMiddleware(rulesets=[rs]) + mw._raise_interrupt = lambda **kw: { # type: ignore[assignment] + "decision_type": "reject", + "feedback": "use a different subject line", + } + state = {"messages": [_msg({"name": "send_email", "args": {}, "id": "1"})]} + with pytest.raises(CorrectedError) as excinfo: + mw.after_model(state, _FakeRuntime()) + assert excinfo.value.feedback == "use a different subject line" + + def test_once_proceeds_without_persisting(self) -> None: + mw = PermissionMiddleware(rulesets=[]) + mw._raise_interrupt = lambda **kw: {"decision_type": "once"} # type: ignore[assignment] + state = {"messages": [_msg({"name": "send_email", "args": {}, "id": "1"})]} + out = mw.after_model(state, _FakeRuntime()) + # No state change because all calls kept + assert out is None + # No new rule persisted + assert mw._runtime_ruleset.rules == [] + + def test_always_persists_runtime_rule(self) -> None: + mw = PermissionMiddleware(rulesets=[]) + mw._raise_interrupt = lambda **kw: {"decision_type": "always"} # type: ignore[assignment] + state = {"messages": [_msg({"name": "send_email", "args": {}, "id": "1"})]} + out = mw.after_model(state, _FakeRuntime()) + assert out is None # call kept + # Runtime ruleset got the always-allow rule + new_rules = [r for r in mw._runtime_ruleset.rules if r.action == "allow"] + assert any( + r.permission == "send_email" for r in new_rules + ) diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_permissions.py b/surfsense_backend/tests/unit/agents/new_chat/test_permissions.py new file mode 100644 index 000000000..4924f2aee --- /dev/null +++ b/surfsense_backend/tests/unit/agents/new_chat/test_permissions.py @@ -0,0 +1,111 @@ +"""Tests for the wildcard matcher and rule evaluator (opencode evaluate.ts parity).""" + +from __future__ import annotations + +import pytest + +from app.agents.new_chat.permissions import ( + Rule, + Ruleset, + aggregate_action, + evaluate, + evaluate_many, + wildcard_match, +) + +pytestmark = pytest.mark.unit + + +class TestWildcardMatch: + @pytest.mark.parametrize( + "value,pattern,expected", + [ + ("edit", "edit", True), + ("edit", "*", True), + ("read", "edit", False), + ("/documents/secrets/x", "/documents/secrets/**", True), + # Single-segment glob: '*' does not cross '/' + ("/documents/secrets/x", "/documents/*/x", True), + ("/documents/foo/bar/x", "/documents/*/x", False), + ("/documents/foo/x", "/documents/*/x", True), + ("linear_create", "linear_*", True), + ("notion_create", "linear_*", False), + # ':' is not a separator, so '*' matches it + ("mcp:notion:create_page", "mcp:*", True), + ("mcp:notion:create_page", "mcp:**", True), + # But '/' IS a separator + ("foo/bar", "foo/*", True), + ("foo/bar/baz", "foo/*", False), + ], + ) + def test_match(self, value: str, pattern: str, expected: bool) -> None: + assert wildcard_match(value, pattern) is expected + + +class TestEvaluate: + def test_default_action_is_ask(self) -> None: + rule = evaluate("edit", "/foo/bar") + assert rule.action == "ask" + assert rule.permission == "edit" + + def test_last_match_wins(self) -> None: + rs = Ruleset( + rules=[ + Rule("edit", "*", "allow"), + Rule("edit", "/secrets/**", "deny"), + ] + ) + # Second rule (deny) is more specific AND specified later + assert evaluate("edit", "/secrets/x", rs).action == "deny" + # First rule (allow) covers the rest + assert evaluate("edit", "/public/x", rs).action == "allow" + + def test_layered_rulesets_later_overrides_earlier(self) -> None: + defaults = Ruleset(rules=[Rule("edit", "*", "ask")], origin="defaults") + space = Ruleset(rules=[Rule("edit", "*", "allow")], origin="space") + thread = Ruleset(rules=[Rule("edit", "*", "deny")], origin="thread") + # All three layered: thread wins + assert evaluate("edit", "x", defaults, space, thread).action == "deny" + # Without thread: space wins + assert evaluate("edit", "x", defaults, space).action == "allow" + + def test_permission_wildcard(self) -> None: + rs = Ruleset(rules=[Rule("linear_*", "*", "allow")]) + assert evaluate("linear_create_issue", "x", rs).action == "allow" + assert evaluate("notion_create", "x", rs).action == "ask" + + def test_pattern_wildcard(self) -> None: + rs = Ruleset(rules=[Rule("edit", "/documents/secrets/**", "deny")]) + assert evaluate("edit", "/documents/secrets/foo", rs).action == "deny" + assert evaluate("edit", "/documents/public/foo", rs).action == "ask" + + def test_evaluate_many(self) -> None: + rs = Ruleset( + rules=[ + Rule("edit", "*", "allow"), + Rule("edit", "/secrets/*", "deny"), + ] + ) + results = evaluate_many("edit", ["/public/x", "/secrets/y"], rs) + assert [r.action for r in results] == ["allow", "deny"] + + +class TestAggregateAction: + def test_any_deny_means_deny(self) -> None: + rules = [ + Rule("a", "*", "allow"), + Rule("a", "*", "deny"), + Rule("a", "*", "ask"), + ] + assert aggregate_action(rules) == "deny" + + def test_any_ask_means_ask_when_no_deny(self) -> None: + rules = [Rule("a", "*", "allow"), Rule("a", "*", "ask")] + assert aggregate_action(rules) == "ask" + + def test_all_allow_means_allow(self) -> None: + rules = [Rule("a", "*", "allow"), Rule("a", "*", "allow")] + assert aggregate_action(rules) == "allow" + + def test_empty_means_ask(self) -> None: + assert aggregate_action([]) == "ask" diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_plugin_loader.py b/surfsense_backend/tests/unit/agents/new_chat/test_plugin_loader.py new file mode 100644 index 000000000..8d98e1328 --- /dev/null +++ b/surfsense_backend/tests/unit/agents/new_chat/test_plugin_loader.py @@ -0,0 +1,187 @@ +"""Unit tests for the SurfSense plugin entry-point loader (Tier 6).""" + +from __future__ import annotations + +from unittest.mock import MagicMock, patch + +from langchain.agents.middleware import AgentMiddleware + +from app.agents.new_chat.plugin_loader import ( + PLUGIN_ENTRY_POINT_GROUP, + PluginContext, + load_allowed_plugin_names_from_env, + load_plugin_middlewares, +) +from app.agents.new_chat.plugins.year_substituter import ( + _YearSubstituterMiddleware, + make_middleware as year_substituter_factory, +) + +# --------------------------------------------------------------------------- +# Helpers +# --------------------------------------------------------------------------- + + +class _DummyMiddleware(AgentMiddleware): + """Trivial middleware used as the success-path return value.""" + + tools = () + + +def _ctx() -> PluginContext: + return PluginContext.build( + search_space_id=1, + user_id="u", + thread_visibility="PRIVATE", # type: ignore[arg-type] + llm=MagicMock(), + ) + + +class _FakeEntryPoint: + """Stand-in for ``importlib.metadata.EntryPoint``.""" + + def __init__(self, name: str, factory) -> None: + self.name = name + self._factory = factory + + def load(self): + return self._factory + + +# --------------------------------------------------------------------------- +# Loader behaviour +# --------------------------------------------------------------------------- + + +class TestPluginLoaderBasics: + def test_returns_empty_when_allowlist_is_empty(self) -> None: + assert load_plugin_middlewares(_ctx(), allowed_plugin_names=[]) == [] + + def test_skips_non_allowlisted_plugin(self) -> None: + called = [] + + def factory(_): # would be an obvious bug if called + called.append(True) + return _DummyMiddleware() + + ep = _FakeEntryPoint("dangerous_plugin", factory) + with patch( + "app.agents.new_chat.plugin_loader.entry_points", + return_value=[ep], + ): + result = load_plugin_middlewares(_ctx(), allowed_plugin_names=["allowed_only"]) + assert result == [] + assert not called + + def test_loads_allowlisted_plugin(self) -> None: + ep = _FakeEntryPoint("year_substituter", year_substituter_factory) + with patch( + "app.agents.new_chat.plugin_loader.entry_points", + return_value=[ep], + ): + result = load_plugin_middlewares( + _ctx(), allowed_plugin_names={"year_substituter"} + ) + assert len(result) == 1 + assert isinstance(result[0], _YearSubstituterMiddleware) + + +class TestPluginLoaderIsolation: + def test_factory_exception_is_isolated(self) -> None: + def crashing_factory(_): + raise RuntimeError("boom") + + ep = _FakeEntryPoint("buggy", crashing_factory) + with patch( + "app.agents.new_chat.plugin_loader.entry_points", + return_value=[ep], + ): + result = load_plugin_middlewares(_ctx(), allowed_plugin_names={"buggy"}) + assert result == [] # construction continued without the plugin + + def test_non_middleware_return_is_rejected(self) -> None: + def bad_factory(_): + return "not a middleware" + + ep = _FakeEntryPoint("liar", bad_factory) + with patch( + "app.agents.new_chat.plugin_loader.entry_points", + return_value=[ep], + ): + result = load_plugin_middlewares(_ctx(), allowed_plugin_names={"liar"}) + assert result == [] + + def test_load_phase_exception_is_isolated(self) -> None: + class _BrokenEP: + name = "broken" + + def load(self): + raise ImportError("cannot import") + + with patch( + "app.agents.new_chat.plugin_loader.entry_points", + return_value=[_BrokenEP()], + ): + result = load_plugin_middlewares(_ctx(), allowed_plugin_names={"broken"}) + assert result == [] + + def test_one_failure_does_not_block_others(self) -> None: + """Two plugins; one crashes during factory; the other still loads.""" + + def crashing_factory(_): + raise RuntimeError("boom") + + eps = [ + _FakeEntryPoint("crashing", crashing_factory), + _FakeEntryPoint("ok", year_substituter_factory), + ] + with patch( + "app.agents.new_chat.plugin_loader.entry_points", return_value=eps + ): + result = load_plugin_middlewares( + _ctx(), allowed_plugin_names={"crashing", "ok"} + ) + assert len(result) == 1 + assert isinstance(result[0], _YearSubstituterMiddleware) + + +class TestAllowlistEnv: + def test_empty_env_returns_empty_set(self, monkeypatch) -> None: + monkeypatch.delenv("SURFSENSE_ALLOWED_PLUGINS", raising=False) + assert load_allowed_plugin_names_from_env() == set() + + def test_parses_comma_separated_value(self, monkeypatch) -> None: + monkeypatch.setenv( + "SURFSENSE_ALLOWED_PLUGINS", " year_substituter , noisy , " + ) + assert load_allowed_plugin_names_from_env() == { + "year_substituter", + "noisy", + } + + +class TestPluginContext: + def test_build_includes_required_fields(self) -> None: + llm = MagicMock() + ctx = PluginContext.build( + search_space_id=42, + user_id="user-1", + thread_visibility="PRIVATE", # type: ignore[arg-type] + llm=llm, + ) + assert ctx["search_space_id"] == 42 + assert ctx["user_id"] == "user-1" + assert ctx["llm"] is llm + + def test_does_not_carry_secrets_or_db_session(self) -> None: + ctx = _ctx() + # If a future change tries to add these keys, this test will fail loudly. + for forbidden in ("api_key", "secret", "db_session", "session"): + assert forbidden not in ctx + + +class TestEntryPointGroup: + def test_group_name_matches_pyproject_convention(self) -> None: + # Plugins register under `surfsense.plugins`; this is part of our + # public contract for plugin authors. + assert PLUGIN_ENTRY_POINT_GROUP == "surfsense.plugins" diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_retry_after.py b/surfsense_backend/tests/unit/agents/new_chat/test_retry_after.py new file mode 100644 index 000000000..39dd9bf00 --- /dev/null +++ b/surfsense_backend/tests/unit/agents/new_chat/test_retry_after.py @@ -0,0 +1,107 @@ +"""Tests for RetryAfterMiddleware Retry-After parsing and retry decision logic.""" + +from __future__ import annotations + +import pytest + +from app.agents.new_chat.middleware.retry_after import ( + RetryAfterMiddleware, + _extract_retry_after_seconds, + _is_non_retryable, +) + +pytestmark = pytest.mark.unit + + +class _FakeResponse: + def __init__(self, headers: dict[str, str]) -> None: + self.headers = headers + + +class _FakeRateLimit(Exception): + def __init__(self, msg: str, headers: dict[str, str] | None = None) -> None: + super().__init__(msg) + if headers is not None: + self.response = _FakeResponse(headers) + + +class TestExtractRetryAfter: + def test_seconds_header(self) -> None: + exc = _FakeRateLimit("rate", {"Retry-After": "30"}) + assert _extract_retry_after_seconds(exc) == 30.0 + + def test_milliseconds_header_overrides_seconds(self) -> None: + exc = _FakeRateLimit("rate", {"retry-after-ms": "1500"}) + assert _extract_retry_after_seconds(exc) == 1.5 + + def test_case_insensitive(self) -> None: + exc = _FakeRateLimit("rate", {"RETRY-AFTER": "12"}) + assert _extract_retry_after_seconds(exc) == 12.0 + + def test_falls_back_to_message_regex(self) -> None: + exc = Exception("Please retry after 7 seconds") + assert _extract_retry_after_seconds(exc) == 7.0 + + def test_returns_none_when_no_hint(self) -> None: + exc = Exception("oops") + assert _extract_retry_after_seconds(exc) is None + + def test_handles_missing_headers_attr(self) -> None: + exc = ValueError("no headers") + assert _extract_retry_after_seconds(exc) is None + + +class TestIsNonRetryable: + @pytest.mark.parametrize( + "name", + ["ContextWindowExceededError", "AuthenticationError", "InvalidRequestError"], + ) + def test_non_retryable_classes(self, name: str) -> None: + cls = type(name, (Exception,), {}) + assert _is_non_retryable(cls("x")) is True + + def test_generic_exception_is_retryable(self) -> None: + assert _is_non_retryable(RuntimeError("transient")) is False + + +class TestDelayCalculation: + def test_takes_max_of_backoff_and_header(self) -> None: + mw = RetryAfterMiddleware(max_retries=3, initial_delay=1.0, jitter=False) + exc = _FakeRateLimit("rl", {"retry-after": "10"}) + delay = mw._delay_for_attempt(0, exc) + assert delay == pytest.approx(10.0) + + def test_uses_backoff_when_no_header(self) -> None: + mw = RetryAfterMiddleware( + max_retries=3, initial_delay=2.0, backoff_factor=2.0, jitter=False + ) + delay = mw._delay_for_attempt(2, RuntimeError("transient")) + # 2 * 2^2 = 8 + assert delay == pytest.approx(8.0) + + def test_caps_at_max_delay(self) -> None: + mw = RetryAfterMiddleware( + max_retries=3, + initial_delay=10.0, + backoff_factor=10.0, + max_delay=15.0, + jitter=False, + ) + delay = mw._delay_for_attempt(5, RuntimeError("x")) + assert delay <= 15.0 + + +class TestShouldRetry: + def test_default_retries_generic(self) -> None: + mw = RetryAfterMiddleware() + assert mw._should_retry(RuntimeError("transient")) is True + + def test_default_skips_non_retryable(self) -> None: + mw = RetryAfterMiddleware() + cls = type("ContextWindowExceededError", (Exception,), {}) + assert mw._should_retry(cls("too big")) is False + + def test_custom_retry_on(self) -> None: + mw = RetryAfterMiddleware(retry_on=lambda exc: isinstance(exc, ValueError)) + assert mw._should_retry(ValueError()) is True + assert mw._should_retry(KeyError()) is False diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_skills_backends.py b/surfsense_backend/tests/unit/agents/new_chat/test_skills_backends.py new file mode 100644 index 000000000..eb9cf396c --- /dev/null +++ b/surfsense_backend/tests/unit/agents/new_chat/test_skills_backends.py @@ -0,0 +1,242 @@ +"""Tests for the skills backends used by SurfSense's SkillsMiddleware.""" + +from __future__ import annotations + +import asyncio +from pathlib import Path + +import pytest + +from app.agents.new_chat.middleware.skills_backends import ( + SKILLS_BUILTIN_PREFIX, + SKILLS_SPACE_PREFIX, + BuiltinSkillsBackend, + SearchSpaceSkillsBackend, + build_skills_backend_factory, + default_skills_sources, +) + + +@pytest.fixture +def skills_root(tmp_path: Path) -> Path: + """Build a small synthetic skill-tree used by the tests.""" + root = tmp_path / "skills" + (root / "alpha").mkdir(parents=True) + (root / "alpha" / "SKILL.md").write_text( + "---\nname: alpha\ndescription: alpha skill\n---\n# Alpha\n" + ) + (root / "beta").mkdir(parents=True) + (root / "beta" / "SKILL.md").write_text( + "---\nname: beta\ndescription: beta skill\n---\n# Beta\n" + ) + (root / "_orphan_file.md").write_text("not a skill, just a stray file") + return root + + +class TestBuiltinSkillsBackendListing: + def test_lists_skill_directories_at_root(self, skills_root: Path) -> None: + backend = BuiltinSkillsBackend(skills_root) + infos = backend.ls_info("/") + names = {info["path"] for info in infos} + assert "/alpha" in names + assert "/beta" in names + assert "/_orphan_file.md" in names + for info in infos: + if info["path"] in {"/alpha", "/beta"}: + assert info["is_dir"] is True + + def test_lists_skill_md_under_skill_directory(self, skills_root: Path) -> None: + backend = BuiltinSkillsBackend(skills_root) + infos = backend.ls_info("/alpha") + paths = {info["path"] for info in infos} + assert paths == {"/alpha/SKILL.md"} + assert infos[0]["is_dir"] is False + assert infos[0]["size"] > 0 + + def test_returns_empty_for_missing_path(self, skills_root: Path) -> None: + backend = BuiltinSkillsBackend(skills_root) + assert backend.ls_info("/nonexistent") == [] + + def test_returns_empty_when_root_missing(self, tmp_path: Path) -> None: + backend = BuiltinSkillsBackend(tmp_path / "definitely-missing") + assert backend.ls_info("/") == [] + assert backend.download_files(["/x/SKILL.md"])[0].error == "file_not_found" + + def test_refuses_path_traversal(self, skills_root: Path) -> None: + backend = BuiltinSkillsBackend(skills_root) + assert backend.ls_info("/../../../etc") == [] + responses = backend.download_files(["/../../../etc/passwd"]) + assert responses[0].error == "invalid_path" + + +class TestBuiltinSkillsBackendDownload: + def test_downloads_skill_md_content(self, skills_root: Path) -> None: + backend = BuiltinSkillsBackend(skills_root) + responses = backend.download_files(["/alpha/SKILL.md", "/beta/SKILL.md"]) + assert len(responses) == 2 + assert responses[0].path == "/alpha/SKILL.md" + assert responses[0].content is not None + assert b"name: alpha" in responses[0].content + assert responses[1].error is None + + def test_marks_directory_as_is_directory_error(self, skills_root: Path) -> None: + backend = BuiltinSkillsBackend(skills_root) + responses = backend.download_files(["/alpha"]) + assert responses[0].error == "is_directory" + + def test_marks_missing_file_as_file_not_found(self, skills_root: Path) -> None: + backend = BuiltinSkillsBackend(skills_root) + responses = backend.download_files(["/alpha/missing.md"]) + assert responses[0].error == "file_not_found" + assert responses[0].content is None + + def test_response_path_matches_input_for_correlation( + self, skills_root: Path + ) -> None: + backend = BuiltinSkillsBackend(skills_root) + inputs = ["/alpha/SKILL.md", "/missing.md", "/beta/SKILL.md"] + responses = backend.download_files(inputs) + assert [r.path for r in responses] == inputs + + +class TestBuiltinSkillsBackendIntegration: + """Mirror the call sequence the SkillsMiddleware actually uses.""" + + def test_skills_middleware_call_pattern(self, skills_root: Path) -> None: + backend = BuiltinSkillsBackend(skills_root) + + infos = asyncio.run(backend.als_info("/")) + skill_dirs = [i["path"] for i in infos if i.get("is_dir")] + assert sorted(skill_dirs) == ["/alpha", "/beta"] + + skill_md_paths = [f"{p}/SKILL.md" for p in skill_dirs] + responses = asyncio.run(backend.adownload_files(skill_md_paths)) + assert all(r.error is None for r in responses) + assert all(r.content is not None for r in responses) + + +class TestBundledSkills: + def test_default_root_resolves_to_repo_skills_dir(self) -> None: + backend = BuiltinSkillsBackend() + assert backend.root.name == "builtin" + assert backend.root.parent.name == "skills" + + def test_bundled_starter_skills_are_present(self) -> None: + backend = BuiltinSkillsBackend() + infos = backend.ls_info("/") + names = {info["path"].lstrip("/") for info in infos if info.get("is_dir")} + # Five starter skills required by the Tier 4 plan. + for required in ( + "kb-research", + "report-writing", + "meeting-prep", + "slack-summary", + "email-drafting", + ): + assert required in names, f"missing starter skill: {required}" + + def test_each_starter_skill_has_valid_skill_md(self) -> None: + backend = BuiltinSkillsBackend() + infos = backend.ls_info("/") + skill_dirs = [info["path"] for info in infos if info.get("is_dir")] + for skill_dir in skill_dirs: + md_path = f"{skill_dir}/SKILL.md" + response = backend.download_files([md_path])[0] + assert response.error is None, f"missing SKILL.md in {skill_dir}" + content = response.content.decode("utf-8").replace("\r\n", "\n") + assert content.startswith("---\n"), f"missing frontmatter in {skill_dir}" + assert "\nname:" in content + assert "\ndescription:" in content + + +class _FakeKBBackend: + """Stand-in for :class:`KBPostgresBackend` with the two methods we need.""" + + def __init__(self, listing: list[dict], file_contents: dict[str, bytes]) -> None: + self._listing = listing + self._file_contents = file_contents + self.last_ls_path: str | None = None + self.last_download_paths: list[str] | None = None + + async def als_info(self, path: str): + self.last_ls_path = path + return self._listing + + async def adownload_files(self, paths): + from deepagents.backends.protocol import FileDownloadResponse + + self.last_download_paths = list(paths) + out: list[FileDownloadResponse] = [] + for p in paths: + content = self._file_contents.get(p) + if content is None: + out.append(FileDownloadResponse(path=p, error="file_not_found")) + else: + out.append(FileDownloadResponse(path=p, content=content)) + return out + + +class TestSearchSpaceSkillsBackend: + def test_remaps_paths_when_listing(self) -> None: + listing = [ + {"path": "/documents/_skills/policy", "is_dir": True}, + {"path": "/documents/_skills/policy/SKILL.md", "is_dir": False}, + {"path": "/documents/other-folder/x.md", "is_dir": False}, + ] + kb = _FakeKBBackend(listing=listing, file_contents={}) + backend = SearchSpaceSkillsBackend(kb) + infos = asyncio.run(backend.als_info("/")) + assert kb.last_ls_path == "/documents/_skills" + paths = [info["path"] for info in infos] + assert "/policy" in paths + assert "/policy/SKILL.md" in paths + # Unrelated KB documents must NOT leak into the skills namespace. + assert all(not p.startswith("/documents") for p in paths) + + def test_remaps_paths_when_downloading(self) -> None: + kb = _FakeKBBackend( + listing=[], + file_contents={ + "/documents/_skills/policy/SKILL.md": b"---\nname: policy\n---\n", + }, + ) + backend = SearchSpaceSkillsBackend(kb) + responses = asyncio.run(backend.adownload_files(["/policy/SKILL.md"])) + assert kb.last_download_paths == ["/documents/_skills/policy/SKILL.md"] + assert responses[0].path == "/policy/SKILL.md" + assert responses[0].error is None + assert responses[0].content is not None + + def test_sync_methods_raise_not_implemented(self) -> None: + backend = SearchSpaceSkillsBackend(_FakeKBBackend([], {})) + with pytest.raises(NotImplementedError): + backend.ls_info("/") + with pytest.raises(NotImplementedError): + backend.download_files(["/x"]) + + def test_custom_kb_root_is_honored(self) -> None: + kb = _FakeKBBackend( + listing=[ + {"path": "/skills_admin/x", "is_dir": True}, + ], + file_contents={}, + ) + backend = SearchSpaceSkillsBackend(kb, kb_root="/skills_admin") + infos = asyncio.run(backend.als_info("/")) + assert kb.last_ls_path == "/skills_admin" + assert infos[0]["path"] == "/x" + + +class TestBackendFactory: + def test_builtin_only_factory_returns_composite(self) -> None: + factory = build_skills_backend_factory() + backend = factory(runtime=None) # type: ignore[arg-type] + from deepagents.backends.composite import CompositeBackend + + assert isinstance(backend, CompositeBackend) + assert SKILLS_BUILTIN_PREFIX in backend.routes + assert SKILLS_SPACE_PREFIX not in backend.routes + + def test_default_skills_sources_lists_builtin_then_space(self) -> None: + sources = default_skills_sources() + assert sources == [SKILLS_BUILTIN_PREFIX, SKILLS_SPACE_PREFIX] diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_specialized_subagents.py b/surfsense_backend/tests/unit/agents/new_chat/test_specialized_subagents.py new file mode 100644 index 000000000..3819b4605 --- /dev/null +++ b/surfsense_backend/tests/unit/agents/new_chat/test_specialized_subagents.py @@ -0,0 +1,338 @@ +"""Tests for the specialized subagents (explore / report_writer / connector_negotiator).""" + +from __future__ import annotations + +from langchain_core.tools import tool + +from app.agents.new_chat.middleware.permission import PermissionMiddleware +from app.agents.new_chat.subagents import ( + build_connector_negotiator_subagent, + build_explore_subagent, + build_report_writer_subagent, + build_specialized_subagents, +) +from app.agents.new_chat.subagents.config import ( + EXPLORE_READ_TOOLS, + REPORT_WRITER_TOOLS, + WRITE_TOOL_DENY_PATTERNS, +) + +# --------------------------------------------------------------------------- +# Fake tools used to verify filtering & permission behavior +# --------------------------------------------------------------------------- + + +@tool +def search_surfsense_docs(query: str) -> str: + """Search the user's KB.""" + return "" + + +@tool +def web_search(query: str) -> str: + """Search the public web.""" + return "" + + +@tool +def scrape_webpage(url: str) -> str: + """Scrape a single webpage.""" + return "" + + +@tool +def read_file(path: str) -> str: + """Read a file.""" + return "" + + +@tool +def ls_tree(path: str) -> str: + """List a tree.""" + return "" + + +@tool +def grep(pattern: str) -> str: + """Grep.""" + return "" + + +@tool +def update_memory(content: str) -> str: + """Update the user's memory.""" + return "" + + +@tool +def edit_file(path: str, old: str, new: str) -> str: + """Edit a file.""" + return "" + + +@tool +def linear_create_issue(title: str) -> str: + """Create a Linear issue.""" + return "" + + +@tool +def slack_send_message(channel: str, text: str) -> str: + """Send a Slack message.""" + return "" + + +@tool +def get_connected_accounts() -> str: + """List connected accounts.""" + return "" + + +@tool +def generate_report(topic: str) -> str: + """Generate a report artifact.""" + return "" + + +ALL_TOOLS = [ + search_surfsense_docs, + web_search, + scrape_webpage, + read_file, + ls_tree, + grep, + update_memory, + edit_file, + linear_create_issue, + slack_send_message, + get_connected_accounts, + generate_report, +] + + +class TestExploreSubagent: + def test_only_read_tools_are_exposed(self) -> None: + spec = build_explore_subagent(tools=ALL_TOOLS) + names = {t.name for t in spec["tools"]} # type: ignore[index] + assert names == EXPLORE_READ_TOOLS & {t.name for t in ALL_TOOLS} + assert "update_memory" not in names + assert "linear_create_issue" not in names + assert "edit_file" not in names + + def test_includes_permission_middleware_with_deny_rules(self) -> None: + spec = build_explore_subagent(tools=ALL_TOOLS) + permission_mws = [ + m for m in spec["middleware"] if isinstance(m, PermissionMiddleware) # type: ignore[index] + ] + assert len(permission_mws) == 1 + ruleset = permission_mws[0]._static_rulesets[0] + assert ruleset.origin == "subagent_explore" + deny_patterns = {r.permission for r in ruleset.rules if r.action == "deny"} + assert "update_memory" in deny_patterns + assert "edit_file" in deny_patterns + assert "*create*" in deny_patterns + assert "*send*" in deny_patterns + + def test_skills_inherits_default_sources(self) -> None: + spec = build_explore_subagent(tools=ALL_TOOLS) + assert spec["skills"] == ["/skills/builtin/", "/skills/space/"] # type: ignore[index] + + def test_name_and_description_match_contract(self) -> None: + spec = build_explore_subagent(tools=ALL_TOOLS) + assert spec["name"] == "explore" + assert "read-only" in spec["description"].lower() + + def test_includes_dedup_and_patch_middleware(self) -> None: + from deepagents.middleware.patch_tool_calls import PatchToolCallsMiddleware + + from app.agents.new_chat.middleware import DedupHITLToolCallsMiddleware + + spec = build_explore_subagent(tools=ALL_TOOLS) + types = {type(m) for m in spec["middleware"]} # type: ignore[index] + assert PatchToolCallsMiddleware in types + assert DedupHITLToolCallsMiddleware in types + + +class TestReportWriterSubagent: + def test_exposes_only_report_writing_tools(self) -> None: + spec = build_report_writer_subagent(tools=ALL_TOOLS) + names = {t.name for t in spec["tools"]} # type: ignore[index] + assert names == REPORT_WRITER_TOOLS & {t.name for t in ALL_TOOLS} + assert "generate_report" in names + assert "search_surfsense_docs" in names + + def test_deny_rules_block_writes_but_allow_generate_report(self) -> None: + spec = build_report_writer_subagent(tools=ALL_TOOLS) + permission_mws = [ + m for m in spec["middleware"] if isinstance(m, PermissionMiddleware) # type: ignore[index] + ] + ruleset = permission_mws[0]._static_rulesets[0] + deny_patterns = {r.permission for r in ruleset.rules if r.action == "deny"} + assert "update_memory" in deny_patterns + # generate_report MUST not be denied — it's the whole point of the subagent. + assert "generate_report" not in deny_patterns + # No deny pattern should match `generate_report` either. + assert all( + not _wildcard_matches(pattern, "generate_report") + for pattern in deny_patterns + ) + + +class TestConnectorNegotiatorSubagent: + def test_inherits_all_parent_tools(self) -> None: + spec = build_connector_negotiator_subagent(tools=ALL_TOOLS) + names = {t.name for t in spec["tools"]} # type: ignore[index] + # Every parent tool is inherited; the deny ruleset enforces behavior + # at execution time instead of trimming the tool list. + assert names == {t.name for t in ALL_TOOLS} + + def test_get_connected_accounts_is_present(self) -> None: + spec = build_connector_negotiator_subagent(tools=ALL_TOOLS) + names = {t.name for t in spec["tools"]} # type: ignore[index] + assert "get_connected_accounts" in names + + def test_deny_ruleset_blocks_mutating_connector_tools(self) -> None: + spec = build_connector_negotiator_subagent(tools=ALL_TOOLS) + permission_mws = [ + m for m in spec["middleware"] if isinstance(m, PermissionMiddleware) # type: ignore[index] + ] + ruleset = permission_mws[0]._static_rulesets[0] + deny_patterns = {r.permission for r in ruleset.rules if r.action == "deny"} + # `linear_create_issue` matches the `*_create` deny pattern. + assert any( + _wildcard_matches(p, "linear_create_issue") for p in deny_patterns + ) + assert any( + _wildcard_matches(p, "slack_send_message") for p in deny_patterns + ) + + +class TestBuildSpecializedSubagents: + def test_returns_three_specs(self) -> None: + specs = build_specialized_subagents(tools=ALL_TOOLS) + names = [s["name"] for s in specs] # type: ignore[index] + assert names == ["explore", "report_writer", "connector_negotiator"] + + def test_all_specs_have_unique_names(self) -> None: + specs = build_specialized_subagents(tools=ALL_TOOLS) + names = [s["name"] for s in specs] # type: ignore[index] + assert len(set(names)) == len(names) + + def test_extra_middleware_is_prepended_to_each_spec(self) -> None: + """Sentinel middleware passed via ``extra_middleware`` must appear + in each subagent's ``middleware`` list, before the local rules. + + This guards against the regression where specialized subagents + promised filesystem tools (``read_file``, ``ls``, ``grep``) in + their system prompts but had no filesystem middleware mounted. + """ + + class _Sentinel: + pass + + sentinel = _Sentinel() + specs = build_specialized_subagents( + tools=ALL_TOOLS, extra_middleware=[sentinel] + ) + for spec in specs: + mws = spec["middleware"] # type: ignore[index] + assert sentinel in mws + # The sentinel must appear *before* the permission middleware + # (subagent-local rules), preserving the documented composition + # order: extra → custom → patch → dedup. + sentinel_idx = mws.index(sentinel) + perm_idx = next( + (i for i, m in enumerate(mws) + if isinstance(m, PermissionMiddleware)), + None, + ) + assert perm_idx is not None + assert sentinel_idx < perm_idx + + +class TestFilterToolsWarningSuppression: + """Names provided by middleware (read_file, ls, grep, …) must not + trigger the spurious "missing" warning in :func:`_filter_tools`.""" + + def test_middleware_provided_names_are_silent(self, caplog) -> None: + import logging + + from app.agents.new_chat.subagents.config import _filter_tools + + with caplog.at_level(logging.INFO, logger="app.agents.new_chat.subagents.config"): + # Allowed set asks for two registry tools (one present, one + # not) plus a bunch of middleware-provided names. + _filter_tools( + [search_surfsense_docs], + allowed_names={ + "search_surfsense_docs", + "scrape_webpage", # legitimately missing → should warn + "read_file", # mw-provided → suppressed + "ls", + "grep", + "glob", + "write_todos", + }, + ) + + warnings = [ + r.message for r in caplog.records if r.levelno >= logging.INFO + ] + # Exactly one warning, and it should mention scrape_webpage but not + # any middleware-provided name. Inspect the rendered "missing" + # list (between the brackets) so we don't false-match substrings + # like ``ls`` inside ``available``. + assert len(warnings) == 1, warnings + msg = warnings[0] + assert "scrape_webpage" in msg + bracket_section = msg.split("missing: ", 1)[1] + for noisy in ("read_file", "ls", "grep", "glob", "write_todos"): + assert f"'{noisy}'" not in bracket_section, msg + + +class TestDenyPatternsCoverage: + def test_deny_patterns_cover_canonical_write_tools(self) -> None: + canonical_writes = [ + "update_memory", + "edit_file", + "write_file", + "move_file", + "mkdir", + "linear_create_issue", + "linear_update_issue", + "linear_delete_issue", + "slack_send_message", + "create_index", + "update_account", + "delete_record", + "send_email", + ] + for tool_name in canonical_writes: + assert any( + _wildcard_matches(pattern, tool_name) + for pattern in WRITE_TOOL_DENY_PATTERNS + ), f"no deny pattern matches {tool_name!r}" + + def test_deny_patterns_do_not_match_safe_read_tools(self) -> None: + canonical_reads = [ + "search_surfsense_docs", + "read_file", + "ls_tree", + "grep", + "web_search", + "scrape_webpage", + "get_connected_accounts", + "generate_report", + ] + for tool_name in canonical_reads: + assert not any( + _wildcard_matches(pattern, tool_name) + for pattern in WRITE_TOOL_DENY_PATTERNS + ), f"deny pattern incorrectly matches read tool {tool_name!r}" + + +def _wildcard_matches(pattern: str, value: str) -> bool: + """Helper using the same matcher the rule evaluator does.""" + from app.agents.new_chat.permissions import wildcard_match + + return wildcard_match(value, pattern) diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_tool_call_repair.py b/surfsense_backend/tests/unit/agents/new_chat/test_tool_call_repair.py new file mode 100644 index 000000000..f792aef60 --- /dev/null +++ b/surfsense_backend/tests/unit/agents/new_chat/test_tool_call_repair.py @@ -0,0 +1,103 @@ +"""Tests for ToolCallNameRepairMiddleware.""" + +from __future__ import annotations + +import pytest +from langchain_core.messages import AIMessage + +from app.agents.new_chat.middleware.tool_call_repair import ( + ToolCallNameRepairMiddleware, +) +from app.agents.new_chat.tools.invalid_tool import INVALID_TOOL_NAME + +pytestmark = pytest.mark.unit + + +def _make_state(message: AIMessage) -> dict: + return {"messages": [message]} + + +class _FakeRuntime: + def __init__(self, context: object | None = None) -> None: + self.context = context + + +class TestRepair: + def test_passthrough_when_name_matches(self) -> None: + mw = ToolCallNameRepairMiddleware( + registered_tool_names={"echo"}, fuzzy_match_threshold=None + ) + msg = AIMessage(content="", tool_calls=[ + {"name": "echo", "args": {}, "id": "1"}, + ]) + out = mw.after_model(_make_state(msg), _FakeRuntime()) + assert out is None # no change + + def test_lowercase_repair(self) -> None: + mw = ToolCallNameRepairMiddleware( + registered_tool_names={"echo"}, fuzzy_match_threshold=None + ) + msg = AIMessage(content="", tool_calls=[ + {"name": "Echo", "args": {"x": 1}, "id": "1"}, + ]) + out = mw.after_model(_make_state(msg), _FakeRuntime()) + assert out is not None + repaired = out["messages"][0] + assert repaired.tool_calls[0]["name"] == "echo" + + def test_invalid_fallback_when_no_match(self) -> None: + mw = ToolCallNameRepairMiddleware( + registered_tool_names={"echo", INVALID_TOOL_NAME}, + fuzzy_match_threshold=None, + ) + msg = AIMessage(content="", tool_calls=[ + {"name": "totally_different_name", "args": {"k": "v"}, "id": "1"}, + ]) + out = mw.after_model(_make_state(msg), _FakeRuntime()) + assert out is not None + repaired_call = out["messages"][0].tool_calls[0] + assert repaired_call["name"] == INVALID_TOOL_NAME + assert repaired_call["args"]["tool"] == "totally_different_name" + assert "totally_different_name" in repaired_call["args"]["error"] + + def test_no_invalid_means_skip_when_unknown(self) -> None: + mw = ToolCallNameRepairMiddleware( + registered_tool_names={"echo"}, fuzzy_match_threshold=None + ) + msg = AIMessage(content="", tool_calls=[ + {"name": "unknown", "args": {}, "id": "1"}, + ]) + out = mw.after_model(_make_state(msg), _FakeRuntime()) + # No repair available; original returned unchanged (no update) + assert out is None + + def test_fuzzy_match_works_when_enabled(self) -> None: + mw = ToolCallNameRepairMiddleware( + registered_tool_names={"search_documents"}, + fuzzy_match_threshold=0.7, + ) + msg = AIMessage(content="", tool_calls=[ + {"name": "search_docments", "args": {}, "id": "1"}, + ]) + out = mw.after_model(_make_state(msg), _FakeRuntime()) + assert out is not None + assert out["messages"][0].tool_calls[0]["name"] == "search_documents" + + def test_skips_when_no_messages(self) -> None: + mw = ToolCallNameRepairMiddleware(registered_tool_names={"echo"}) + out = mw.after_model({"messages": []}, _FakeRuntime()) + assert out is None + + def test_runtime_context_extends_registered(self) -> None: + from types import SimpleNamespace + + mw = ToolCallNameRepairMiddleware( + registered_tool_names={"echo"}, fuzzy_match_threshold=None + ) + msg = AIMessage(content="", tool_calls=[ + {"name": "DynamicTool", "args": {}, "id": "1"}, + ]) + runtime = _FakeRuntime(SimpleNamespace(registered_tool_names=["dynamictool"])) + out = mw.after_model(_make_state(msg), runtime) + assert out is not None + assert out["messages"][0].tool_calls[0]["name"] == "dynamictool" diff --git a/surfsense_backend/tests/unit/middleware/test_dedup_hitl_tool_calls.py b/surfsense_backend/tests/unit/middleware/test_dedup_hitl_tool_calls.py index add0105e4..467ba6d5f 100644 --- a/surfsense_backend/tests/unit/middleware/test_dedup_hitl_tool_calls.py +++ b/surfsense_backend/tests/unit/middleware/test_dedup_hitl_tool_calls.py @@ -1,8 +1,10 @@ import pytest from langchain_core.messages import AIMessage +from langchain_core.tools import StructuredTool from app.agents.new_chat.middleware.dedup_tool_calls import ( DedupHITLToolCallsMiddleware, + wrap_dedup_key_by_arg_name, ) pytestmark = pytest.mark.unit @@ -14,9 +16,34 @@ def _make_state(tool_calls: list[dict]) -> dict: return {"messages": [msg]} +def _hitl_tool(name: str, *, dedup_arg: str) -> StructuredTool: + """Build a tool with declarative ``dedup_key`` metadata. + + Mirrors the ``ToolDefinition.dedup_key`` -> ``tool.metadata["dedup_key"]`` + propagation done by :func:`build_tools` after the cleanup tier. + """ + + def _fn(**kwargs): + return "ok" + + return StructuredTool.from_function( + func=_fn, + name=name, + description="x", + metadata={"dedup_key": wrap_dedup_key_by_arg_name(dedup_arg)}, + ) + + def test_duplicate_hitl_calls_reduced_to_first(): - """When the LLM emits the same HITL tool call twice, only the first is kept.""" - mw = DedupHITLToolCallsMiddleware() + """When the LLM emits the same HITL tool call twice, only the first is kept. + + After the cleanup tier removed ``_NATIVE_HITL_TOOL_DEDUP_KEYS``, the + resolver is sourced from ``ToolDefinition.dedup_key`` propagated onto + ``tool.metadata`` — which the registry does at agent build time. The + test mirrors that wiring with an in-memory tool. + """ + tool = _hitl_tool("delete_calendar_event", dedup_arg="event_title_or_id") + mw = DedupHITLToolCallsMiddleware(agent_tools=[tool]) state = _make_state( [ diff --git a/surfsense_backend/tests/unit/observability/__init__.py b/surfsense_backend/tests/unit/observability/__init__.py new file mode 100644 index 000000000..8b1378917 --- /dev/null +++ b/surfsense_backend/tests/unit/observability/__init__.py @@ -0,0 +1 @@ + diff --git a/surfsense_backend/tests/unit/observability/test_otel.py b/surfsense_backend/tests/unit/observability/test_otel.py new file mode 100644 index 000000000..583142098 --- /dev/null +++ b/surfsense_backend/tests/unit/observability/test_otel.py @@ -0,0 +1,84 @@ +"""Tests for the SurfSense OpenTelemetry shim (Tier 3b).""" + +from __future__ import annotations + +import pytest + +from app.observability import otel + +pytestmark = pytest.mark.unit + + +@pytest.fixture(autouse=True) +def _reset_otel_state(monkeypatch: pytest.MonkeyPatch): + """Force a clean OTel disabled state per test, then restore after.""" + for env in ("OTEL_EXPORTER_OTLP_ENDPOINT", "SURFSENSE_DISABLE_OTEL"): + monkeypatch.delenv(env, raising=False) + monkeypatch.setenv("SURFSENSE_DISABLE_OTEL", "true") + otel.reload_for_tests() + yield + otel.reload_for_tests() + + +def test_disabled_by_default_when_no_endpoint() -> None: + assert otel.is_enabled() is False + + +def test_enabled_when_endpoint_configured(monkeypatch: pytest.MonkeyPatch) -> None: + monkeypatch.delenv("SURFSENSE_DISABLE_OTEL", raising=False) + monkeypatch.setenv("OTEL_EXPORTER_OTLP_ENDPOINT", "http://localhost:4317") + assert otel.reload_for_tests() is True + + +def test_kill_switch_overrides_endpoint(monkeypatch: pytest.MonkeyPatch) -> None: + monkeypatch.setenv("OTEL_EXPORTER_OTLP_ENDPOINT", "http://localhost:4317") + monkeypatch.setenv("SURFSENSE_DISABLE_OTEL", "true") + assert otel.reload_for_tests() is False + + +class TestNoopSpansWhenDisabled: + def test_generic_span_yields_noop(self) -> None: + with otel.span("any.thing", attributes={"x": 1}) as sp: + sp.set_attribute("y", 2) + sp.set_attributes({"a": "b"}) + sp.add_event("evt") + sp.record_exception(RuntimeError("ignored")) + sp.set_status("ignored") + # Reaching here without raising means the no-op is well-formed + + def test_exception_propagates_through_span(self) -> None: + with pytest.raises(ValueError), otel.span("err"): + raise ValueError("boom") + + def test_each_helper_is_a_no_op_when_disabled(self) -> None: + helpers = [ + otel.tool_call_span("write_file", input_size=42), + otel.model_call_span(model_id="openai:gpt-4o", provider="openai"), + otel.kb_search_span(search_space_id=1, query_chars=99), + otel.kb_persist_span(document_type="NOTE", document_id=7), + otel.compaction_span(reason="overflow", messages_in=120), + otel.interrupt_span(interrupt_type="permission_ask"), + otel.permission_asked_span(permission="edit", pattern="/x/**"), + ] + for cm in helpers: + with cm as sp: + assert sp is not None + sp.set_attribute("ok", True) + + +class TestEnabledIntegration: + """When OTel is wired but no SDK exporter is bound, the API still works.""" + + def test_span_attaches_attributes(self, monkeypatch: pytest.MonkeyPatch) -> None: + # Use the API tracer (no-op-ish but real Span objects). + monkeypatch.delenv("SURFSENSE_DISABLE_OTEL", raising=False) + monkeypatch.setenv("OTEL_EXPORTER_OTLP_ENDPOINT", "http://localhost:4317") + assert otel.reload_for_tests() is True + + # Should not raise even when set_attributes/record_exception fall through + # to an SDK that isn't actually installed. + with otel.tool_call_span("scrape_webpage", input_size=10) as sp: + sp.set_attribute("tool.output.size", 200) + sp.set_attribute("tool.truncated", False) + with otel.model_call_span(model_id="m", provider="p") as sp: + sp.set_attribute("retry.count", 3) diff --git a/surfsense_backend/tests/unit/services/test_revert_service.py b/surfsense_backend/tests/unit/services/test_revert_service.py new file mode 100644 index 000000000..cb8443291 --- /dev/null +++ b/surfsense_backend/tests/unit/services/test_revert_service.py @@ -0,0 +1,56 @@ +"""Unit tests for the agent revert service (Tier 5.3).""" + +from __future__ import annotations + +from typing import Any + +from app.services.revert_service import can_revert + + +class _FakeAction: + def __init__(self, *, user_id: Any, tool_name: str = "edit_file") -> None: + self.user_id = user_id + self.tool_name = tool_name + + +class TestCanRevert: + def test_owner_can_revert_their_own_action(self) -> None: + action = _FakeAction(user_id="user-123") + assert can_revert( + requester_user_id="user-123", action=action, is_admin=False + ) + + def test_other_user_cannot_revert(self) -> None: + action = _FakeAction(user_id="user-123") + assert not can_revert( + requester_user_id="someone-else", action=action, is_admin=False + ) + + def test_admin_always_allowed(self) -> None: + action = _FakeAction(user_id="user-123") + assert can_revert( + requester_user_id="anybody", action=action, is_admin=True + ) + + def test_admin_can_revert_anonymous_action(self) -> None: + action = _FakeAction(user_id=None) + assert can_revert( + requester_user_id="admin", action=action, is_admin=True + ) + + def test_anonymous_action_blocks_non_admin(self) -> None: + action = _FakeAction(user_id=None) + assert not can_revert( + requester_user_id="user-1", action=action, is_admin=False + ) + + def test_uuid_string_normalization(self) -> None: + """``user_id`` may be a UUID object; comparison should still work.""" + import uuid + + u = uuid.uuid4() + action = _FakeAction(user_id=u) + # Same UUID, passed as string from the requesting side. + assert can_revert( + requester_user_id=str(u), action=action, is_admin=False + ) diff --git a/surfsense_web/app/dashboard/[search_space_id]/user-settings/components/AgentPermissionsContent.tsx b/surfsense_web/app/dashboard/[search_space_id]/user-settings/components/AgentPermissionsContent.tsx new file mode 100644 index 000000000..b01f556ad --- /dev/null +++ b/surfsense_web/app/dashboard/[search_space_id]/user-settings/components/AgentPermissionsContent.tsx @@ -0,0 +1,451 @@ +"use client"; + +import { useMutation, useQuery, useQueryClient } from "@tanstack/react-query"; +import { useAtomValue } from "jotai"; +import { AlertTriangle, Check, Plus, ShieldCheck, Trash2, X } from "lucide-react"; +import { useCallback, useMemo, useState } from "react"; +import { toast } from "sonner"; +import { agentFlagsAtom } from "@/atoms/agent/agent-flags-query.atom"; +import { activeSearchSpaceIdAtom } from "@/atoms/search-spaces/search-space-query.atoms"; +import { Alert, AlertDescription, AlertTitle } from "@/components/ui/alert"; +import { + AlertDialog, + AlertDialogAction, + AlertDialogCancel, + AlertDialogContent, + AlertDialogDescription, + AlertDialogFooter, + AlertDialogHeader, + AlertDialogTitle, +} from "@/components/ui/alert-dialog"; +import { Badge } from "@/components/ui/badge"; +import { Button } from "@/components/ui/button"; +import { Input } from "@/components/ui/input"; +import { Label } from "@/components/ui/label"; +import { + Select, + SelectContent, + SelectItem, + SelectTrigger, + SelectValue, +} from "@/components/ui/select"; +import { Spinner } from "@/components/ui/spinner"; +import { + type AgentPermissionAction, + type AgentPermissionRule, + type AgentPermissionRuleCreate, + agentPermissionsApiService, +} from "@/lib/apis/agent-permissions-api.service"; +import { AppError } from "@/lib/error"; +import { formatRelativeDate } from "@/lib/format-date"; +import { cn } from "@/lib/utils"; + +const ACTION_DESCRIPTIONS: Record = { + allow: "Always run without prompting", + deny: "Block silently", + ask: "Pause and ask for approval", +}; + +const ACTION_BADGE: Record = { + allow: { label: "Allow", className: "bg-emerald-500/10 text-emerald-600 border-emerald-500/30" }, + deny: { label: "Deny", className: "bg-destructive/10 text-destructive border-destructive/30" }, + ask: { label: "Ask", className: "bg-amber-500/10 text-amber-600 border-amber-500/30" }, +}; + +const EMPTY_FORM: AgentPermissionRuleCreate = { + permission: "", + pattern: "*", + action: "ask", + user_id: null, + thread_id: null, +}; + +function permissionRulesQueryKey(searchSpaceId: number) { + return ["agent-permission-rules", searchSpaceId] as const; +} + +function ScopeBadge({ rule }: { rule: AgentPermissionRule }) { + if (rule.thread_id !== null) { + return ( + + Thread #{rule.thread_id} + + ); + } + if (rule.user_id !== null) { + return ( + + User-specific + + ); + } + return ( + + Search space + + ); +} + +export function AgentPermissionsContent() { + const searchSpaceIdRaw = useAtomValue(activeSearchSpaceIdAtom); + const searchSpaceId = searchSpaceIdRaw ? Number(searchSpaceIdRaw) : null; + + const { data: flags } = useAtomValue(agentFlagsAtom); + const featureEnabled = !!flags?.enable_permission && !flags?.disable_new_agent_stack; + + const queryClient = useQueryClient(); + + const { + data: rules, + isLoading, + isError, + error, + } = useQuery({ + queryKey: searchSpaceId + ? permissionRulesQueryKey(searchSpaceId) + : ["agent-permission-rules", "none"], + queryFn: () => agentPermissionsApiService.list(searchSpaceId as number), + enabled: !!searchSpaceId && featureEnabled, + staleTime: 60 * 1000, + }); + + const createMutation = useMutation({ + mutationFn: (payload: AgentPermissionRuleCreate) => + agentPermissionsApiService.create(searchSpaceId as number, payload), + onSuccess: () => { + toast.success("Rule created."); + queryClient.invalidateQueries({ + queryKey: permissionRulesQueryKey(searchSpaceId as number), + }); + }, + onError: (err: unknown) => { + toast.error(err instanceof Error ? err.message : "Failed to create rule."); + }, + }); + + const updateMutation = useMutation({ + mutationFn: (params: { ruleId: number; action: AgentPermissionAction; pattern?: string }) => + agentPermissionsApiService.update(searchSpaceId as number, params.ruleId, { + action: params.action, + pattern: params.pattern, + }), + onSuccess: () => { + queryClient.invalidateQueries({ + queryKey: permissionRulesQueryKey(searchSpaceId as number), + }); + }, + onError: (err: unknown) => { + toast.error(err instanceof Error ? err.message : "Failed to update rule."); + }, + }); + + const deleteMutation = useMutation({ + mutationFn: (ruleId: number) => + agentPermissionsApiService.remove(searchSpaceId as number, ruleId), + onSuccess: () => { + toast.success("Rule deleted."); + queryClient.invalidateQueries({ + queryKey: permissionRulesQueryKey(searchSpaceId as number), + }); + }, + onError: (err: unknown) => { + toast.error(err instanceof Error ? err.message : "Failed to delete rule."); + }, + }); + + const [showForm, setShowForm] = useState(false); + const [formData, setFormData] = useState(EMPTY_FORM); + const [deleteTarget, setDeleteTarget] = useState(null); + + const sortedRules = useMemo(() => rules ?? [], [rules]); + + const handleCreate = useCallback(async () => { + if (!formData.permission.trim()) { + toast.error("Permission is required."); + return; + } + try { + await createMutation.mutateAsync({ + ...formData, + permission: formData.permission.trim(), + pattern: formData.pattern.trim() || "*", + }); + setShowForm(false); + setFormData(EMPTY_FORM); + } catch (err) { + if (err instanceof AppError && err.message) { + // already toasted by onError + } + } + }, [createMutation, formData]); + + const handleConfirmDelete = useCallback(async () => { + if (deleteTarget === null) return; + try { + await deleteMutation.mutateAsync(deleteTarget); + } finally { + setDeleteTarget(null); + } + }, [deleteMutation, deleteTarget]); + + if (!featureEnabled) { + return ( + + + Permission middleware is disabled + + Flip{" "} + SURFSENSE_ENABLE_PERMISSION on + the backend to manage allow/deny/ask rules from this panel. + + + ); + } + + if (!searchSpaceId) { + return ( +

Open a search space to manage agent rules.

+ ); + } + + if (isLoading) { + return ( +
+ +
+ ); + } + + if (isError) { + return ( +
+ +

Failed to load rules

+

+ {error instanceof Error ? error.message : "Unknown error."} +

+
+ ); + } + + return ( +
+
+
+

+ Tell the agent which tools to allow, deny, or ask before running. Rules use wildcard + patterns and are evaluated at the most specific scope first. +

+
+ {!showForm && ( + + )} +
+ + {showForm && ( +
+
+

New permission rule

+ +
+
+ + setFormData((p) => ({ ...p, permission: e.target.value }))} + /> +

+ Match a tool capability. Use * for wildcards. +

+
+ +
+ + setFormData((p) => ({ ...p, pattern: e.target.value }))} + /> +

+ Wildcard against the canonical argument (e.g. prod-*). +

+
+
+ +
+ + +

+ {ACTION_DESCRIPTIONS[formData.action]} +

+
+ +
+ + +
+
+
+ )} + + {sortedRules.length === 0 && !showForm && ( +
+ +

No rules yet

+

+ Without rules the agent uses the deployment default for every tool. +

+
+ )} + + {sortedRules.length > 0 && ( +
+ {sortedRules.map((rule) => { + const badge = ACTION_BADGE[rule.action]; + const isUpdating = + updateMutation.isPending && updateMutation.variables?.ruleId === rule.id; + const isDeleting = deleteMutation.isPending && deleteMutation.variables === rule.id; + + return ( +
+
+
+
+ + {rule.permission} + + {rule.pattern !== "*" && ( + + → {rule.pattern} + + )} + +
+

+ Created {formatRelativeDate(rule.created_at)} +

+
+ +
+ + + +
+
+
+ ); + })} +
+ )} + + !open && setDeleteTarget(null)} + > + + + Delete this rule? + + The agent will fall back to deployment defaults for matching tool calls. + + + + Cancel + { + e.preventDefault(); + handleConfirmDelete(); + }} + disabled={deleteMutation.isPending} + > + {deleteMutation.isPending ? "Deleting…" : "Delete"} + + + + +
+ ); +} diff --git a/surfsense_web/app/dashboard/[search_space_id]/user-settings/components/AgentStatusContent.tsx b/surfsense_web/app/dashboard/[search_space_id]/user-settings/components/AgentStatusContent.tsx new file mode 100644 index 000000000..bd8f03a70 --- /dev/null +++ b/surfsense_web/app/dashboard/[search_space_id]/user-settings/components/AgentStatusContent.tsx @@ -0,0 +1,309 @@ +"use client"; + +import { useAtomValue } from "jotai"; +import { CircleCheck, CircleSlash, Cog, RotateCcw } from "lucide-react"; +import { useMemo } from "react"; +import { agentFlagsAtom } from "@/atoms/agent/agent-flags-query.atom"; +import { Alert, AlertDescription, AlertTitle } from "@/components/ui/alert"; +import { Badge } from "@/components/ui/badge"; +import { Separator } from "@/components/ui/separator"; +import { Skeleton } from "@/components/ui/skeleton"; +import type { AgentFeatureFlags } from "@/lib/apis/agent-flags-api.service"; +import { cn } from "@/lib/utils"; + +type FlagKey = keyof AgentFeatureFlags; + +interface FlagDef { + key: FlagKey; + label: string; + description: string; + envVar: string; +} + +interface FlagGroup { + id: string; + title: string; + subtitle: string; + flags: FlagDef[]; +} + +const FLAG_GROUPS: FlagGroup[] = [ + { + id: "tier1", + title: "Tier 1 — Agent quality", + subtitle: "Context editing, retries, fallbacks, doom-loop, tool-call repair.", + flags: [ + { + key: "enable_context_editing", + label: "Context editing", + description: "Trim tool outputs and spill old text into backend storage.", + envVar: "SURFSENSE_ENABLE_CONTEXT_EDITING", + }, + { + key: "enable_compaction_v2", + label: "Compaction v2", + description: "SurfSense-aware compaction replacing safe summarization.", + envVar: "SURFSENSE_ENABLE_COMPACTION_V2", + }, + { + key: "enable_retry_after", + label: "Retry-After", + description: "Honour rate-limit retry-after headers automatically.", + envVar: "SURFSENSE_ENABLE_RETRY_AFTER", + }, + { + key: "enable_model_fallback", + label: "Model fallback", + description: "Fail over to a backup model on persistent errors.", + envVar: "SURFSENSE_ENABLE_MODEL_FALLBACK", + }, + { + key: "enable_model_call_limit", + label: "Model call limit", + description: "Cap total model calls per turn to prevent budget run-aways.", + envVar: "SURFSENSE_ENABLE_MODEL_CALL_LIMIT", + }, + { + key: "enable_tool_call_limit", + label: "Tool call limit", + description: "Cap total tool calls per turn.", + envVar: "SURFSENSE_ENABLE_TOOL_CALL_LIMIT", + }, + { + key: "enable_tool_call_repair", + label: "Tool-call name repair", + description: "Recover from lower-cased / fuzzy tool names emitted by smaller models.", + envVar: "SURFSENSE_ENABLE_TOOL_CALL_REPAIR", + }, + { + key: "enable_doom_loop", + label: "Doom-loop detection", + description: "Detect repeated identical tool calls and ask the user to confirm.", + envVar: "SURFSENSE_ENABLE_DOOM_LOOP", + }, + ], + }, + { + id: "tier2", + title: "Tier 2 — Safety", + subtitle: "Permission rules, busy-mutex, smarter tool selection.", + flags: [ + { + key: "enable_permission", + label: "Permission middleware", + description: "Apply allow/deny/ask rules from the Agent Permissions tab.", + envVar: "SURFSENSE_ENABLE_PERMISSION", + }, + { + key: "enable_busy_mutex", + label: "Busy mutex", + description: "Prevent two concurrent runs from corrupting the same thread.", + envVar: "SURFSENSE_ENABLE_BUSY_MUTEX", + }, + { + key: "enable_llm_tool_selector", + label: "LLM tool selector", + description: "Use a smaller model to pre-filter the tool list per turn.", + envVar: "SURFSENSE_ENABLE_LLM_TOOL_SELECTOR", + }, + ], + }, + { + id: "tier4", + title: "Tier 4 — Skills + subagents", + subtitle: "Built-in skills, specialized subagents, KB planner runnable.", + flags: [ + { + key: "enable_skills", + label: "Skills", + description: "Load on-demand skill packs (kb-research, report-writing, …).", + envVar: "SURFSENSE_ENABLE_SKILLS", + }, + { + key: "enable_specialized_subagents", + label: "Specialized subagents", + description: "Spin up explore / report_writer / connector_negotiator subagents.", + envVar: "SURFSENSE_ENABLE_SPECIALIZED_SUBAGENTS", + }, + { + key: "enable_kb_planner_runnable", + label: "KB planner runnable", + description: "Compile a private planner sub-agent for KB search.", + envVar: "SURFSENSE_ENABLE_KB_PLANNER_RUNNABLE", + }, + ], + }, + { + id: "tier5", + title: "Tier 5 — Audit + revert", + subtitle: "Action log + revert route used by the Agent Actions sheet.", + flags: [ + { + key: "enable_action_log", + label: "Action log", + description: "Persist every tool call to agent_action_log.", + envVar: "SURFSENSE_ENABLE_ACTION_LOG", + }, + { + key: "enable_revert_route", + label: "Revert route", + description: "Allow reverting reversible actions from the action log.", + envVar: "SURFSENSE_ENABLE_REVERT_ROUTE", + }, + ], + }, + { + id: "tier6", + title: "Tier 6 — Plugins", + subtitle: "Optional middleware loaded from entry points.", + flags: [ + { + key: "enable_plugin_loader", + label: "Plugin loader", + description: "Load surfsense.plugins entry-point middleware.", + envVar: "SURFSENSE_ENABLE_PLUGIN_LOADER", + }, + ], + }, + { + id: "obs", + title: "Observability", + subtitle: "Telemetry pipelines (orthogonal to feature gating).", + flags: [ + { + key: "enable_otel", + label: "OpenTelemetry", + description: "Emit OTel spans (also requires OTEL_EXPORTER_OTLP_ENDPOINT).", + envVar: "SURFSENSE_ENABLE_OTEL", + }, + ], + }, +]; + +function FlagRow({ def, value }: { def: FlagDef; value: boolean }) { + return ( +
+
+
+ {def.label} + + {def.envVar} + +
+

{def.description}

+
+ + {value ? : } + {value ? "On" : "Off"} + +
+ ); +} + +export function AgentStatusContent() { + const { data: flags, isLoading, isError, error, refetch } = useAtomValue(agentFlagsAtom); + + const enabledCount = useMemo(() => { + if (!flags) return 0; + return Object.entries(flags).filter(([k, v]) => k !== "disable_new_agent_stack" && v === true) + .length; + }, [flags]); + + if (isLoading) { + return ( +
+ + + +
+ ); + } + + if (isError || !flags) { + return ( + + Failed to load agent status + + {error instanceof Error ? error.message : "Unknown error."} + + + + ); + } + + const masterOff = flags.disable_new_agent_stack; + + return ( +
+ {masterOff ? ( + + + Master kill-switch is on + + + SURFSENSE_DISABLE_NEW_AGENT_STACK=true + + forces every new middleware off, regardless of the individual flags below. Restart the + backend after changing it. + + + ) : ( + + + + Agent stack + + {enabledCount} on + + + + Read-only mirror of the backend's AgentFeatureFlags. Flip an env var and + restart the backend to change a value. + + + )} + + {FLAG_GROUPS.map((group, groupIdx) => { + const allOff = group.flags.every((f) => !flags[f.key]); + return ( +
+ {groupIdx > 0 && } +
+
+
+

{group.title}

+

{group.subtitle}

+
+ {allOff && ( + + all off + + )} +
+
+ {group.flags.map((def) => ( + + ))} +
+
+
+ ); + })} +
+ ); +} diff --git a/surfsense_web/atoms/agent/action-log-sheet.atom.ts b/surfsense_web/atoms/agent/action-log-sheet.atom.ts new file mode 100644 index 000000000..f88d3ed1e --- /dev/null +++ b/surfsense_web/atoms/agent/action-log-sheet.atom.ts @@ -0,0 +1,19 @@ +import { atom } from "jotai"; + +interface ActionLogSheetState { + open: boolean; + threadId: number | null; +} + +export const actionLogSheetAtom = atom({ + open: false, + threadId: null, +}); + +export const openActionLogSheetAtom = atom(null, (_get, set, threadId: number) => { + set(actionLogSheetAtom, { open: true, threadId }); +}); + +export const closeActionLogSheetAtom = atom(null, (_get, set) => { + set(actionLogSheetAtom, { open: false, threadId: null }); +}); diff --git a/surfsense_web/atoms/agent/agent-flags-query.atom.ts b/surfsense_web/atoms/agent/agent-flags-query.atom.ts new file mode 100644 index 000000000..30158deaa --- /dev/null +++ b/surfsense_web/atoms/agent/agent-flags-query.atom.ts @@ -0,0 +1,17 @@ +import { atomWithQuery } from "jotai-tanstack-query"; +import { agentFlagsApiService } from "@/lib/apis/agent-flags-api.service"; +import { getBearerToken } from "@/lib/auth-utils"; + +export const AGENT_FLAGS_QUERY_KEY = ["agent", "flags"] as const; + +/** + * Reads the backend agent feature flags. Cached for the lifetime of the + * page (flags only change on backend restart) so we can drive UI gating + * without re-hitting the API. + */ +export const agentFlagsAtom = atomWithQuery(() => ({ + queryKey: AGENT_FLAGS_QUERY_KEY, + staleTime: 10 * 60 * 1000, + enabled: !!getBearerToken(), + queryFn: () => agentFlagsApiService.get(), +})); diff --git a/surfsense_web/components/agent-action-log/action-log-button.tsx b/surfsense_web/components/agent-action-log/action-log-button.tsx new file mode 100644 index 000000000..1c0383136 --- /dev/null +++ b/surfsense_web/components/agent-action-log/action-log-button.tsx @@ -0,0 +1,50 @@ +"use client"; + +import { useAtomValue, useSetAtom } from "jotai"; +import { Activity } from "lucide-react"; +import { useCallback } from "react"; +import { openActionLogSheetAtom } from "@/atoms/agent/action-log-sheet.atom"; +import { agentFlagsAtom } from "@/atoms/agent/agent-flags-query.atom"; +import { Button } from "@/components/ui/button"; +import { Tooltip, TooltipContent, TooltipTrigger } from "@/components/ui/tooltip"; + +interface ActionLogButtonProps { + threadId: number | null; +} + +/** + * Header button that opens the agent action log sheet for the current + * thread. Renders nothing when: + * - the action log feature flag is off (graceful no-op for older + * deployments), OR + * - there is no active thread (lazy-created chats haven't started). + */ +export function ActionLogButton({ threadId }: ActionLogButtonProps) { + const { data: flags } = useAtomValue(agentFlagsAtom); + const open = useSetAtom(openActionLogSheetAtom); + + const enabled = !!flags?.enable_action_log && !flags?.disable_new_agent_stack; + + const handleClick = useCallback(() => { + if (threadId !== null) open(threadId); + }, [open, threadId]); + + if (!enabled || threadId === null) return null; + + return ( + + + + + Agent actions + + ); +} diff --git a/surfsense_web/components/agent-action-log/action-log-item.tsx b/surfsense_web/components/agent-action-log/action-log-item.tsx new file mode 100644 index 000000000..425714c1f --- /dev/null +++ b/surfsense_web/components/agent-action-log/action-log-item.tsx @@ -0,0 +1,215 @@ +"use client"; + +import { ChevronRight, RotateCcw, ShieldOff, Undo2 } from "lucide-react"; +import { useState } from "react"; +import { toast } from "sonner"; +import { + AlertDialog, + AlertDialogAction, + AlertDialogCancel, + AlertDialogContent, + AlertDialogDescription, + AlertDialogFooter, + AlertDialogHeader, + AlertDialogTitle, + AlertDialogTrigger, +} from "@/components/ui/alert-dialog"; +import { Badge } from "@/components/ui/badge"; +import { Button } from "@/components/ui/button"; +import { Separator } from "@/components/ui/separator"; +import { getToolIcon } from "@/contracts/enums/toolIcons"; +import { type AgentAction, agentActionsApiService } from "@/lib/apis/agent-actions-api.service"; +import { AppError } from "@/lib/error"; +import { formatRelativeDate } from "@/lib/format-date"; +import { cn } from "@/lib/utils"; + +function formatToolName(name: string): string { + return name.replace(/_/g, " ").replace(/\b\w/g, (c) => c.toUpperCase()); +} + +interface ActionLogItemProps { + action: AgentAction; + threadId: number; + onRevertSuccess: () => void; +} + +export function ActionLogItem({ action, threadId, onRevertSuccess }: ActionLogItemProps) { + const [isExpanded, setIsExpanded] = useState(false); + const [isReverting, setIsReverting] = useState(false); + const [confirmOpen, setConfirmOpen] = useState(false); + + const isAlreadyReverted = action.reverted_by_action_id !== null; + const isRevertAction = action.is_revert_action; + const hasError = action.error !== null && action.error !== undefined; + + const Icon = getToolIcon(action.tool_name); + const displayName = formatToolName(action.tool_name); + + const argsPreview = action.args ? JSON.stringify(action.args, null, 2) : null; + const truncatedArgs = + argsPreview && argsPreview.length > 600 ? `${argsPreview.slice(0, 600)}…` : argsPreview; + + const canRevert = action.reversible && !isAlreadyReverted && !isRevertAction && !hasError; + + const handleRevert = async () => { + setIsReverting(true); + try { + const response = await agentActionsApiService.revert(threadId, action.id); + toast.success(response.message || "Action reverted successfully."); + onRevertSuccess(); + } catch (err) { + const message = + err instanceof AppError + ? err.message + : err instanceof Error + ? err.message + : "Failed to revert action."; + toast.error(message); + } finally { + setIsReverting(false); + setConfirmOpen(false); + } + }; + + return ( +
+ + + {isExpanded && ( +
+ {truncatedArgs && ( +
+

+ Arguments +

+ 
+								{truncatedArgs}
+
+
+ )} + {action.error && ( +
+

+ Error +

+ 
+								{JSON.stringify(action.error, null, 2)}
+
+
+ )} + {action.reverse_descriptor && ( +
+

+ Reverse plan +

+ 
+								{JSON.stringify(action.reverse_descriptor, null, 2)}
+
+
+ )} + + + +
+

+ Action ID: {action.id} +

+ {canRevert ? ( + + + + + + + Revert this action? + + This will undo {displayName} and append a + new audit entry. The agent's chat history is preserved — only the tool's + effects on your knowledge base or connectors will be reversed where possible. + + + + Cancel + { + e.preventDefault(); + handleRevert(); + }} + disabled={isReverting} + > + {isReverting ? "Reverting…" : "Revert"} + + + + + ) : ( +
+ + {isAlreadyReverted + ? "Already reverted" + : isRevertAction + ? "Revert entry" + : hasError + ? "Cannot revert errored action" + : "Not reversible"} +
+ )} +
+
+ )} +
+ ); +} diff --git a/surfsense_web/components/agent-action-log/action-log-sheet.tsx b/surfsense_web/components/agent-action-log/action-log-sheet.tsx new file mode 100644 index 000000000..68d2ffef3 --- /dev/null +++ b/surfsense_web/components/agent-action-log/action-log-sheet.tsx @@ -0,0 +1,185 @@ +"use client"; + +import { useQuery, useQueryClient } from "@tanstack/react-query"; +import { useAtom, useAtomValue } from "jotai"; +import { Activity, RefreshCcw } from "lucide-react"; +import { useCallback, useMemo } from "react"; +import { actionLogSheetAtom } from "@/atoms/agent/action-log-sheet.atom"; +import { agentFlagsAtom } from "@/atoms/agent/agent-flags-query.atom"; +import { Badge } from "@/components/ui/badge"; +import { Button } from "@/components/ui/button"; +import { Separator } from "@/components/ui/separator"; +import { + Sheet, + SheetContent, + SheetDescription, + SheetHeader, + SheetTitle, +} from "@/components/ui/sheet"; +import { Skeleton } from "@/components/ui/skeleton"; +import { agentActionsApiService } from "@/lib/apis/agent-actions-api.service"; +import { ActionLogItem } from "./action-log-item"; + +const ACTION_LOG_PAGE_SIZE = 50; + +function actionLogQueryKey(threadId: number) { + return ["agent-actions", threadId] as const; +} + +function EmptyState() { + return ( +
+
+ +
+
+

No actions logged yet

+

+ Once the agent calls a tool in this thread, it will show up here. From the log you can + inspect arguments and revert reversible actions. +

+
+
+ ); +} + +function DisabledState() { + return ( +
+
+ +
+
+

Action log is disabled

+

+ This deployment hasn't enabled the agent action log. An admin can flip + + SURFSENSE_ENABLE_ACTION_LOG + + . +

+
+
+ ); +} + +const SKELETON_KEYS = ["s1", "s2", "s3", "s4"] as const; + +function LoadingState() { + return ( +
+ {SKELETON_KEYS.map((key) => ( + + ))} +
+ ); +} + +export function ActionLogSheet() { + const [state, setState] = useAtom(actionLogSheetAtom); + const queryClient = useQueryClient(); + + const { data: flags } = useAtomValue(agentFlagsAtom); + const actionLogEnabled = !!flags?.enable_action_log && !flags?.disable_new_agent_stack; + const revertEnabled = !!flags?.enable_revert_route && !flags?.disable_new_agent_stack; + + const threadId = state.threadId; + + const { data, isLoading, isFetching, isError, error, refetch } = useQuery({ + queryKey: threadId !== null ? actionLogQueryKey(threadId) : ["agent-actions", "none"], + queryFn: () => + agentActionsApiService.listForThread(threadId as number, { + page: 0, + pageSize: ACTION_LOG_PAGE_SIZE, + }), + enabled: state.open && threadId !== null && actionLogEnabled, + staleTime: 15 * 1000, + }); + + const handleRevertSuccess = useCallback(() => { + if (threadId !== null) { + queryClient.invalidateQueries({ queryKey: actionLogQueryKey(threadId) }); + } + }, [queryClient, threadId]); + + const items = useMemo(() => data?.items ?? [], [data]); + + return ( + setState((s) => ({ ...s, open }))}> + + +
+
+ + Agent actions + {data?.total !== undefined && data.total > 0 && ( + + {data.total} + + )} +
+ +
+ + Audit trail of every tool call the agent made in this thread. + {revertEnabled + ? " Reversible actions can be undone in place." + : " Reverts are read-only on this deployment."} + + + + + +
+ {!actionLogEnabled ? ( + + ) : threadId === null ? ( + + ) : isLoading ? ( + + ) : isError ? ( +
+

Failed to load actions

+

+ {error instanceof Error ? error.message : "Unknown error"} +

+ +
+ ) : items.length === 0 ? ( + + ) : ( +
+ {items.map((action) => ( + + ))} + {data?.has_more && ( +

+ Showing {items.length} of {data.total}. Older actions are paginated. +

+ )} +
+ )} +
+ + + ); +} diff --git a/surfsense_web/components/assistant-ui/markdown-text.tsx b/surfsense_web/components/assistant-ui/markdown-text.tsx index 8bb228580..7655e10cc 100644 --- a/surfsense_web/components/assistant-ui/markdown-text.tsx +++ b/surfsense_web/components/assistant-ui/markdown-text.tsx @@ -85,10 +85,13 @@ function preprocessMarkdown(content: string): string { } ); + // All math forms are normalised to $$...$$ so we can disable single-dollar + // inline math in remark-math (otherwise currency like "$3,120.00 and $0.00" + // gets parsed as a LaTeX expression). // 1. Block math: \[...\] → $$...$$ content = content.replace(/\\\[([\s\S]*?)\\\]/g, (_, inner) => `$$${inner}$$`); - // 2. Inline math: \(...\) → $...$ - content = content.replace(/\\\(([\s\S]*?)\\\)/g, (_, inner) => `$${inner}$`); + // 2. Inline math: \(...\) → $$...$$ + content = content.replace(/\\\(([\s\S]*?)\\\)/g, (_, inner) => `$$${inner}$$`); // 3. Block: \begin{equation}...\end{equation} → $$...$$ content = content.replace( /\\begin\{equation\}([\s\S]*?)\\end\{equation\}/g, @@ -99,8 +102,11 @@ function preprocessMarkdown(content: string): string { /\\begin\{displaymath\}([\s\S]*?)\\end\{displaymath\}/g, (_, inner) => `$$${inner}$$` ); - // 5. Inline: \begin{math}...\end{math} → $...$ - content = content.replace(/\\begin\{math\}([\s\S]*?)\\end\{math\}/g, (_, inner) => `$${inner}$`); + // 5. Inline: \begin{math}...\end{math} → $$...$$ + content = content.replace( + /\\begin\{math\}([\s\S]*?)\\end\{math\}/g, + (_, inner) => `$$${inner}$$` + ); // 6. Strip backtick wrapping around math: `$$...$$` → $$...$$ and `$...$` → $...$ content = content.replace(/`(\${1,2})((?:(?!\1).)+)\1`/g, "$1$2$1"); @@ -180,7 +186,7 @@ const MarkdownTextImpl = () => { return ( { if (isInterruptResult(props.result)) { + if (isDoomLoopInterrupt(props.result)) { + return ; + } return ; } return ; diff --git a/surfsense_web/components/layout/providers/LayoutDataProvider.tsx b/surfsense_web/components/layout/providers/LayoutDataProvider.tsx index aecf55a27..3efdab03b 100644 --- a/surfsense_web/components/layout/providers/LayoutDataProvider.tsx +++ b/surfsense_web/components/layout/providers/LayoutDataProvider.tsx @@ -28,6 +28,7 @@ import { import { currentUserAtom } from "@/atoms/user/user-query.atoms"; import { SearchSpaceSettingsDialog } from "@/components/settings/search-space-settings-dialog"; import { TeamDialog } from "@/components/settings/team-dialog"; +import { ActionLogSheet } from "@/components/agent-action-log/action-log-sheet"; import { UserSettingsDialog } from "@/components/settings/user-settings-dialog"; import { AlertDialog, @@ -909,6 +910,9 @@ export function LayoutDataProvider({ searchSpaceId, children }: LayoutDataProvid + + {/* Agent action log + revert sheet */} + ); } diff --git a/surfsense_web/components/layout/ui/header/Header.tsx b/surfsense_web/components/layout/ui/header/Header.tsx index ec54cb901..f49d7fb88 100644 --- a/surfsense_web/components/layout/ui/header/Header.tsx +++ b/surfsense_web/components/layout/ui/header/Header.tsx @@ -5,6 +5,7 @@ import { usePathname } from "next/navigation"; import { currentThreadAtom } from "@/atoms/chat/current-thread.atom"; import { activeSearchSpaceIdAtom } from "@/atoms/search-spaces/search-space-query.atoms"; import { activeTabAtom, tabsAtom } from "@/atoms/tabs/tabs.atom"; +import { ActionLogButton } from "@/components/agent-action-log/action-log-button"; import { ChatHeader } from "@/components/new-chat/chat-header"; import { ChatShareButton } from "@/components/new-chat/chat-share-button"; import { useIsMobile } from "@/hooks/use-mobile"; @@ -69,6 +70,7 @@ export function Header({ mobileMenuTrigger }: HeaderProps) { {/* Right side - Actions */}
+ {hasThread && } {hasThread && ( )} diff --git a/surfsense_web/components/markdown-viewer.tsx b/surfsense_web/components/markdown-viewer.tsx index 5775fe083..c4d73e30b 100644 --- a/surfsense_web/components/markdown-viewer.tsx +++ b/surfsense_web/components/markdown-viewer.tsx @@ -10,7 +10,11 @@ const code = createCodePlugin({ }); const math = createMathPlugin({ - singleDollarTextMath: true, + // Disabled so currency like "$3,120.00 and ... $0.00" isn't parsed as + // inline LaTeX. convertLatexDelimiters() below normalises any genuine + // inline math (\(...\), $...$ starting with a LaTeX command, etc.) to + // $$...$$, so this flip doesn't lose any math rendering. + singleDollarTextMath: false, }); interface MarkdownViewerProps { diff --git a/surfsense_web/components/settings/user-settings-dialog.tsx b/surfsense_web/components/settings/user-settings-dialog.tsx index 6740aad92..a04ce16dd 100644 --- a/surfsense_web/components/settings/user-settings-dialog.tsx +++ b/surfsense_web/components/settings/user-settings-dialog.tsx @@ -2,6 +2,7 @@ import { useAtom } from "jotai"; import { + Activity, Brain, CircleUser, Globe, @@ -9,6 +10,7 @@ import { KeyRound, Monitor, ReceiptText, + ShieldCheck, Sparkles, } from "lucide-react"; import dynamic from "next/dynamic"; @@ -74,6 +76,20 @@ const MemoryContent = dynamic( ), { ssr: false } ); +const AgentPermissionsContent = dynamic( + () => + import( + "@/app/dashboard/[search_space_id]/user-settings/components/AgentPermissionsContent" + ).then((m) => ({ default: m.AgentPermissionsContent })), + { ssr: false } +); +const AgentStatusContent = dynamic( + () => + import("@/app/dashboard/[search_space_id]/user-settings/components/AgentStatusContent").then( + (m) => ({ default: m.AgentStatusContent }) + ), + { ssr: false } +); export function UserSettingsDialog() { const t = useTranslations("userSettings"); @@ -103,6 +119,16 @@ export function UserSettingsDialog() { label: "Memory", icon: , }, + { + value: "agent-permissions", + label: "Agent Permissions", + icon: , + }, + { + value: "agent-status", + label: "Agent Status", + icon: , + }, { value: "purchases", label: "Purchase History", @@ -141,6 +167,8 @@ export function UserSettingsDialog() { {state.initialTab === "prompts" && } {state.initialTab === "community-prompts" && } {state.initialTab === "memory" && } + {state.initialTab === "agent-permissions" && } + {state.initialTab === "agent-status" && } {state.initialTab === "purchases" && } {state.initialTab === "desktop" && } {state.initialTab === "desktop-shortcuts" && } diff --git a/surfsense_web/components/tool-ui/doom-loop-approval.tsx b/surfsense_web/components/tool-ui/doom-loop-approval.tsx new file mode 100644 index 000000000..6132a71ed --- /dev/null +++ b/surfsense_web/components/tool-ui/doom-loop-approval.tsx @@ -0,0 +1,187 @@ +"use client"; + +import type { ToolCallMessagePartComponent } from "@assistant-ui/react"; +import { CornerDownLeftIcon, OctagonAlert } from "lucide-react"; +import { useCallback, useEffect, useMemo } from "react"; +import { TextShimmerLoader } from "@/components/prompt-kit/loader"; +import { Alert, AlertDescription, AlertTitle } from "@/components/ui/alert"; +import { Badge } from "@/components/ui/badge"; +import { Button } from "@/components/ui/button"; +import { Separator } from "@/components/ui/separator"; +import { useHitlPhase } from "@/hooks/use-hitl-phase"; +import type { HitlDecision, InterruptResult } from "@/lib/hitl"; +import { isInterruptResult, useHitlDecision } from "@/lib/hitl"; + +/** + * Specialized HITL card for ``DoomLoopMiddleware`` interrupts. The + * backend signals these by setting ``context.permission === "doom_loop"`` + * on the ``permission_ask`` interrupt. + * + * The card replaces the generic "approve/reject" framing with a + * "continue/stop" affordance that better matches the user's mental + * model: the agent is stuck repeating itself, not asking permission + * for a destructive action. + */ +function DoomLoopCard({ + toolName, + args, + interruptData, + onDecision, +}: { + toolName: string; + args: Record; + interruptData: InterruptResult; + onDecision: (decision: HitlDecision) => void; +}) { + const { phase, setProcessing, setRejected } = useHitlPhase(interruptData); + + const context = (interruptData.context ?? {}) as Record; + const threshold = typeof context.threshold === "number" ? context.threshold : 3; + const stuckTool = (typeof context.tool === "string" && context.tool) || toolName; + const recentSignatures = Array.isArray(context.recent_signatures) + ? (context.recent_signatures as string[]) + : []; + const displayName = stuckTool.replace(/_/g, " ").replace(/\b\w/g, (c) => c.toUpperCase()); + + const argPreview = useMemo(() => { + if (!args || Object.keys(args).length === 0) return null; + try { + const json = JSON.stringify(args, null, 2); + return json.length > 600 ? `${json.slice(0, 600)}…` : json; + } catch { + return null; + } + }, [args]); + + const handleContinue = useCallback(() => { + if (phase !== "pending") return; + setProcessing(); + onDecision({ type: "approve" }); + }, [phase, setProcessing, onDecision]); + + const handleStop = useCallback(() => { + if (phase !== "pending") return; + setRejected(); + onDecision({ type: "reject", message: "Doom loop: user requested stop." }); + }, [phase, setRejected, onDecision]); + + useEffect(() => { + const handler = (e: KeyboardEvent) => { + if (phase !== "pending") return; + if (e.key === "Enter" && !e.shiftKey && !e.ctrlKey && !e.metaKey) { + e.preventDefault(); + handleStop(); + } + }; + window.addEventListener("keydown", handler); + return () => window.removeEventListener("keydown", handler); + }, [phase, handleStop]); + + const isResolved = phase === "complete" || phase === "rejected"; + + return ( + + + + + {phase === "rejected" + ? "Stopped" + : phase === "processing" + ? "Continuing…" + : phase === "complete" + ? "Continued" + : "I might be stuck"} + + {!isResolved && ( + + doom-loop + + )} + + + {phase === "processing" ? ( + + ) : phase === "rejected" ? ( +

+ I stopped retrying {displayName} as you asked. +

+ ) : phase === "complete" ? ( +

+ Continuing to call {displayName} as you asked. +

+ ) : ( +

+ I called {displayName} {threshold} times in a row + with similar arguments. Should I keep going or stop and rethink? +

+ )} + + {argPreview && phase === "pending" && ( + <> + +
+

+ Last arguments +

+ 
+								{argPreview}
+
+
+ + )} + + {recentSignatures.length > 0 && phase === "pending" && ( +
+ + Show repeated signatures ({recentSignatures.length}) + +
    + {recentSignatures.map((sig) => ( +
  • + {sig} +
    • + ))} +
+
+ )} + + {phase === "pending" && ( +
+ + +
+ )} + + + ); +} + +export const DoomLoopApprovalToolUI: ToolCallMessagePartComponent = ({ + toolName, + args, + result, +}) => { + const { dispatch } = useHitlDecision(); + + if (!result || !isInterruptResult(result)) return null; + + return ( + } + interruptData={result} + onDecision={(decision) => dispatch([decision])} + /> + ); +}; + +export function isDoomLoopInterrupt(result: unknown): boolean { + if (!isInterruptResult(result)) return false; + const ctx = (result.context ?? {}) as Record; + return ctx.permission === "doom_loop"; +} diff --git a/surfsense_web/lib/apis/agent-actions-api.service.ts b/surfsense_web/lib/apis/agent-actions-api.service.ts new file mode 100644 index 000000000..007bb131e --- /dev/null +++ b/surfsense_web/lib/apis/agent-actions-api.service.ts @@ -0,0 +1,64 @@ +import { z } from "zod"; +import { baseApiService } from "./base-api.service"; + +const AgentActionReadSchema = z.object({ + id: z.number(), + thread_id: z.number(), + user_id: z.string().nullable(), + search_space_id: z.number(), + tool_name: z.string(), + args: z.record(z.string(), z.unknown()).nullable(), + result_id: z.string().nullable(), + reversible: z.boolean(), + reverse_descriptor: z.record(z.string(), z.unknown()).nullable(), + error: z.record(z.string(), z.unknown()).nullable(), + reverse_of: z.number().nullable(), + reverted_by_action_id: z.number().nullable(), + is_revert_action: z.boolean(), + created_at: z.string(), +}); + +export type AgentAction = z.infer; + +const AgentActionListResponseSchema = z.object({ + items: z.array(AgentActionReadSchema), + total: z.number(), + page: z.number(), + page_size: z.number(), + has_more: z.boolean(), +}); + +export type AgentActionListResponse = z.infer; + +const RevertResponseSchema = z.object({ + status: z.literal("ok"), + message: z.string(), + new_action_id: z.number().nullable().optional(), +}); + +export type RevertResponse = z.infer; + +class AgentActionsApiService { + listForThread = async ( + threadId: number, + opts: { page?: number; pageSize?: number } = {} + ): Promise => { + const params = new URLSearchParams(); + params.set("page", String(opts.page ?? 0)); + params.set("page_size", String(opts.pageSize ?? 50)); + return baseApiService.get( + `/api/v1/threads/${threadId}/actions?${params.toString()}`, + AgentActionListResponseSchema + ); + }; + + revert = async (threadId: number, actionId: number): Promise => { + return baseApiService.post( + `/api/v1/threads/${threadId}/revert/${actionId}`, + RevertResponseSchema, + { body: {} } + ); + }; +} + +export const agentActionsApiService = new AgentActionsApiService(); diff --git a/surfsense_web/lib/apis/agent-flags-api.service.ts b/surfsense_web/lib/apis/agent-flags-api.service.ts new file mode 100644 index 000000000..87332ca9f --- /dev/null +++ b/surfsense_web/lib/apis/agent-flags-api.service.ts @@ -0,0 +1,40 @@ +import { z } from "zod"; +import { baseApiService } from "./base-api.service"; + +const AgentFeatureFlagsSchema = z.object({ + disable_new_agent_stack: z.boolean(), + + enable_context_editing: z.boolean(), + enable_compaction_v2: z.boolean(), + enable_retry_after: z.boolean(), + enable_model_fallback: z.boolean(), + enable_model_call_limit: z.boolean(), + enable_tool_call_limit: z.boolean(), + enable_tool_call_repair: z.boolean(), + enable_doom_loop: z.boolean(), + + enable_permission: z.boolean(), + enable_busy_mutex: z.boolean(), + enable_llm_tool_selector: z.boolean(), + + enable_skills: z.boolean(), + enable_specialized_subagents: z.boolean(), + enable_kb_planner_runnable: z.boolean(), + + enable_action_log: z.boolean(), + enable_revert_route: z.boolean(), + + enable_plugin_loader: z.boolean(), + + enable_otel: z.boolean(), +}); + +export type AgentFeatureFlags = z.infer; + +class AgentFlagsApiService { + get = async (): Promise => { + return baseApiService.get(`/api/v1/agent/flags`, AgentFeatureFlagsSchema); + }; +} + +export const agentFlagsApiService = new AgentFlagsApiService(); diff --git a/surfsense_web/lib/apis/agent-permissions-api.service.ts b/surfsense_web/lib/apis/agent-permissions-api.service.ts new file mode 100644 index 000000000..6927c55d0 --- /dev/null +++ b/surfsense_web/lib/apis/agent-permissions-api.service.ts @@ -0,0 +1,90 @@ +import { z } from "zod"; +import { ValidationError } from "@/lib/error"; +import { baseApiService } from "./base-api.service"; + +const ActionEnum = z.enum(["allow", "deny", "ask"]); +export type AgentPermissionAction = z.infer; + +const AgentPermissionRuleSchema = z.object({ + id: z.number(), + search_space_id: z.number(), + user_id: z.string().nullable(), + thread_id: z.number().nullable(), + permission: z.string(), + pattern: z.string(), + action: ActionEnum, + created_at: z.string(), +}); + +export type AgentPermissionRule = z.infer; + +const AgentPermissionRuleListSchema = z.array(AgentPermissionRuleSchema); + +const AgentPermissionRuleCreateSchema = z.object({ + permission: z + .string() + .min(1, "Permission is required") + .max(255) + .regex(/^[a-zA-Z0-9_:.\-*]+$/, "Use letters, digits, '.', '_', ':', '-', or '*' wildcards."), + pattern: z.string().min(1).max(255).default("*"), + action: ActionEnum, + user_id: z.string().nullable().optional(), + thread_id: z.number().nullable().optional(), +}); + +export type AgentPermissionRuleCreate = z.infer; + +const AgentPermissionRuleUpdateSchema = z.object({ + pattern: z.string().min(1).max(255).optional(), + action: ActionEnum.optional(), +}); + +export type AgentPermissionRuleUpdate = z.infer; + +class AgentPermissionsApiService { + list = async (searchSpaceId: number): Promise => { + return baseApiService.get( + `/api/v1/searchspaces/${searchSpaceId}/agent/permissions/rules`, + AgentPermissionRuleListSchema + ); + }; + + create = async ( + searchSpaceId: number, + payload: AgentPermissionRuleCreate + ): Promise => { + const parsed = AgentPermissionRuleCreateSchema.safeParse(payload); + if (!parsed.success) { + throw new ValidationError(parsed.error.issues.map((i) => i.message).join(", ")); + } + return baseApiService.post( + `/api/v1/searchspaces/${searchSpaceId}/agent/permissions/rules`, + AgentPermissionRuleSchema, + { body: parsed.data } + ); + }; + + update = async ( + searchSpaceId: number, + ruleId: number, + payload: AgentPermissionRuleUpdate + ): Promise => { + const parsed = AgentPermissionRuleUpdateSchema.safeParse(payload); + if (!parsed.success) { + throw new ValidationError(parsed.error.issues.map((i) => i.message).join(", ")); + } + return baseApiService.patch( + `/api/v1/searchspaces/${searchSpaceId}/agent/permissions/rules/${ruleId}`, + AgentPermissionRuleSchema, + { body: parsed.data } + ); + }; + + remove = async (searchSpaceId: number, ruleId: number): Promise => { + await baseApiService.delete( + `/api/v1/searchspaces/${searchSpaceId}/agent/permissions/rules/${ruleId}` + ); + }; +} + +export const agentPermissionsApiService = new AgentPermissionsApiService(); From 76c91adebc0b30102e0d6df026f62b47716d3ac2 Mon Sep 17 00:00:00 2001 From: Anish Sarkar <104695310+AnishSarkar22@users.noreply.github.com> Date: Wed, 29 Apr 2026 04:12:42 +0530 Subject: [PATCH 09/17] refactor(mentions): replace sidebarSelectedDocumentsAtom with mentionedDocumentsAtom and introduce getMentionDocKey utility for consistent document key generation --- .../atoms/chat/mentioned-documents.atom.ts | 23 ------ .../assistant-ui/inline-mention-editor.tsx | 82 ++++++++++++------- .../components/assistant-ui/thread.tsx | 53 ++++++------ .../layout/ui/sidebar/DocumentsSidebar.tsx | 29 ++++--- surfsense_web/lib/chat/mention-doc-key.ts | 8 ++ 5 files changed, 102 insertions(+), 93 deletions(-) create mode 100644 surfsense_web/lib/chat/mention-doc-key.ts diff --git a/surfsense_web/atoms/chat/mentioned-documents.atom.ts b/surfsense_web/atoms/chat/mentioned-documents.atom.ts index 47401995d..9c4546237 100644 --- a/surfsense_web/atoms/chat/mentioned-documents.atom.ts +++ b/surfsense_web/atoms/chat/mentioned-documents.atom.ts @@ -9,29 +9,6 @@ import type { Document } from "@/contracts/types/document.types"; */ export const mentionedDocumentsAtom = atom[]>([]); -/** - * Back-compat alias for sidebar checkbox selection. - * This now points to mentionedDocumentsAtom so the app has a single source - * of truth for mentioned/selected documents. - */ -export const sidebarSelectedDocumentsAtom = atom< - Pick[], - [ - | Pick[] - | (( - prev: Pick[] - ) => Pick[]), - ], - void ->( - (get) => get(mentionedDocumentsAtom), - (get, set, update) => { - const prev = get(mentionedDocumentsAtom); - const next = typeof update === "function" ? update(prev) : update; - set(mentionedDocumentsAtom, next); - } -); - /** * Derived read-only atom that maps deduplicated mentioned docs * into backend payload fields. diff --git a/surfsense_web/components/assistant-ui/inline-mention-editor.tsx b/surfsense_web/components/assistant-ui/inline-mention-editor.tsx index e75a840c0..05277f508 100644 --- a/surfsense_web/components/assistant-ui/inline-mention-editor.tsx +++ b/surfsense_web/components/assistant-ui/inline-mention-editor.tsx @@ -14,6 +14,7 @@ import { import { renderToStaticMarkup } from "react-dom/server"; import { getConnectorIcon } from "@/contracts/enums/connectorIcons"; import type { Document } from "@/contracts/types/document.types"; +import { getMentionDocKey } from "@/lib/chat/mention-doc-key"; import { cn } from "@/lib/utils"; function renderElementToHTML(element: ReactElement): string { @@ -57,7 +58,6 @@ interface InlineMentionEditorProps { onKeyDown?: (e: React.KeyboardEvent) => void; disabled?: boolean; className?: string; - initialDocuments?: MentionedDocument[]; initialText?: string; } @@ -109,7 +109,6 @@ export const InlineMentionEditor = forwardRef(null); const [isEmpty, setIsEmpty] = useState(true); const [mentionedDocs, setMentionedDocs] = useState>( - () => new Map(initialDocuments.map((d) => [`${d.document_type ?? "UNKNOWN"}:${d.id}`, d])) + () => new Map() ); const isComposingRef = useRef(false); const lastSelectionRangeRef = useRef(null); + const isRangeInsideEditor = useCallback((range: Range | null): range is Range => { + if (!range || !editorRef.current) return false; + return ( + editorRef.current.contains(range.startContainer) && + editorRef.current.contains(range.endContainer) + ); + }, []); const isSelectionInsideEditor = useCallback( (selection: Selection | null): selection is Selection => { if (!selection || selection.rangeCount === 0 || !editorRef.current) return false; const range = selection.getRangeAt(0); - return editorRef.current.contains(range.startContainer); + return isRangeInsideEditor(range); }, - [] + [isRangeInsideEditor] ); const rememberSelection = useCallback(() => { @@ -139,11 +145,11 @@ export const InlineMentionEditor = forwardRef { const selection = window.getSelection(); if (!selection) return null; - if (!lastSelectionRangeRef.current) return selection; + if (!isRangeInsideEditor(lastSelectionRangeRef.current)) return null; selection.removeAllRanges(); selection.addRange(lastSelectionRangeRef.current.cloneRange()); return selection; - }, []); + }, [isRangeInsideEditor]); useEffect(() => { const handleSelectionChange = () => { @@ -154,23 +160,13 @@ export const InlineMentionEditor = forwardRef document.removeEventListener("selectionchange", handleSelectionChange); }, [rememberSelection]); - - // Sync initial documents - useEffect(() => { - if (initialDocuments.length > 0) { - setMentionedDocs( - new Map(initialDocuments.map((d) => [`${d.document_type ?? "UNKNOWN"}:${d.id}`, d])) - ); - } - }, [initialDocuments]); - useEffect(() => { if (!initialText || !editorRef.current) return; editorRef.current.innerText = initialText; editorRef.current.appendChild(document.createElement("br")); editorRef.current.appendChild(document.createElement("br")); setIsEmpty(false); - onChange?.(initialText, initialDocuments); + onChange?.(initialText, []); editorRef.current.focus(); const sel = window.getSelection(); const range = document.createRange(); @@ -182,7 +178,7 @@ export const InlineMentionEditor = forwardRef { @@ -284,7 +280,7 @@ export const InlineMentionEditor = forwardRef { const next = new Map(prev); next.delete(docKey); @@ -358,7 +354,7 @@ export const InlineMentionEditor = forwardRef new Map(prev).set(docKey, mentionDoc)); const nextDocs = new Map(mentionedDocs); nextDocs.set(docKey, mentionDoc); @@ -367,12 +363,33 @@ export const InlineMentionEditor = forwardRef { if (!editorRef.current) return; - const chipKey = `${docType ?? "UNKNOWN"}:${docId}`; + const chipKey = getMentionDocKey({ id: docId, document_type: docType }); const chips = editorRef.current.querySelectorAll( `span[${CHIP_DATA_ATTR}="true"]` ); @@ -696,7 +712,10 @@ export const InlineMentionEditor = forwardRef { const next = new Map(prev); next.delete(chipKey); @@ -734,7 +753,10 @@ export const InlineMentionEditor = forwardRef { const next = new Map(prev); next.delete(chipKey); diff --git a/surfsense_web/components/assistant-ui/thread.tsx b/surfsense_web/components/assistant-ui/thread.tsx index dcc068bd1..f9e5ca7fb 100644 --- a/surfsense_web/components/assistant-ui/thread.tsx +++ b/surfsense_web/components/assistant-ui/thread.tsx @@ -87,6 +87,7 @@ import { useBatchCommentsPreload } from "@/hooks/use-comments"; import { useCommentsSync } from "@/hooks/use-comments-sync"; import { useMediaQuery } from "@/hooks/use-media-query"; import { useElectronAPI } from "@/hooks/use-platform"; +import { getMentionDocKey } from "@/lib/chat/mention-doc-key"; import { SLIDEOUT_PANEL_OPENED_EVENT } from "@/lib/layout-events"; import { cn } from "@/lib/utils"; @@ -338,6 +339,9 @@ const Composer: FC = () => { const [mentionQuery, setMentionQuery] = useState(""); const [actionQuery, setActionQuery] = useState(""); const editorRef = useRef(null); + const prevMentionedDocsRef = useRef< + Map> + >(new Map()); const documentPickerRef = useRef(null); const promptPickerRef = useRef(null); const viewportRef = useRef(null); @@ -633,51 +637,50 @@ const Composer: FC = () => { const handleDocumentsMention = useCallback( (documents: Pick[]) => { - const existingKeys = new Set(mentionedDocuments.map((d) => `${d.document_type}:${d.id}`)); - const newDocs = documents.filter( - (doc) => !existingKeys.has(`${doc.document_type}:${doc.id}`) - ); + const editorMentionedDocs = editorRef.current?.getMentionedDocuments() ?? []; + const editorDocKeys = new Set(editorMentionedDocs.map((doc) => getMentionDocKey(doc))); - for (const doc of newDocs) { + for (const doc of documents) { + const key = getMentionDocKey(doc); + if (editorDocKeys.has(key)) continue; editorRef.current?.insertDocumentChip(doc); } setMentionedDocuments((prev) => { - const existingKeySet = new Set(prev.map((d) => `${d.document_type}:${d.id}`)); - const uniqueNewDocs = documents.filter( - (doc) => !existingKeySet.has(`${doc.document_type}:${doc.id}`) - ); + const existingKeySet = new Set(prev.map((d) => getMentionDocKey(d))); + const uniqueNewDocs = documents.filter((doc) => !existingKeySet.has(getMentionDocKey(doc))); return [...prev, ...uniqueNewDocs]; }); setMentionQuery(""); }, - [mentionedDocuments, setMentionedDocuments] + [setMentionedDocuments] ); useEffect(() => { const editor = editorRef.current; - if (!editor) return; + const nextDocsMap = new Map(mentionedDocuments.map((doc) => [getMentionDocKey(doc), doc])); + const prevDocsMap = prevMentionedDocsRef.current; - const toKey = (doc: { id: number; document_type?: string }) => - `${doc.document_type ?? "UNKNOWN"}:${doc.id}`; - - const atomDocs = mentionedDocuments; - const editorDocs = editor.getMentionedDocuments(); - const atomKeys = new Set(atomDocs.map(toKey)); - const editorKeys = new Set(editorDocs.map(toKey)); - - for (const doc of atomDocs) { - if (!editorKeys.has(toKey(doc))) { - editor.insertDocumentChip(doc, { removeTriggerText: false }); - } + if (!editor) { + prevMentionedDocsRef.current = nextDocsMap; + return; } - for (const doc of editorDocs) { - if (!atomKeys.has(toKey(doc))) { + const editorKeys = new Set(editor.getMentionedDocuments().map(getMentionDocKey)); + + for (const [key, doc] of nextDocsMap) { + if (prevDocsMap.has(key) || editorKeys.has(key)) continue; + editor.insertDocumentChip(doc, { removeTriggerText: false }); + } + + for (const [key, doc] of prevDocsMap) { + if (!nextDocsMap.has(key)) { editor.removeDocumentChip(doc.id, doc.document_type); } } + + prevMentionedDocsRef.current = nextDocsMap; }, [mentionedDocuments]); return ( diff --git a/surfsense_web/components/layout/ui/sidebar/DocumentsSidebar.tsx b/surfsense_web/components/layout/ui/sidebar/DocumentsSidebar.tsx index 3c5a64b0e..63b6dc1b7 100644 --- a/surfsense_web/components/layout/ui/sidebar/DocumentsSidebar.tsx +++ b/surfsense_web/components/layout/ui/sidebar/DocumentsSidebar.tsx @@ -24,7 +24,7 @@ import type React from "react"; import { useCallback, useEffect, useMemo, useRef, useState } from "react"; import { toast } from "sonner"; import { - sidebarSelectedDocumentsAtom, + mentionedDocumentsAtom, } from "@/atoms/chat/mentioned-documents.atom"; import { connectorDialogOpenAtom } from "@/atoms/connector-dialog/connector-dialog.atoms"; import { connectorsAtom } from "@/atoms/connectors/connector-query.atoms"; @@ -74,6 +74,7 @@ import type { DocumentTypeEnum } from "@/contracts/types/document.types"; import { useDebouncedValue } from "@/hooks/use-debounced-value"; import { useMediaQuery } from "@/hooks/use-media-query"; import { usePlatform, useElectronAPI } from "@/hooks/use-platform"; +import { getMentionDocKey } from "@/lib/chat/mention-doc-key"; import { anonymousChatApiService } from "@/lib/apis/anonymous-chat-api.service"; import { documentsApiService } from "@/lib/apis/documents-api.service"; import { foldersApiService } from "@/lib/apis/folders-api.service"; @@ -414,7 +415,7 @@ function AuthenticatedDocumentsSidebarBase({ }, [refreshWatchedIds]); const { mutateAsync: deleteDocumentMutation } = useAtomValue(deleteDocumentMutationAtom); - const [sidebarDocs, setSidebarDocs] = useAtom(sidebarSelectedDocumentsAtom); + const [sidebarDocs, setSidebarDocs] = useAtom(mentionedDocumentsAtom); const mentionedDocIds = useMemo(() => new Set(sidebarDocs.map((d) => d.id)), [sidebarDocs]); // Folder state @@ -859,12 +860,12 @@ function AuthenticatedDocumentsSidebarBase({ const handleToggleChatMention = useCallback( (doc: { id: number; title: string; document_type: string }, isMentioned: boolean) => { - const key = `${doc.document_type}:${doc.id}`; + const key = getMentionDocKey(doc); if (isMentioned) { - setSidebarDocs((prev) => prev.filter((d) => `${d.document_type}:${d.id}` !== key)); + setSidebarDocs((prev) => prev.filter((d) => getMentionDocKey(d) !== key)); } else { setSidebarDocs((prev) => { - if (prev.some((d) => `${d.document_type}:${d.id}` === key)) return prev; + if (prev.some((d) => getMentionDocKey(d) === key)) return prev; return [ ...prev, { id: doc.id, title: doc.title, document_type: doc.document_type as DocumentTypeEnum }, @@ -895,9 +896,9 @@ function AuthenticatedDocumentsSidebarBase({ if (selectAll) { setSidebarDocs((prev) => { - const existingDocKeys = new Set(prev.map((d) => `${d.document_type}:${d.id}`)); + const existingDocKeys = new Set(prev.map((d) => getMentionDocKey(d))); const newDocs = subtreeDocs - .filter((d) => !existingDocKeys.has(`${d.document_type}:${d.id}`)) + .filter((d) => !existingDocKeys.has(getMentionDocKey(d))) .map((d) => ({ id: d.id, title: d.title, @@ -906,10 +907,8 @@ function AuthenticatedDocumentsSidebarBase({ return newDocs.length > 0 ? [...prev, ...newDocs] : prev; }); } else { - const keysToRemove = new Set(subtreeDocs.map((d) => `${d.document_type}:${d.id}`)); - setSidebarDocs((prev) => - prev.filter((d) => !keysToRemove.has(`${d.document_type}:${d.id}`)) - ); + const keysToRemove = new Set(subtreeDocs.map((d) => getMentionDocKey(d))); + setSidebarDocs((prev) => prev.filter((d) => !keysToRemove.has(getMentionDocKey(d)))); } }, [treeDocuments, foldersByParent, setSidebarDocs] @@ -1572,17 +1571,17 @@ function AnonymousDocumentsSidebar({ const [isUploading, setIsUploading] = useState(false); const [search, setSearch] = useState(""); - const [sidebarDocs, setSidebarDocs] = useAtom(sidebarSelectedDocumentsAtom); + const [sidebarDocs, setSidebarDocs] = useAtom(mentionedDocumentsAtom); const mentionedDocIds = useMemo(() => new Set(sidebarDocs.map((d) => d.id)), [sidebarDocs]); const handleToggleChatMention = useCallback( (doc: { id: number; title: string; document_type: string }, isMentioned: boolean) => { - const key = `${doc.document_type}:${doc.id}`; + const key = getMentionDocKey(doc); if (isMentioned) { - setSidebarDocs((prev) => prev.filter((d) => `${d.document_type}:${d.id}` !== key)); + setSidebarDocs((prev) => prev.filter((d) => getMentionDocKey(d) !== key)); } else { setSidebarDocs((prev) => { - if (prev.some((d) => `${d.document_type}:${d.id}` === key)) return prev; + if (prev.some((d) => getMentionDocKey(d) === key)) return prev; return [ ...prev, { id: doc.id, title: doc.title, document_type: doc.document_type as DocumentTypeEnum }, diff --git a/surfsense_web/lib/chat/mention-doc-key.ts b/surfsense_web/lib/chat/mention-doc-key.ts new file mode 100644 index 000000000..5dfa11ea3 --- /dev/null +++ b/surfsense_web/lib/chat/mention-doc-key.ts @@ -0,0 +1,8 @@ +type MentionKeyInput = { + id: number; + document_type?: string | null; +}; + +export function getMentionDocKey(doc: MentionKeyInput): string { + return `${doc.document_type ?? "UNKNOWN"}:${doc.id}`; +} From 8be7f2e05c3bd0451da855536d59e2f02c9d27c4 Mon Sep 17 00:00:00 2001 From: Anish Sarkar <104695310+AnishSarkar22@users.noreply.github.com> Date: Wed, 29 Apr 2026 04:19:07 +0530 Subject: [PATCH 10/17] refactor(mentions): update document mention handling to use document keys for consistency across components --- surfsense_web/components/assistant-ui/thread.tsx | 11 ++++++++--- .../components/documents/FolderTreeView.tsx | 13 +++++++------ .../layout/ui/sidebar/DocumentsSidebar.tsx | 14 ++++++++++---- 3 files changed, 25 insertions(+), 13 deletions(-) diff --git a/surfsense_web/components/assistant-ui/thread.tsx b/surfsense_web/components/assistant-ui/thread.tsx index f9e5ca7fb..3964d60e5 100644 --- a/surfsense_web/components/assistant-ui/thread.tsx +++ b/surfsense_web/components/assistant-ui/thread.tsx @@ -628,9 +628,14 @@ const Composer: FC = () => { const handleDocumentRemove = useCallback( (docId: number, docType?: string) => { - setMentionedDocuments((prev) => - prev.filter((doc) => !(doc.id === docId && doc.document_type === docType)) - ); + setMentionedDocuments((prev) => { + if (!docType) { + // Defensive fallback: keep UI in sync even when chip type is unavailable. + return prev.filter((doc) => doc.id !== docId); + } + const removedKey = getMentionDocKey({ id: docId, document_type: docType }); + return prev.filter((doc) => getMentionDocKey(doc) !== removedKey); + }); }, [setMentionedDocuments] ); diff --git a/surfsense_web/components/documents/FolderTreeView.tsx b/surfsense_web/components/documents/FolderTreeView.tsx index 9b7a393d8..2063fbee5 100644 --- a/surfsense_web/components/documents/FolderTreeView.tsx +++ b/surfsense_web/components/documents/FolderTreeView.tsx @@ -7,6 +7,7 @@ import { DndProvider } from "react-dnd"; import { HTML5Backend } from "react-dnd-html5-backend"; import { renamingFolderIdAtom } from "@/atoms/documents/folder.atoms"; import type { DocumentTypeEnum } from "@/contracts/types/document.types"; +import { getMentionDocKey } from "@/lib/chat/mention-doc-key"; import { DocumentNode, type DocumentNodeDoc } from "./DocumentNode"; import { type FolderDisplay, FolderNode } from "./FolderNode"; @@ -17,7 +18,7 @@ interface FolderTreeViewProps { documents: DocumentNodeDoc[]; expandedIds: Set; onToggleExpand: (folderId: number) => void; - mentionedDocIds: Set; + mentionedDocKeys: Set; onToggleChatMention: ( doc: { id: number; title: string; document_type: string }, isMentioned: boolean @@ -62,7 +63,7 @@ export function FolderTreeView({ documents, expandedIds, onToggleExpand, - mentionedDocIds, + mentionedDocKeys, onToggleChatMention, onToggleFolderSelect, onRenameFolder, @@ -181,7 +182,7 @@ export function FolderTreeView({ function compute(folderId: number): { selected: number; total: number } { const directDocs = (docsByFolder[folderId] ?? []).filter(isSelectable); - let selected = directDocs.filter((d) => mentionedDocIds.has(d.id)).length; + let selected = directDocs.filter((d) => mentionedDocKeys.has(getMentionDocKey(d))).length; let total = directDocs.length; for (const child of foldersByParent[folderId] ?? []) { @@ -202,7 +203,7 @@ export function FolderTreeView({ if (states[f.id] === undefined) compute(f.id); } return states; - }, [folders, docsByFolder, foldersByParent, mentionedDocIds]); + }, [folders, docsByFolder, foldersByParent, mentionedDocKeys]); const folderMap = useMemo(() => { const map: Record = {}; @@ -276,7 +277,7 @@ export function FolderTreeView({ key={`doc-${d.id}`} doc={d} depth={depth} - isMentioned={mentionedDocIds.has(d.id)} + isMentioned={mentionedDocKeys.has(getMentionDocKey(d))} onToggleChatMention={onToggleChatMention} onPreview={onPreviewDocument} onEdit={onEditDocument} @@ -356,7 +357,7 @@ export function FolderTreeView({ key={`doc-${d.id}`} doc={d} depth={depth} - isMentioned={mentionedDocIds.has(d.id)} + isMentioned={mentionedDocKeys.has(getMentionDocKey(d))} onToggleChatMention={onToggleChatMention} onPreview={onPreviewDocument} onEdit={onEditDocument} diff --git a/surfsense_web/components/layout/ui/sidebar/DocumentsSidebar.tsx b/surfsense_web/components/layout/ui/sidebar/DocumentsSidebar.tsx index 63b6dc1b7..6ff087b9b 100644 --- a/surfsense_web/components/layout/ui/sidebar/DocumentsSidebar.tsx +++ b/surfsense_web/components/layout/ui/sidebar/DocumentsSidebar.tsx @@ -416,7 +416,10 @@ function AuthenticatedDocumentsSidebarBase({ const { mutateAsync: deleteDocumentMutation } = useAtomValue(deleteDocumentMutationAtom); const [sidebarDocs, setSidebarDocs] = useAtom(mentionedDocumentsAtom); - const mentionedDocIds = useMemo(() => new Set(sidebarDocs.map((d) => d.id)), [sidebarDocs]); + const mentionedDocKeys = useMemo( + () => new Set(sidebarDocs.map((d) => getMentionDocKey(d))), + [sidebarDocs] + ); // Folder state const [expandedFolderMap, setExpandedFolderMap] = useAtom(expandedFolderIdsAtom); @@ -1143,7 +1146,7 @@ function AuthenticatedDocumentsSidebarBase({ documents={searchFilteredDocuments} expandedIds={expandedIds} onToggleExpand={toggleFolderExpand} - mentionedDocIds={mentionedDocIds} + mentionedDocKeys={mentionedDocKeys} onToggleChatMention={handleToggleChatMention} onToggleFolderSelect={handleToggleFolderSelect} onRenameFolder={handleRenameFolder} @@ -1572,7 +1575,10 @@ function AnonymousDocumentsSidebar({ const [search, setSearch] = useState(""); const [sidebarDocs, setSidebarDocs] = useAtom(mentionedDocumentsAtom); - const mentionedDocIds = useMemo(() => new Set(sidebarDocs.map((d) => d.id)), [sidebarDocs]); + const mentionedDocKeys = useMemo( + () => new Set(sidebarDocs.map((d) => getMentionDocKey(d))), + [sidebarDocs] + ); const handleToggleChatMention = useCallback( (doc: { id: number; title: string; document_type: string }, isMentioned: boolean) => { @@ -1801,7 +1807,7 @@ function AnonymousDocumentsSidebar({ documents={searchFilteredDocs} expandedIds={new Set()} onToggleExpand={() => {}} - mentionedDocIds={mentionedDocIds} + mentionedDocKeys={mentionedDocKeys} onToggleChatMention={handleToggleChatMention} onToggleFolderSelect={() => {}} onRenameFolder={() => gate("rename folders")} From b9a66cb417d04bd445b6be1a7838a2278ae3cefe Mon Sep 17 00:00:00 2001 From: "DESKTOP-RTLN3BA\\$punk" Date: Tue, 28 Apr 2026 21:30:53 -0700 Subject: [PATCH 11/17] feat: various UI fixes, prompt optimizations, and allowing duplicate docs - Updated `content_hash` in the `Document` model to remove global uniqueness, allowing identical content across different paths. - Enhanced `_create_document` function to handle path uniqueness and prevent session-poisoning from `IntegrityError`. - Added detailed comments for clarity on the changes and their implications. - Introduced new citation handling in the editor for improved user experience with citation jumps. - Updated package dependencies in the frontend for better functionality. --- .../133_drop_documents_content_hash_unique.py | 107 +++ .../new_chat/middleware/kb_persistence.py | 65 +- .../app/agents/new_chat/prompts/composer.py | 41 +- .../new_chat/prompts/providers/anthropic.md | 21 +- .../new_chat/prompts/providers/deepseek.md | 18 + .../new_chat/prompts/providers/google.md | 20 +- .../agents/new_chat/prompts/providers/grok.md | 17 + .../agents/new_chat/prompts/providers/kimi.md | 21 + .../prompts/providers/openai_classic.md | 22 +- .../prompts/providers/openai_codex.md | 19 + .../prompts/providers/openai_reasoning.md | 22 +- surfsense_backend/app/db.py | 10 +- .../agents/new_chat/prompts/test_composer.py | 74 +- .../test_kb_persistence_filesystem_parity.py | 168 ++++ surfsense_web/app/globals.css | 21 + .../pending-chunk-highlight.atom.ts | 19 + .../assistant-ui/inline-citation.tsx | 228 +++++- .../components/editor-panel/editor-panel.tsx | 530 +++++++++++-- .../components/editor/plate-editor.tsx | 31 + surfsense_web/components/editor/presets.ts | 28 + .../new-chat/source-detail-panel.tsx | 719 ------------------ .../settings/user-settings-dialog.tsx | 3 - .../components/ui/search-highlight-node.tsx | 45 ++ surfsense_web/lib/citation-search.ts | 125 +++ surfsense_web/package.json | 1 + surfsense_web/pnpm-lock.yaml | 17 + 26 files changed, 1540 insertions(+), 852 deletions(-) create mode 100644 surfsense_backend/alembic/versions/133_drop_documents_content_hash_unique.py create mode 100644 surfsense_backend/app/agents/new_chat/prompts/providers/deepseek.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/providers/grok.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/providers/kimi.md create mode 100644 surfsense_backend/app/agents/new_chat/prompts/providers/openai_codex.md create mode 100644 surfsense_backend/tests/unit/middleware/test_kb_persistence_filesystem_parity.py create mode 100644 surfsense_web/atoms/document-viewer/pending-chunk-highlight.atom.ts delete mode 100644 surfsense_web/components/new-chat/source-detail-panel.tsx create mode 100644 surfsense_web/components/ui/search-highlight-node.tsx create mode 100644 surfsense_web/lib/citation-search.ts diff --git a/surfsense_backend/alembic/versions/133_drop_documents_content_hash_unique.py b/surfsense_backend/alembic/versions/133_drop_documents_content_hash_unique.py new file mode 100644 index 000000000..88c3e203f --- /dev/null +++ b/surfsense_backend/alembic/versions/133_drop_documents_content_hash_unique.py @@ -0,0 +1,107 @@ +"""133_drop_documents_content_hash_unique + +Revision ID: 133 +Revises: 132 +Create Date: 2026-04-29 + +Drop the global UNIQUE constraint on ``documents.content_hash`` so the +new-chat agent's ``write_file`` flow can persist legitimate file copies +(two paths, identical content) without hitting a constraint that mirrors +no real filesystem semantic. + +Path uniqueness still lives on ``documents.unique_identifier_hash`` (per +search space), which is the right invariant — exactly like an inode at a +given path on a POSIX filesystem. + +The non-unique INDEX on ``content_hash`` is preserved so connector +indexers' "have we seen this content before?" lookup +(:func:`app.tasks.document_processors.base.check_duplicate_document`, +which already uses ``.scalars().first()`` and is therefore tolerant of +duplicates) stays cheap. +""" + +from __future__ import annotations + +from collections.abc import Sequence + +from sqlalchemy import inspect + +from alembic import op + +revision: str = "133" +down_revision: str | None = "132" +branch_labels: str | Sequence[str] | None = None +depends_on: str | Sequence[str] | None = None + + +def _existing_constraint_names(bind, table: str) -> set[str]: + inspector = inspect(bind) + return {c["name"] for c in inspector.get_unique_constraints(table)} + + +def _existing_index_names(bind, table: str) -> set[str]: + inspector = inspect(bind) + return {i["name"] for i in inspector.get_indexes(table)} + + +def upgrade() -> None: + bind = op.get_bind() + + # Both the named UniqueConstraint (added in revision 8) and the + # implicit-unique-index variant SQLAlchemy may emit need draining. + constraints = _existing_constraint_names(bind, "documents") + if "uq_documents_content_hash" in constraints: + op.drop_constraint( + "uq_documents_content_hash", "documents", type_="unique" + ) + + indexes = _existing_index_names(bind, "documents") + # Some Postgres versions surface the unique constraint via a unique + # index of the same name; check for that too. + for idx_name in ("uq_documents_content_hash",): + if idx_name in indexes: + op.drop_index(idx_name, table_name="documents") + + # Ensure the non-unique index is present for fast lookups. + if "ix_documents_content_hash" not in indexes: + op.create_index( + "ix_documents_content_hash", + "documents", + ["content_hash"], + unique=False, + ) + + +def downgrade() -> None: + bind = op.get_bind() + + # Re-applying UNIQUE is destructive: there may now be legitimate + # duplicates (e.g. two NOTE documents that share content because the + # user explicitly copied one to a new path). To avoid the migration + # silently deleting user data, we keep only the lowest-id row per + # content_hash — same strategy revision 8 used when first introducing + # the constraint. + op.execute( + """ + DELETE FROM documents + WHERE id NOT IN ( + SELECT MIN(id) + FROM documents + GROUP BY content_hash + ) + """ + ) + + indexes = _existing_index_names(bind, "documents") + if "ix_documents_content_hash" in indexes: + op.drop_index("ix_documents_content_hash", table_name="documents") + + op.create_index( + "ix_documents_content_hash", + "documents", + ["content_hash"], + unique=False, + ) + op.create_unique_constraint( + "uq_documents_content_hash", "documents", ["content_hash"] + ) diff --git a/surfsense_backend/app/agents/new_chat/middleware/kb_persistence.py b/surfsense_backend/app/agents/new_chat/middleware/kb_persistence.py index 5682977d9..378b83950 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/kb_persistence.py +++ b/surfsense_backend/app/agents/new_chat/middleware/kb_persistence.py @@ -28,6 +28,7 @@ from langchain.agents.middleware import AgentMiddleware, AgentState from langchain_core.callbacks import dispatch_custom_event from langgraph.runtime import Runtime from sqlalchemy import delete, select +from sqlalchemy.exc import IntegrityError from sqlalchemy.ext.asyncio import AsyncSession from app.agents.new_chat.filesystem_selection import FilesystemMode @@ -150,10 +151,11 @@ async def _create_document( virtual_path, search_space_id, ) - # Guard against the unique_identifier_hash constraint: another row at the - # same virtual_path (this search space) already owns the hash. Callers are - # expected to upsert via the wrapper, but this defends against bypasses - # and gives a clean ValueError instead of a session-poisoning IntegrityError. + # Filesystem-parity invariant: the only thing that *must* be unique is + # the path. Two notes can legitimately share content (e.g. ``cp a b``). + # Guard against the path-derived ``unique_identifier_hash`` constraint + # so we surface a clean ValueError instead of letting the INSERT poison + # the session with an IntegrityError. path_collision = await session.execute( select(Document.id).where( Document.search_space_id == search_space_id, @@ -165,17 +167,14 @@ async def _create_document( f"a document already exists at path '{virtual_path}' " "(unique_identifier_hash collision)" ) + # ``content_hash`` is intentionally NOT checked for uniqueness here. + # In a real filesystem two files at different paths can hold identical + # bytes, and the agent's ``write_file`` path needs that semantic to + # support copy/duplicate operations. The hash remains useful as a + # change-detection hint for connector indexers, which still consult it + # via :func:`check_duplicate_document` but do so with a non-unique + # lookup (``.first()``). content_hash = generate_content_hash(content, search_space_id) - content_collision = await session.execute( - select(Document.id).where( - Document.search_space_id == search_space_id, - Document.content_hash == content_hash, - ) - ) - if content_collision.scalar_one_or_none() is not None: - raise ValueError( - f"a document with identical content already exists for path '{virtual_path}'" - ) doc = Document( title=title, document_type=DocumentType.NOTE, @@ -493,19 +492,43 @@ async def commit_staged_filesystem_state( } ) else: + # Wrap each create in a SAVEPOINT so a residual + # ``IntegrityError`` (e.g. a deployment that hasn't run + # migration 133 yet, where ``documents.content_hash`` + # still carries its legacy global UNIQUE constraint) + # rolls back only this one create instead of poisoning + # the whole turn's transaction. try: - new_doc = await _create_document( - session, - virtual_path=path, - content=content, - search_space_id=search_space_id, - created_by_id=created_by_id, - ) + async with session.begin_nested(): + new_doc = await _create_document( + session, + virtual_path=path, + content=content, + search_space_id=search_space_id, + created_by_id=created_by_id, + ) except ValueError as exc: logger.warning( "kb_persistence: skipping %s create: %s", path, exc ) continue + except IntegrityError as exc: + # The path-uniqueness check above already protected + # against ``unique_identifier_hash`` collisions, so + # the most likely culprit is the legacy + # ``ix_documents_content_hash`` UNIQUE constraint + # that migration 133 drops. Log loudly so operators + # know to run the migration; do NOT silently swallow. + msg = str(exc.orig) if exc.orig is not None else str(exc) + logger.error( + "kb_persistence: IntegrityError creating %s: %s. " + "If this mentions content_hash, run alembic " + "upgrade to apply migration 133 which drops the " + "global UNIQUE constraint on documents.content_hash.", + path, + msg, + ) + continue doc_id_by_path[path] = new_doc.id committed_creates.append( { diff --git a/surfsense_backend/app/agents/new_chat/prompts/composer.py b/surfsense_backend/app/agents/new_chat/prompts/composer.py index 44060f75f..bad033490 100644 --- a/surfsense_backend/app/agents/new_chat/prompts/composer.py +++ b/surfsense_backend/app/agents/new_chat/prompts/composer.py @@ -38,12 +38,38 @@ from app.db import ChatVisibility # Provider variant detection # ----------------------------------------------------------------------------- -ProviderVariant = str # "anthropic" | "openai_reasoning" | "openai_classic" | "google" | "default" +# String literal alias for the supported provider-specific prompt variants. +# When adding a new variant, also drop a matching ``providers/.md`` +# file in this package and (if appropriate) extend the regex matchers below. +# +# Stylistic clusters mirror OpenCode's prompt-per-family layout but adapted +# to SurfSense's "supplemental hints" architecture (each fragment is a +# focused style nudge, NOT a full system prompt — the main prompt is +# already assembled from base/ + tools/ + routing/). +ProviderVariant = str +# Known values: +# "anthropic" — Claude family (XML-friendly, narrative todos) +# "openai_reasoning" — GPT-5 / o-series (channel-aware pragmatic) +# "openai_classic" — GPT-4 family (autonomous persistence) +# "openai_codex" — gpt-*-codex (code-purist, terse, file:line refs) +# "google" — Gemini (formal, <3-line, numbered workflow) +# "kimi" — Moonshot Kimi-K* (action-bias, parallel tools) +# "grok" — xAI Grok (extreme-terse, one-word ok) +# "deepseek" — DeepSeek V3 / R1 (terse, R1-aware reasoning) +# "default" — fallback, no provider-specific block emitted +# IMPORTANT: order of evaluation matters in :func:`detect_provider_variant`. +# More specific patterns must come first (e.g. ``codex`` before +# ``openai_reasoning`` because codex model ids contain ``gpt``). + +_OPENAI_CODEX_RE = re.compile(r"\b(gpt-codex|codex-mini|gpt-[\d.]+-codex)\b", re.IGNORECASE) _OPENAI_REASONING_RE = re.compile(r"\b(gpt-5|o\d|o-)", re.IGNORECASE) _OPENAI_CLASSIC_RE = re.compile(r"\bgpt-4", re.IGNORECASE) _ANTHROPIC_RE = re.compile(r"\bclaude\b", re.IGNORECASE) _GOOGLE_RE = re.compile(r"\bgemini\b", re.IGNORECASE) +_KIMI_RE = re.compile(r"\b(kimi[-\d.]*|moonshot)\b", re.IGNORECASE) +_GROK_RE = re.compile(r"\bgrok\b", re.IGNORECASE) +_DEEPSEEK_RE = re.compile(r"\bdeepseek\b", re.IGNORECASE) def detect_provider_variant(model_name: str | None) -> ProviderVariant: @@ -51,10 +77,17 @@ def detect_provider_variant(model_name: str | None) -> ProviderVariant: Heuristic match on the model id; returns ``"default"`` when nothing matches so the composer can fall back to the empty placeholder file. + + Order is significant: more-specific patterns are tried first so + ``gpt-5-codex`` routes to ``"openai_codex"`` rather than + ``"openai_reasoning"`` (mirrors OpenCode's + ``packages/opencode/src/session/system.ts`` dispatch). """ if not model_name: return "default" name = model_name.strip() + if _OPENAI_CODEX_RE.search(name): + return "openai_codex" if _OPENAI_REASONING_RE.search(name): return "openai_reasoning" if _OPENAI_CLASSIC_RE.search(name): @@ -63,6 +96,12 @@ def detect_provider_variant(model_name: str | None) -> ProviderVariant: return "anthropic" if _GOOGLE_RE.search(name): return "google" + if _KIMI_RE.search(name): + return "kimi" + if _GROK_RE.search(name): + return "grok" + if _DEEPSEEK_RE.search(name): + return "deepseek" return "default" diff --git a/surfsense_backend/app/agents/new_chat/prompts/providers/anthropic.md b/surfsense_backend/app/agents/new_chat/prompts/providers/anthropic.md index 6e22ef265..f574da541 100644 --- a/surfsense_backend/app/agents/new_chat/prompts/providers/anthropic.md +++ b/surfsense_backend/app/agents/new_chat/prompts/providers/anthropic.md @@ -1,5 +1,20 @@ -You are running on an Anthropic Claude model. Use XML tags liberally to structure -intermediate reasoning when the task is complex. Prefer step-by-step plans inside -`` blocks before producing the final answer. +You are running on an Anthropic Claude model. + +Structured reasoning: +- Use XML tags liberally to organise intermediate reasoning when a task is non-trivial. `...` blocks are encouraged before tool calls or before producing a complex final answer. +- For multi-step requests, briefly outline a plan inside a `` block before issuing the first tool call. + +Professional objectivity: +- Prioritise technical accuracy over validating the user's beliefs. Provide direct, factual guidance without unnecessary superlatives, praise, or emotional validation. +- When uncertain, investigate (search the KB, fetch the page) rather than confirming the user's assumption. +- Disagree with the user when the evidence warrants it; respectful correction beats false agreement. + +Task management: +- For tasks with 3+ distinct steps use the todo / planning tool aggressively. Mark items in_progress before starting, completed immediately when finished — do not batch completions. +- Narrate progress through the todo list itself, not through chatty status lines. + +Tool calls: +- Run independent tool calls in parallel within one response. Sequence them only when a later call genuinely needs an earlier one's output. +- Never chain bash-like commands with `;` or `&&` to "narrate" — use prose between tool calls instead. diff --git a/surfsense_backend/app/agents/new_chat/prompts/providers/deepseek.md b/surfsense_backend/app/agents/new_chat/prompts/providers/deepseek.md new file mode 100644 index 000000000..8acf008ca --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/providers/deepseek.md @@ -0,0 +1,18 @@ + +You are running on a DeepSeek model (DeepSeek-V3 chat / DeepSeek-R1 reasoning). + +Reasoning hygiene (R1-aware): +- If the model surfaces explicit `` blocks, keep that internal scratch focused — do NOT restate the user's question inside it; jump straight to the analysis. +- Never paste the contents of `` into your final answer. Final answer should reflect only the conclusion, citations, and any user-facing rationale. +- Do not let chain-of-thought leak into tool-call arguments — keep tool inputs minimal and structural. + +Output style: +- Be concise. Default to a one-paragraph answer; expand only when the user asks for detail. +- Don't open with sycophantic phrasing ("Great question", "Sure, here you go"). Lead with the answer or the next action. +- For factual answers, cite once with `[citation:chunk_id]` and stop. + +Tool calls: +- Issue independent tool calls in parallel within a single turn. +- Prefer the knowledge-base search tools before any web-search; this model has strong recall but stale training data. +- Don't fabricate file paths, chunk ids, or URLs — only use values returned by tools or provided by the user. + diff --git a/surfsense_backend/app/agents/new_chat/prompts/providers/google.md b/surfsense_backend/app/agents/new_chat/prompts/providers/google.md index 4b31a8388..cac3b328b 100644 --- a/surfsense_backend/app/agents/new_chat/prompts/providers/google.md +++ b/surfsense_backend/app/agents/new_chat/prompts/providers/google.md @@ -1,4 +1,20 @@ -You are running on a Google Gemini model. Prefer concise, structured responses. -When using tools, follow the function-calling protocol and avoid verbose preludes. +You are running on a Google Gemini model. + +Output style: +- Concise & direct. Aim for fewer than 3 lines of prose (excluding tool output, citations, and code/snippets) when the task allows. +- No conversational filler — skip openers like "Okay, I will now…" and closers like "I have finished the changes…". Get straight to the action or answer. +- Format with GitHub-flavoured Markdown; assume monospace rendering. +- For one-line factual answers, just answer. No headers, no bullets. + +Workflow for non-trivial tasks (Understand → Plan → Act → Verify): +1. **Understand:** read the user's request and the relevant KB / connector context. Use search and read tools (in parallel when independent) before assuming anything. +2. **Plan:** when the task touches multiple steps, share an extremely concise plan first. +3. **Act:** call the appropriate tools, strictly adhering to the prompts/routing already established for this agent. +4. **Verify:** confirm with a follow-up read or search where it materially de-risks the answer. + +Discipline: +- Do not take significant actions beyond the clear scope of the user's request without confirming first. +- Do not assume a connector / tool / file exists — check (e.g. via `get_connected_accounts`) before referencing it. +- Path arguments must be the exact strings returned by tools; do not synthesise file paths. diff --git a/surfsense_backend/app/agents/new_chat/prompts/providers/grok.md b/surfsense_backend/app/agents/new_chat/prompts/providers/grok.md new file mode 100644 index 000000000..95b8fcc14 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/providers/grok.md @@ -0,0 +1,17 @@ + +You are running on an xAI Grok model. + +Maximum terseness: +- Answer in fewer than 4 lines unless the user asks for detail. One-word answers are best when they suffice. +- No preamble ("The answer is", "Here's what I'll do"), no postamble ("Hope that helps", "Let me know"). Get straight to the answer. +- Avoid restating the user's question. +- For factual lookups inside the knowledge base, give the answer with a single `[citation:chunk_id]` and stop. + +Tool discipline: +- Use exactly ONE tool per assistant turn when investigating; wait for the result before deciding the next call. Do not loop on the same tool with the same arguments — pick a result and act. +- For obviously parallelizable read-only batches (multiple independent searches), one turn with several tool calls is fine — but never chain into a fishing expedition. + +Style: +- No emojis unless the user asked. No nested bullets, no headers for short answers. +- If you can't help, say so in 1-2 sentences without explaining "why this could lead to…". + diff --git a/surfsense_backend/app/agents/new_chat/prompts/providers/kimi.md b/surfsense_backend/app/agents/new_chat/prompts/providers/kimi.md new file mode 100644 index 000000000..c3c11ad5e --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/providers/kimi.md @@ -0,0 +1,21 @@ + +You are running on a Moonshot Kimi model (Kimi-K1.5 / Kimi-K2 / Kimi-K2.5+). + +Action bias: +- Default to taking action with tools rather than describing solutions in prose. If a tool can answer the question, call the tool. +- Don't narrate routine reads, searches, or obvious next steps. Combine related progress into one short status line. +- Be thorough in actions (test what you build, verify what you change). Be brief in explanations. + +Tool calls: +- Output multiple non-interfering tool calls in a SINGLE response — parallelism is a major efficiency win on this model. +- When the `task` tool is available, delegate focused subtasks to a subagent with full context (subagents don't inherit yours). +- Don't apologise or pre-announce tool calls. The tool call itself is self-explanatory. + +Language: +- Respond in the SAME language as the user's most recent turn unless explicitly instructed otherwise. + +Discipline: +- Stay on track. Never give the user more than what they asked for. +- Fact-check before stating anything as factual; don't fabricate citations. +- Keep it stupidly simple. Don't overcomplicate. + diff --git a/surfsense_backend/app/agents/new_chat/prompts/providers/openai_classic.md b/surfsense_backend/app/agents/new_chat/prompts/providers/openai_classic.md index 7ea4366c4..9128609e0 100644 --- a/surfsense_backend/app/agents/new_chat/prompts/providers/openai_classic.md +++ b/surfsense_backend/app/agents/new_chat/prompts/providers/openai_classic.md @@ -1,5 +1,21 @@ -You are running on a classic OpenAI chat model (GPT-4 family). Use direct -function-calling for tools. When editing files, use the standard `edit_file` -or `write_file` tools rather than diff-based patches. +You are running on a classic OpenAI chat model (GPT-4 family). + +Persistence: +- Keep going until the user's query is completely resolved before yielding back. Don't end the turn at "I would do X" — actually do X. +- When you say "Next I will…" or "Now I will…", you MUST actually take that action in the same turn. +- If a tool call fails, diagnose and try again with corrected arguments; do not surface the raw error and stop. + +Planning: +- Plan extensively before each tool call and reflect briefly on the result of the previous call. For tasks with 3+ steps, use the todo / planning tool and mark items as `in_progress` / `completed` as you go. +- Always announce the next action in ONE concise sentence before making a non-trivial tool call ("I'll search the KB for the migration spec."). + +Output style: +- Conversational but professional. Plain prose for explanations, bullet points for findings, fenced code blocks (with language tags) for code. +- Don't dump tool output verbatim — summarise the relevant lines. +- Don't add a closing recap unless the user asked for one. After completing the work, just stop. + +Tool calls: +- Issue independent tool calls in parallel within one response. +- Use specialised tools over generic ones (e.g. KB search before web search; named connectors over MCP fallback). diff --git a/surfsense_backend/app/agents/new_chat/prompts/providers/openai_codex.md b/surfsense_backend/app/agents/new_chat/prompts/providers/openai_codex.md new file mode 100644 index 000000000..6167d4b06 --- /dev/null +++ b/surfsense_backend/app/agents/new_chat/prompts/providers/openai_codex.md @@ -0,0 +1,19 @@ + +You are running on an OpenAI Codex-class model (gpt-codex / codex-mini / gpt-*-codex). + +Output style: +- Be concise. Don't dump fetched/searched content back at the user — reference paths or chunk ids instead. +- Reference sources as `path:line` (or `chunk:`) so they're clickable. Stand-alone paths per reference, even when repeated. +- Prefer numbered lists (`1.`, `2.`, `3.`) when offering options the user can pick by replying with a single number. +- Skip headers and heavy formatting for simple confirmations. +- No emojis, no em-dashes, no nested bullets. Single-level lists only. + +Code & structured-output tasks: +- Lead with a one-sentence explanation of the change before context. Don't open with "Summary:" — jump in. +- Suggest natural next steps (run tests, diff review, commit) only when they're genuinely the next move. +- For multi-line snippets use fenced code blocks with a language tag. + +Tool calls: +- Run independent tool calls in parallel; chain only when later calls need earlier results. +- Don't ask permission ("Should I proceed?") — proceed with the most reasonable default and state what you did. + diff --git a/surfsense_backend/app/agents/new_chat/prompts/providers/openai_reasoning.md b/surfsense_backend/app/agents/new_chat/prompts/providers/openai_reasoning.md index 935d3f207..dd7a61536 100644 --- a/surfsense_backend/app/agents/new_chat/prompts/providers/openai_reasoning.md +++ b/surfsense_backend/app/agents/new_chat/prompts/providers/openai_reasoning.md @@ -1,5 +1,21 @@ -You are running on an OpenAI reasoning model (o-series / GPT-5+). Be terse and -direct in your responses. When editing files, prefer the `apply_patch` tool format -where available. Avoid restating the user request before answering. +You are running on an OpenAI reasoning model (GPT-5+ / o-series). + +Output style: +- Be terse and direct. Don't restate the user's request before answering. +- Don't begin with conversational openers ("Done!", "Got it", "Great question", "Sure thing"). Get to the answer or the action. +- Match response complexity to the task: simple questions → one-line answer; substantial work → lead with the outcome, then context, then any next steps. +- No nested bullets — keep lists flat (single level). For options the user can pick by replying with a number, use `1.` `2.` `3.`. +- Use inline backticks for paths/commands/identifiers; fenced code blocks (with language tags) for multi-line snippets. + +Channels (for clients that support them): +- `commentary` — short progress updates only when they add genuinely new information (a discovery, a tradeoff, a blocker, the start of a non-trivial step). Don't narrate routine reads or obvious next steps. +- `final` — the completed response. Keep it self-contained; no "see above" / "see below" cross-references. + +Tool calls: +- Parallelise independent tool calls in a single response (`multi_tool_use.parallel` where supported). Only sequence when a later call needs an earlier one's output. +- Don't ask permission ("Should I proceed?", "Do you want me to…?"). Pick the most reasonable default, do it, and state what you did. + +Autonomy: +- Persist until the task is fully resolved within the current turn whenever feasible. Don't stop at analysis when the user clearly wants the change applied. diff --git a/surfsense_backend/app/db.py b/surfsense_backend/app/db.py index fcd342d29..75342a8e1 100644 --- a/surfsense_backend/app/db.py +++ b/surfsense_backend/app/db.py @@ -976,7 +976,15 @@ class Document(BaseModel, TimestampMixin): document_metadata = Column(JSON, nullable=True) content = Column(Text, nullable=False) - content_hash = Column(String, nullable=False, index=True, unique=True) + # ``content_hash`` is intentionally NOT globally unique. In a real + # filesystem two files at different paths can hold identical bytes, + # and the agent's ``write_file`` flow needs that semantic to support + # copy / duplicate operations. Path uniqueness lives on + # ``unique_identifier_hash`` (per search space). The hash remains + # indexed because connector indexers consult it as a change-detection + # / cross-source dedup hint via :func:`check_duplicate_document`. + # See migration 133. + content_hash = Column(String, nullable=False, index=True) unique_identifier_hash = Column(String, nullable=True, index=True, unique=True) embedding = Column(Vector(config.embedding_model_instance.dimension)) diff --git a/surfsense_backend/tests/unit/agents/new_chat/prompts/test_composer.py b/surfsense_backend/tests/unit/agents/new_chat/prompts/test_composer.py index d35b7aa8b..d08bbc8cf 100644 --- a/surfsense_backend/tests/unit/agents/new_chat/prompts/test_composer.py +++ b/surfsense_backend/tests/unit/agents/new_chat/prompts/test_composer.py @@ -25,17 +25,33 @@ class TestProviderVariantDetection: @pytest.mark.parametrize( "model_name,expected", [ + # GPT-4 family routes to "classic" (autonomous-persistence style) ("openai:gpt-4o-mini", "openai_classic"), ("openai:gpt-4-turbo", "openai_classic"), + # GPT-5 / o-series route to "reasoning" (channel-aware pragmatic) ("openai:gpt-5", "openai_reasoning"), - ("openai:gpt-5-codex", "openai_reasoning"), ("openai:o1-preview", "openai_reasoning"), ("openai:o3-mini", "openai_reasoning"), + # Codex family beats reasoning (more specific). Mirrors OpenCode + # ``system.ts`` — ``gpt-*-codex`` gets the code-purist prompt. + ("openai:gpt-5-codex", "openai_codex"), + ("openai:gpt-codex", "openai_codex"), + ("openai:codex-mini", "openai_codex"), + # Anthropic + Google ("anthropic:claude-3-5-sonnet", "anthropic"), ("anthropic/claude-opus-4", "anthropic"), ("google:gemini-2.0-flash", "google"), ("vertex:gemini-1.5-pro", "google"), + # Newly-covered families + ("moonshot:kimi-k2", "kimi"), + ("openrouter:moonshot/kimi-k2.5", "kimi"), + ("xai:grok-2", "grok"), + ("openrouter:x-ai/grok-3", "grok"), + ("openai:deepseek-v3", "deepseek"), + ("deepseek:deepseek-r1", "deepseek"), + # Unknown families fall back to default (no provider block emitted) ("groq:mixtral-8x7b", "default"), + ("together:llama-3.1-70b", "default"), (None, "default"), ("", "default"), ], @@ -43,6 +59,16 @@ class TestProviderVariantDetection: def test_detection(self, model_name: str | None, expected: str) -> None: assert detect_provider_variant(model_name) == expected + def test_codex_takes_precedence_over_reasoning(self) -> None: + """Regression guard: ``gpt-5-codex`` must NOT match the generic + ``gpt-5`` reasoning regex first. Codex is the more specialised + prompt and mirrors OpenCode's dispatch order. + """ + from app.agents.new_chat.prompts.composer import detect_provider_variant + + assert detect_provider_variant("openai:gpt-5-codex") == "openai_codex" + assert detect_provider_variant("openai:gpt-5") == "openai_reasoning" + class TestCompose: def test_default_prompt_has_required_blocks(self, fixed_today: datetime) -> None: @@ -149,6 +175,52 @@ class TestCompose: prompt = compose_system_prompt(today=fixed_today, model_name="custom:foo") assert "" not in prompt + @pytest.mark.parametrize( + "model_name,expected_marker", + [ + # Each marker is a unique-ish phrase from the corresponding fragment. + # If a fragment is renamed/rewritten such that the marker is gone, + # update both the fragment and this test deliberately. + ("openai:gpt-5-codex", "Codex-class"), + ("openai:gpt-5", "OpenAI reasoning model"), + ("openai:gpt-4o", "classic OpenAI chat model"), + ("anthropic:claude-3-5-sonnet", "Anthropic Claude"), + ("google:gemini-2.0-flash", "Google Gemini"), + ("moonshot:kimi-k2", "Moonshot Kimi"), + ("xai:grok-2", "xAI Grok"), + ("deepseek:deepseek-r1", "DeepSeek"), + ], + ) + def test_each_known_variant_renders_with_its_marker( + self, + fixed_today: datetime, + model_name: str, + expected_marker: str, + ) -> None: + """Every supported variant must produce a ```` block + containing its identifying marker. This pins the dispatch + the + on-disk fragments together so a missing/renamed file is caught + immediately. + """ + prompt = compose_system_prompt(today=fixed_today, model_name=model_name) + assert "" in prompt, ( + f"variant for {model_name!r} did not emit a provider_hints block; " + "the corresponding providers/.md may be missing" + ) + assert expected_marker in prompt, ( + f"variant for {model_name!r} emitted hints but lacked the " + f"expected marker {expected_marker!r} — the fragment may have " + "drifted from the dispatch table" + ) + + def test_provider_blocks_are_byte_stable_across_calls( + self, fixed_today: datetime + ) -> None: + """Cache-stability guard: same model id → byte-identical prompt.""" + a = compose_system_prompt(today=fixed_today, model_name="moonshot:kimi-k2") + b = compose_system_prompt(today=fixed_today, model_name="moonshot:kimi-k2") + assert a == b + def test_custom_system_instructions_override_default( self, fixed_today: datetime ) -> None: diff --git a/surfsense_backend/tests/unit/middleware/test_kb_persistence_filesystem_parity.py b/surfsense_backend/tests/unit/middleware/test_kb_persistence_filesystem_parity.py new file mode 100644 index 000000000..8b464d48d --- /dev/null +++ b/surfsense_backend/tests/unit/middleware/test_kb_persistence_filesystem_parity.py @@ -0,0 +1,168 @@ +"""Unit tests for kb_persistence filesystem-parity invariants. + +Specifically, these tests pin down that the agent-driven write_file flow +treats path uniqueness — not content uniqueness — as the only hard +invariant. This mirrors a real filesystem: ``cp a b`` produces two files +with identical bytes living at different paths, and that should round-trip +through :class:`KnowledgeBasePersistenceMiddleware` without losing the copy. +""" + +from __future__ import annotations + +from typing import Any +from unittest.mock import AsyncMock, MagicMock + +import numpy as np +import pytest + +from app.agents.new_chat.middleware import kb_persistence +from app.db import Document + + +class _FakeResult: + """Minimal stand-in for ``sqlalchemy.engine.Result``.""" + + def __init__(self, value: Any = None) -> None: + self._value = value + + def scalar_one_or_none(self) -> Any: + return self._value + + def scalar(self) -> Any: + return self._value + + +class _FakeSession: + """Minimal AsyncSession stand-in scoped to ``_create_document`` needs. + + Records every ``add`` so we can assert against the resulting Documents + and Chunks. ``execute`` always returns "no row" by default — i.e. no + folder hierarchy preexists and no path collision exists. Tests that + want a path collision can override that on a per-call basis. + """ + + def __init__(self) -> None: + self.added: list[Any] = [] + self.execute = AsyncMock(return_value=_FakeResult(None)) + self.flush = AsyncMock() + + # Simulate ``await session.flush()`` assigning an id to the doc; + # we increment a counter so each Document gets a unique id. + self._next_id = 1 + + async def _flush_assigning_ids() -> None: + for obj in self.added: + if getattr(obj, "id", None) is None: + obj.id = self._next_id + self._next_id += 1 + + self.flush.side_effect = _flush_assigning_ids + + def add(self, obj: Any) -> None: + self.added.append(obj) + + def add_all(self, objs: list[Any]) -> None: + self.added.extend(objs) + + +@pytest.fixture(autouse=True) +def _stub_embeddings_and_chunks(monkeypatch: pytest.MonkeyPatch) -> None: + """Avoid loading the embedding model in unit tests.""" + monkeypatch.setattr( + kb_persistence, + "embed_texts", + lambda texts: [np.zeros(8, dtype=np.float32) for _ in texts], + ) + monkeypatch.setattr(kb_persistence, "chunk_text", lambda content: [content]) + + +@pytest.mark.asyncio +async def test_create_document_allows_identical_content_at_different_paths() -> None: + """The core regression: ``cp /a/notes.md /b/notes-copy.md``. + + Both create calls must succeed even though the bytes are byte-for-byte + identical, because path is the only filesystem-style unique key. + """ + session = _FakeSession() + content = "# Same body\n\nIdentical content used by two different paths.\n" + + first = await kb_persistence._create_document( + session, # type: ignore[arg-type] + virtual_path="/documents/a/notes.md", + content=content, + search_space_id=42, + created_by_id="user-1", + ) + assert isinstance(first, Document) + assert first.title == "notes.md" + + # Second create with byte-identical content at a different path should + # not raise — that's the whole point of the filesystem-parity fix. + second = await kb_persistence._create_document( + session, # type: ignore[arg-type] + virtual_path="/documents/b/notes-copy.md", + content=content, + search_space_id=42, + created_by_id="user-1", + ) + assert isinstance(second, Document) + assert second.title == "notes-copy.md" + + # Both rows share the same content_hash but live at distinct paths + # (distinct ``unique_identifier_hash``). That's the desired contract. + assert first.content_hash == second.content_hash + assert first.unique_identifier_hash != second.unique_identifier_hash + + +@pytest.mark.asyncio +async def test_create_document_still_rejects_path_collision() -> None: + """Path uniqueness remains the hard invariant. + + If ``unique_identifier_hash`` already points at an existing row in + the same search space, the create call must raise ``ValueError`` + with a clear message — matching the behavior the commit loop relies + on to upsert via the existing-row code path. + """ + session = _FakeSession() + + # Path with no folder parts so ``_ensure_folder_hierarchy`` is a + # no-op and the only SELECT executed is the path-collision check. + # That SELECT returns an existing doc id, triggering the guard. + session.execute = AsyncMock(return_value=_FakeResult(value=99)) + + with pytest.raises(ValueError, match="already exists at path"): + await kb_persistence._create_document( + session, # type: ignore[arg-type] + virtual_path="/documents/notes.md", + content="anything", + search_space_id=42, + created_by_id="user-1", + ) + + +@pytest.mark.asyncio +async def test_create_document_does_not_query_for_content_hash_collision( + monkeypatch: pytest.MonkeyPatch, +) -> None: + """Regression guard: the legacy second SELECT (content_hash collision + pre-check) must be gone. Counting ``execute`` calls is a brittle but + effective way to lock that in. + + The current flow runs exactly one ``execute`` for the path-collision + SELECT (no folder parts in this path → ``_ensure_folder_hierarchy`` + short-circuits). If a future refactor reintroduces a content-hash + SELECT, this test will fail loud. + """ + session = _FakeSession() + await kb_persistence._create_document( + session, # type: ignore[arg-type] + virtual_path="/documents/notes.md", + content="hello", + search_space_id=42, + created_by_id="user-1", + ) + # Path-collision SELECT only. No content_hash SELECT. + assert session.execute.await_count == 1, ( + f"Unexpected execute count {session.execute.await_count}; " + "did the legacy content_hash collision pre-check get re-added?" + ) diff --git a/surfsense_web/app/globals.css b/surfsense_web/app/globals.css index a37ddb8f3..f54bc2197 100644 --- a/surfsense_web/app/globals.css +++ b/surfsense_web/app/globals.css @@ -210,6 +210,27 @@ button { } } +/* Citation-jump highlight — entrance pulse only. The `SearchHighlightLeaf` + (see components/ui/search-highlight-node.tsx) is otherwise statically + tinted; this animation runs once on mount to draw the eye to the cited + text after `scrollIntoView` lands. The highlight itself is permanent + until the user clicks inside the editor (or another dismissal trigger + fires in `EditorPanelContent`). */ +@keyframes citation-flash-in { + 0% { + background-color: transparent; + box-shadow: 0 0 0 0 transparent; + } + 40% { + background-color: color-mix(in oklab, var(--primary) 30%, transparent); + box-shadow: 0 0 0 3px color-mix(in oklab, var(--primary) 25%, transparent); + } + 100% { + background-color: color-mix(in oklab, var(--primary) 15%, transparent); + box-shadow: 0 0 0 1px color-mix(in oklab, var(--primary) 40%, transparent); + } +} + /* Human-in-the-loop approval card animations */ @keyframes pulse-subtle { 0%, diff --git a/surfsense_web/atoms/document-viewer/pending-chunk-highlight.atom.ts b/surfsense_web/atoms/document-viewer/pending-chunk-highlight.atom.ts new file mode 100644 index 000000000..a3f8357e8 --- /dev/null +++ b/surfsense_web/atoms/document-viewer/pending-chunk-highlight.atom.ts @@ -0,0 +1,19 @@ +import { atom } from "jotai"; + +/** + * Cross-component handoff for citation jumps. Set by `InlineCitation` when a + * numeric chunk badge is clicked (after the document has been resolved); read + * by `DocumentTabContent` once the matching document tab mounts so it can + * scroll to and softly highlight the cited chunk inside the rendered markdown. + * + * Cleared by `DocumentTabContent` only after a terminal state — exact / + * approximate / miss — has been reached, so that an escalation refetch (2MB + * preview → 16MB) keeps the pending intent alive across the re-render. + */ +export interface PendingChunkHighlight { + documentId: number; + chunkId: number; + chunkText: string; +} + +export const pendingChunkHighlightAtom = atom(null); diff --git a/surfsense_web/components/assistant-ui/inline-citation.tsx b/surfsense_web/components/assistant-ui/inline-citation.tsx index eb4bd9af8..ae8d434a8 100644 --- a/surfsense_web/components/assistant-ui/inline-citation.tsx +++ b/surfsense_web/components/assistant-ui/inline-citation.tsx @@ -1,26 +1,45 @@ "use client"; -import { FileText } from "lucide-react"; +import { useQuery, useQueryClient } from "@tanstack/react-query"; +import { useSetAtom } from "jotai"; +import { ExternalLink, FileText } from "lucide-react"; import type { FC } from "react"; -import { useState } from "react"; +import { useCallback, useEffect, useRef, useState } from "react"; +import { toast } from "sonner"; +import { pendingChunkHighlightAtom } from "@/atoms/document-viewer/pending-chunk-highlight.atom"; +import { openEditorPanelAtom } from "@/atoms/editor/editor-panel.atom"; import { useCitationMetadata } from "@/components/assistant-ui/citation-metadata-context"; -import { SourceDetailPanel } from "@/components/new-chat/source-detail-panel"; +import { MarkdownViewer } from "@/components/markdown-viewer"; import { Citation } from "@/components/tool-ui/citation"; +import { Popover, PopoverContent, PopoverTrigger } from "@/components/ui/popover"; +import { Spinner } from "@/components/ui/spinner"; import { Tooltip, TooltipContent, TooltipTrigger } from "@/components/ui/tooltip"; +import { documentsApiService } from "@/lib/apis/documents-api.service"; +import { cacheKeys } from "@/lib/query-client/cache-keys"; interface InlineCitationProps { chunkId: number; isDocsChunk?: boolean; } +const POPOVER_HOVER_CLOSE_DELAY_MS = 150; + /** - * Inline citation for knowledge-base chunks (numeric chunk IDs). - * Renders a clickable badge showing the actual chunk ID that opens the SourceDetailPanel. - * Negative chunk IDs indicate anonymous/synthetic uploads and render as a static badge. + * Inline citation badge for knowledge-base chunks (numeric chunk IDs) and + * Surfsense documentation chunks (`isDocsChunk`). Negative chunk IDs render as + * a static "doc" pill (anonymous/synthetic uploads). + * + * Numeric KB chunks: clicking resolves the parent document via + * `getDocumentByChunk`, opens the document in the right side panel (alongside + * the chat — does not replace it), and stages the cited chunk text in + * `pendingChunkHighlightAtom` so `EditorPanelContent` can scroll to and softly + * highlight it inside the rendered markdown. + * + * Surfsense docs chunks: rendered as a hover-controlled shadcn Popover that + * lazily fetches and previews the cited chunk inline, since those docs aren't + * indexed into the user's search space and have no tab to open. */ export const InlineCitation: FC = ({ chunkId, isDocsChunk = false }) => { - const [isOpen, setIsOpen] = useState(false); - if (chunkId < 0) { return ( @@ -38,26 +57,185 @@ export const InlineCitation: FC = ({ chunkId, isDocsChunk = ); } + if (isDocsChunk) { + return ; + } + + return ; +}; + +const NumericChunkCitation: FC<{ chunkId: number }> = ({ chunkId }) => { + const queryClient = useQueryClient(); + const setPendingHighlight = useSetAtom(pendingChunkHighlightAtom); + const openEditorPanel = useSetAtom(openEditorPanelAtom); + const [resolving, setResolving] = useState(false); + + const handleClick = useCallback(async () => { + if (resolving) return; + setResolving(true); + console.log("[citation:click] start", { chunkId }); + try { + const data = await queryClient.fetchQuery({ + // Local key with explicit window. The shared `cacheKeys.documents.byChunk` + // is window-agnostic (latent footgun); namespace the call to avoid + // reusing a different-window cached result. + queryKey: ["documents", "by-chunk", chunkId, "w0"] as const, + queryFn: () => + documentsApiService.getDocumentByChunk({ chunk_id: chunkId, chunk_window: 0 }), + staleTime: 5 * 60 * 1000, + }); + const cited = data.chunks.find((c) => c.id === chunkId) ?? data.chunks[0]; + console.log("[citation:click] fetched doc-by-chunk", { + docId: data.id, + docTitle: data.title, + chunksReturned: data.chunks.length, + citedChunkId: cited?.id, + citedChunkContentLen: cited?.content?.length ?? 0, + citedChunkPreview: + cited?.content && cited.content.length > 120 + ? `${cited.content.slice(0, 120)}…(+${cited.content.length - 120})` + : (cited?.content ?? ""), + }); + // Stage the highlight BEFORE opening the panel so `EditorPanelContent` + // already sees the pending intent on its very first render — avoids a + // "fetch → render → no-pending → next-tick render with pending" race. + setPendingHighlight({ + documentId: data.id, + chunkId, + chunkText: cited?.content ?? "", + }); + openEditorPanel({ + documentId: data.id, + searchSpaceId: data.search_space_id, + title: data.title, + }); + console.log("[citation:click] staged highlight + opened editor panel", { + documentId: data.id, + }); + } catch (err) { + console.warn("[citation:click] failed", err); + toast.error(err instanceof Error ? err.message : "Couldn't open cited document"); + } finally { + setResolving(false); + } + }, [chunkId, openEditorPanel, queryClient, resolving, setPendingHighlight]); + return ( - - + ); +}; + +const SurfsenseDocCitation: FC<{ chunkId: number }> = ({ chunkId }) => { + const [open, setOpen] = useState(false); + const closeTimerRef = useRef | null>(null); + + const cancelClose = useCallback(() => { + if (closeTimerRef.current) { + clearTimeout(closeTimerRef.current); + closeTimerRef.current = null; + } + }, []); + + const scheduleClose = useCallback(() => { + cancelClose(); + closeTimerRef.current = setTimeout(() => { + setOpen(false); + closeTimerRef.current = null; + }, POPOVER_HOVER_CLOSE_DELAY_MS); + }, [cancelClose]); + + useEffect(() => () => cancelClose(), [cancelClose]); + + const { data, isLoading, error } = useQuery({ + queryKey: cacheKeys.documents.byChunk(`doc-${chunkId}`), + queryFn: () => documentsApiService.getSurfsenseDocByChunk(chunkId), + enabled: open, + staleTime: 5 * 60 * 1000, + }); + + const citedChunk = data?.chunks.find((c) => c.id === chunkId) ?? data?.chunks[0]; + + return ( + + + + + e.preventDefault()} > - {chunkId} - - +
+
+

+ {data?.title ?? "Surfsense documentation"} +

+

Chunk #{chunkId}

+
+ {data?.source && ( + + + Open + + )} +
+
+ {isLoading && ( +
+ + Loading… +
+ )} + {error && ( +

+ {error instanceof Error ? error.message : "Failed to load chunk"} +

+ )} + {!isLoading && !error && citedChunk?.content && ( + + )} + {!isLoading && !error && !citedChunk?.content && ( +

No content available.

+ )} +
+ + ); }; diff --git a/surfsense_web/components/editor-panel/editor-panel.tsx b/surfsense_web/components/editor-panel/editor-panel.tsx index 3b69ae6e0..0c4e9485b 100644 --- a/surfsense_web/components/editor-panel/editor-panel.tsx +++ b/surfsense_web/components/editor-panel/editor-panel.tsx @@ -1,5 +1,6 @@ "use client"; +import { FindReplacePlugin } from "@platejs/find-replace"; import { useAtomValue, useSetAtom } from "jotai"; import { Check, @@ -14,17 +15,21 @@ import { import dynamic from "next/dynamic"; import { useCallback, useEffect, useRef, useState } from "react"; import { toast } from "sonner"; +import { pendingChunkHighlightAtom } from "@/atoms/document-viewer/pending-chunk-highlight.atom"; import { closeEditorPanelAtom, editorPanelAtom } from "@/atoms/editor/editor-panel.atom"; import { VersionHistoryButton } from "@/components/documents/version-history"; +import type { PlateEditorInstance } from "@/components/editor/plate-editor"; import { SourceCodeEditor } from "@/components/editor/source-code-editor"; import { MarkdownViewer } from "@/components/markdown-viewer"; import { Alert, AlertDescription } from "@/components/ui/alert"; import { Button } from "@/components/ui/button"; import { Drawer, DrawerContent, DrawerHandle, DrawerTitle } from "@/components/ui/drawer"; +import { CITATION_HIGHLIGHT_CLASS } from "@/components/ui/search-highlight-node"; import { Spinner } from "@/components/ui/spinner"; import { useMediaQuery } from "@/hooks/use-media-query"; import { useElectronAPI } from "@/hooks/use-platform"; import { authenticatedFetch, getBearerToken, redirectToLogin } from "@/lib/auth-utils"; +import { buildCitationSearchCandidates } from "@/lib/citation-search"; import { inferMonacoLanguageFromPath } from "@/lib/editor-language"; const PlateEditor = dynamic( @@ -32,7 +37,10 @@ const PlateEditor = dynamic( { ssr: false, loading: () => } ); +type CitationHighlightStatus = "exact" | "miss"; + const LARGE_DOCUMENT_THRESHOLD = 2 * 1024 * 1024; // 2MB +const CITATION_MAX_LENGTH = 16 * 1024 * 1024; // 16MB on-demand cap for citation jumps interface EditorContent { document_id: number; @@ -136,6 +144,61 @@ export function EditorPanelContent({ const [displayTitle, setDisplayTitle] = useState(title || "Untitled"); const isLocalFileMode = kind === "local_file"; const editorRenderMode: EditorRenderMode = isLocalFileMode ? "source_code" : "rich_markdown"; + + // --- Citation-jump highlight wiring ---------------------------------- + // `EditorPanelContent` is the consumer of `pendingChunkHighlightAtom`: when + // a citation badge is clicked, the badge stages `{documentId, chunkId, + // chunkText}` and opens this panel. We drive Plate's `FindReplacePlugin` + // (registered in every preset) to highlight the cited text natively via + // Slate decorations — no DOM walking, no Range gymnastics. The state + // machine below escalates the document fetch from 2MB → 16MB once if no + // candidate snippet matched in the preview, and surfaces miss outcomes + // via an inline alert. + const pending = useAtomValue(pendingChunkHighlightAtom); + const setPendingHighlight = useSetAtom(pendingChunkHighlightAtom); + const [fetchKey, setFetchKey] = useState(0); + const [maxLengthOverride, setMaxLengthOverride] = useState(null); + const [highlightResult, setHighlightResult] = useState(null); + const editorRef = useRef(null); + const escalatedForRef = useRef(null); + const lastAppliedChunkIdRef = useRef(null); + // Tracks whether a citation highlight is currently decorated in the + // editor. We use a ref (not state) because the click-to-dismiss handler + // runs in a stable callback that would otherwise close over stale state. + const isHighlightActiveRef = useRef(false); + // Once a citation jump targets this doc we have to keep `PlateEditor` + // mounted for the *rest of the doc session* — even after the highlight + // effect clears `pendingChunkHighlightAtom` (which it does as soon as + // the decoration is applied, so a follow-up citation on the same chunk + // can re-trigger). Without this latch, non-editable docs would re-render + // back into `MarkdownViewer` the instant `pending` is released, tearing + // down the Plate decorations and dropping the highlight after a frame. + const [stickyPlateMode, setStickyPlateMode] = useState(false); + + const clearCitationSearch = useCallback(() => { + isHighlightActiveRef.current = false; + const editor = editorRef.current; + if (!editor) return; + try { + editor.setOption(FindReplacePlugin, "search", ""); + editor.api.redecorate(); + } catch (err) { + console.warn("[EditorPanelContent] clearCitationSearch failed:", err); + } + }, []); + + // Dismiss the highlight when the user interacts with the editor surface. + // `onPointerDown` fires before focus / selection changes so the click + // itself feels responsive — the highlight clears in the same event tick + // that places the cursor. No-op when nothing is highlighted, so we don't + // thrash `redecorate` on every click in normal editing. + const handleEditorPointerDown = useCallback(() => { + if (!isHighlightActiveRef.current) return; + clearCitationSearch(); + setHighlightResult(null); + }, [clearCitationSearch]); + + const isCitationTarget = !!pending && !isLocalFileMode && pending.documentId === documentId; const resolveLocalVirtualPath = useCallback( async (candidatePath: string): Promise => { if (!electronAPI?.getAgentFilesystemMounts) { @@ -155,6 +218,8 @@ export function EditorPanelContent({ const isLargeDocument = (editorDoc?.content_size_bytes ?? 0) > LARGE_DOCUMENT_THRESHOLD; + // `fetchKey` is an explicit re-fetch trigger (escalation bumps it to force + // a new request even when documentId/searchSpaceId haven't changed). useEffect(() => { const controller = new AbortController(); setIsLoading(true); @@ -166,6 +231,12 @@ export function EditorPanelContent({ setIsEditing(false); initialLoadDone.current = false; changeCountRef.current = 0; + // Clear any in-flight FindReplacePlugin search before the editor + // re-mounts on new content (a fresh editor key is generated below + // from documentId + isEditing, so the previous editor + its + // decorations are about to be discarded anyway, but we belt-and- + // brace here for the case where only `fetchKey` changed). + clearCitationSearch(); const doFetch = async () => { try { @@ -210,7 +281,11 @@ export function EditorPanelContent({ const url = new URL( `${process.env.NEXT_PUBLIC_FASTAPI_BACKEND_URL}/api/v1/search-spaces/${searchSpaceId}/documents/${documentId}/editor-content` ); - url.searchParams.set("max_length", String(LARGE_DOCUMENT_THRESHOLD)); + url.searchParams.set("max_length", String(maxLengthOverride ?? LARGE_DOCUMENT_THRESHOLD)); + // `fetchKey` participates here so biome's noUnusedVariables sees it + // as consumed; bumping it forces a fresh request even when the URL + // is otherwise identical. + if (fetchKey > 0) url.searchParams.set("_n", String(fetchKey)); const response = await authenticatedFetch(url.toString(), { method: "GET" }); @@ -256,8 +331,259 @@ export function EditorPanelContent({ resolveLocalVirtualPath, searchSpaceId, title, + fetchKey, + maxLengthOverride, + clearCitationSearch, ]); + // Reset citation-jump bookkeeping whenever the panel switches to a different + // document (or local file). Body only writes setters — the deps are the + // real triggers we want to react to. + // biome-ignore lint/correctness/useExhaustiveDependencies: documentId/localFilePath are intentional triggers. + useEffect(() => { + clearCitationSearch(); + escalatedForRef.current = null; + lastAppliedChunkIdRef.current = null; + setHighlightResult(null); + setMaxLengthOverride(null); + setFetchKey(0); + // Drop sticky Plate mode when the panel moves to a different doc + // — the next doc starts in its preferred render mode (Plate for + // editable, MarkdownViewer for everything else) until/unless a + // citation jump targets it. + setStickyPlateMode(false); + }, [documentId, localFilePath, clearCitationSearch]); + + // Latch sticky Plate mode the first time a citation jump targets this + // doc. We keep it sticky for the remainder of this doc session so the + // highlight effect's `setPendingHighlight(null)` doesn't unmount the + // editor mid-flight (see comment on `stickyPlateMode` declaration). + useEffect(() => { + if (isCitationTarget) setStickyPlateMode(true); + }, [isCitationTarget]); + + // `isEditorReady` is what `useEffect` actually depends on — `editorRef` + // is a ref so changes don't trigger re-runs. We flip this to `true` once + // `PlateEditor` calls back with its live editor instance (its + // `usePlateEditor` value-init runs synchronously, so by the time this + // flips true the markdown is already deserialized into the Slate tree). + const [isEditorReady, setIsEditorReady] = useState(false); + const handleEditorReady = useCallback((editor: PlateEditorInstance | null) => { + console.log("[citation:editor] handleEditorReady", { ready: !!editor }); + editorRef.current = editor; + setIsEditorReady(!!editor); + }, []); + + // --- Citation jump highlight effect ----------------------------------- + // Drives Plate's FindReplacePlugin to highlight the cited chunk: + // 1. Build candidate snippets from the chunk text (first sentence, + // first 8 words, full chunk if short). Plate's decorate runs per- + // block and won't cross block boundaries, so the shorter + // candidates exist to give us something that fits in one + // paragraph / heading. + // 2. For each candidate: setOption('search', ...) → redecorate → + // wait two animation frames for React to flush → query the editor + // DOM for `.${CITATION_HIGHLIGHT_CLASS}`. First hit wins. + // + // Why a className and not a `data-*` attribute? Plate's + // `PlateLeaf` runs its props through `useNodeAttributes`, which + // only forwards `attributes`, `className`, `ref`, and `style` — + // arbitrary `data-*` attributes are silently dropped. `className` + // is the only escape hatch guaranteed to survive into the DOM. + // 3. On hit: smooth-scroll the first match into view, mark the + // highlight active (so a click inside the editor can dismiss it), + // release the pending atom. + // 4. On terminal miss: if the doc was truncated and we haven't + // escalated yet, bump the fetch's `max_length` to the citation + // cap and re-fetch — the post-refetch render will re-run this + // effect against the larger preview. Otherwise, release the + // atom and show the miss alert. + useEffect(() => { + console.log("[citation:effect] fired", { + isCitationTarget, + pendingDocId: pending?.documentId, + pendingChunkId: pending?.chunkId, + pendingChunkTextLen: pending?.chunkText?.length, + documentId, + isLocalFileMode, + isEditing, + hasMarkdown: !!editorDoc?.source_markdown, + markdownLen: editorDoc?.source_markdown?.length, + truncated: editorDoc?.truncated, + isEditorReady, + editorRefSet: !!editorRef.current, + maxLengthOverride, + }); + if (!isCitationTarget || !pending) { + console.log("[citation:effect] guard ✗ no citation target / no pending"); + return; + } + if (isLocalFileMode || isEditing) { + console.log("[citation:effect] guard ✗ localFileMode/editing"); + return; + } + if (!editorDoc?.source_markdown) { + console.log("[citation:effect] guard ✗ source_markdown not ready"); + return; + } + if (!isEditorReady) { + console.log("[citation:effect] guard ✗ editor not ready yet"); + return; + } + const editor = editorRef.current; + if (!editor) { + console.log("[citation:effect] guard ✗ editorRef.current is null"); + return; + } + + if (lastAppliedChunkIdRef.current !== pending.chunkId) { + lastAppliedChunkIdRef.current = pending.chunkId; + } + + let cancelled = false; + + const finishMiss = () => { + console.log("[citation:effect] terminal miss — no candidate matched"); + try { + editor.setOption(FindReplacePlugin, "search", ""); + editor.api.redecorate(); + } catch (err) { + console.warn("[EditorPanelContent] reset search after miss failed:", err); + } + const canEscalate = + editorDoc.truncated === true && + (maxLengthOverride ?? LARGE_DOCUMENT_THRESHOLD) < CITATION_MAX_LENGTH && + escalatedForRef.current !== pending.chunkId; + console.log("[citation:effect] miss decision", { + truncated: editorDoc.truncated, + currentMaxLength: maxLengthOverride ?? LARGE_DOCUMENT_THRESHOLD, + canEscalate, + }); + if (canEscalate) { + escalatedForRef.current = pending.chunkId; + setMaxLengthOverride(CITATION_MAX_LENGTH); + setFetchKey((k) => k + 1); + // Keep the atom set so the post-refetch render re-runs. + return; + } + setHighlightResult("miss"); + setPendingHighlight(null); + }; + + const tryCandidates = async () => { + const candidates = buildCitationSearchCandidates(pending.chunkText); + console.log("[citation:effect] candidates built", { + count: candidates.length, + previews: candidates.map((c) => c.slice(0, 60)), + }); + if (candidates.length === 0) { + if (!cancelled) finishMiss(); + return; + } + // Resolve the editor's rendered DOM root via Slate's stable + // `[data-slate-editor="true"]` attribute (set by slate-react's + // ``). Scoping queries to this root prevents + // `` elements rendered elsewhere on the page (e.g. chat + // search-highlight leaves in another mounted PlateEditor) from + // being mistaken for citation hits. + const editorRoot = document.querySelector('[data-slate-editor="true"]'); + console.log("[citation:effect] editor root", { + hasRoot: !!editorRoot, + }); + const root: ParentNode = editorRoot ?? document; + + for (let i = 0; i < candidates.length; i++) { + const candidate = candidates[i]; + if (cancelled) return; + try { + editor.setOption(FindReplacePlugin, "search", candidate); + editor.api.redecorate(); + console.log(`[citation:effect] try #${i} setOption + redecorate`, { + len: candidate.length, + preview: candidate.slice(0, 80), + }); + } catch (err) { + console.warn("[EditorPanelContent] setOption/redecorate failed:", err); + continue; + } + // Two rAFs: first lets Slate flush its onChange, second lets + // React commit the decoration leaves into the DOM. + await new Promise((resolve) => + requestAnimationFrame(() => requestAnimationFrame(() => resolve())) + ); + if (cancelled) return; + // Primary probe: by our stable class on the rendered . + let el = root.querySelector(`.${CITATION_HIGHLIGHT_CLASS}`); + const classMarkCount = root.querySelectorAll(`.${CITATION_HIGHLIGHT_CLASS}`).length; + // Diagnostic fallback: any inside the editor root. + // If we ever see allMarks > 0 but classMarkCount === 0, + // the className was stripped again and we need to revisit + // `useNodeAttributes` filtering. + const allMarkCount = root.querySelectorAll("mark").length; + if (!el && allMarkCount > 0) { + el = root.querySelector("mark"); + } + console.log(`[citation:effect] try #${i} DOM probe`, { + foundEl: !!el, + classMarkCount, + allMarkCount, + usedFallback: !!el && classMarkCount === 0, + }); + if (el) { + try { + el.scrollIntoView({ block: "center", behavior: "smooth" }); + } catch { + el.scrollIntoView(); + } + isHighlightActiveRef.current = true; + setHighlightResult("exact"); + console.log(`[citation:effect] ✓ exact via candidate #${i} — atom released`); + // No auto-clear timer — the highlight is intentionally + // permanent until the user clicks inside the editor (see + // `handleEditorPointerDown`) or another dismissal trigger + // fires (doc switch, edit-mode toggle, panel unmount, + // next citation jump). Sticky Plate mode keeps the + // editor mounted after the atom clears. + setPendingHighlight(null); + return; + } + } + if (!cancelled) finishMiss(); + }; + + void tryCandidates(); + + return () => { + cancelled = true; + }; + }, [ + isCitationTarget, + pending, + documentId, + editorDoc?.source_markdown, + editorDoc?.truncated, + isLocalFileMode, + isEditing, + isEditorReady, + maxLengthOverride, + clearCitationSearch, + setPendingHighlight, + ]); + + // Cleanup any active highlight on unmount. + useEffect(() => { + return () => clearCitationSearch(); + }, [clearCitationSearch]); + + // Toggling into edit mode swaps Plate out of readOnly. Clear the citation + // search so stale leaves don't linger in the editing surface. + useEffect(() => { + if (isEditing) { + clearCitationSearch(); + setHighlightResult(null); + } + }, [isEditing, clearCitationSearch]); + useEffect(() => { return () => { if (copyResetTimeoutRef.current) { @@ -367,6 +693,15 @@ export function EditorPanelContent({ EDITABLE_DOCUMENT_TYPES.has(editorDoc.document_type ?? "")) && !isLargeDocument : false; + // Use PlateEditor for any of: + // - Editable doc types (FILE/NOTE) — existing editing UX. + // - Active citation jump in flight (`isCitationTarget`) — covers the + // mount in the very first render where the atom is set but the + // sticky effect hasn't fired yet. + // - Sticky Plate mode latched on a previous citation jump — keeps + // the editor mounted (with its decorations) after the highlight + // effect clears the atom. Resets when the doc changes. + const renderInPlateEditor = isEditableType || isCitationTarget || stickyPlateMode; const hasUnsavedChanges = editedMarkdown !== null; const showDesktopHeader = !!onClose; const showEditingActions = isEditableType && isEditing; @@ -381,6 +716,90 @@ export function EditorPanelContent({ setIsEditing(false); }, [editorDoc?.source_markdown]); + const handleDownloadMarkdown = useCallback(async () => { + if (!searchSpaceId || !documentId) return; + setDownloading(true); + try { + const response = await authenticatedFetch( + `${process.env.NEXT_PUBLIC_FASTAPI_BACKEND_URL}/api/v1/search-spaces/${searchSpaceId}/documents/${documentId}/download-markdown`, + { method: "GET" } + ); + if (!response.ok) throw new Error("Download failed"); + const blob = await response.blob(); + const url = URL.createObjectURL(blob); + const a = document.createElement("a"); + a.href = url; + const disposition = response.headers.get("content-disposition"); + const match = disposition?.match(/filename="(.+)"/); + a.download = match?.[1] ?? `${editorDoc?.title || "document"}.md`; + document.body.appendChild(a); + a.click(); + a.remove(); + URL.revokeObjectURL(url); + toast.success("Download started"); + } catch { + toast.error("Failed to download document"); + } finally { + setDownloading(false); + } + }, [documentId, editorDoc?.title, searchSpaceId]); + + // We no longer surface an "approximate" status — Plate's FindReplacePlugin + // either decorates an exact match or it doesn't, and the candidate snippet + // strategy (first sentence → first 8 words → full chunk) means we either + // land on the citation start or fall through to the miss alert. + const showMissAlert = isCitationTarget && highlightResult === "miss"; + + const citationAlerts = showMissAlert && ( + + + + Cited section couldn't be located in this view. + {editorDoc?.truncated && ( + + )} + + + ); + + const largeDocAlert = isLargeDocument && !isLocalFileMode && editorDoc && ( + + + + + This document is too large for the editor ( + {Math.round((editorDoc.content_size_bytes ?? 0) / 1024 / 1024)}MB,{" "} + {editorDoc.chunk_count ?? 0} chunks). Showing a preview below. + + + + + ); + return ( <> {showDesktopHeader ? ( @@ -565,61 +984,6 @@ export function EditorPanelContent({

- ) : isLargeDocument && !isLocalFileMode ? ( -
- - - - - This document is too large for the editor ( - {Math.round((editorDoc.content_size_bytes ?? 0) / 1024 / 1024)}MB,{" "} - {editorDoc.chunk_count ?? 0} chunks). Showing a preview below. - - - - - -
) : editorRenderMode === "source_code" ? (
- ) : isEditableType ? ( - + ) : isLargeDocument && !isLocalFileMode && !isCitationTarget ? ( + // Large doc, no active citation — fast Streamdown preview + // + download CTA. We only fall back to MarkdownViewer here + // because Plate is heavy on multi-MB docs and the user + // isn't waiting on a specific citation to render. +
+ {largeDocAlert} + +
+ ) : renderInPlateEditor ? ( + // Editable doc (FILE/NOTE) OR active citation jump (any + // doc type). The citation path uses Plate's + // FindReplacePlugin for native, decoration-based + // highlighting — see the citation-jump highlight effect + // above for how `editorRef` and `handleEditorReady` are + // wired. +
+ {(citationAlerts || (isLargeDocument && isCitationTarget && !isLocalFileMode)) && ( +
+ {isLargeDocument && isCitationTarget && largeDocAlert} + {citationAlerts} +
+ )} +
+ +
+
) : (
diff --git a/surfsense_web/components/editor/plate-editor.tsx b/surfsense_web/components/editor/plate-editor.tsx index 481a420fb..eef18ef6a 100644 --- a/surfsense_web/components/editor/plate-editor.tsx +++ b/surfsense_web/components/editor/plate-editor.tsx @@ -12,6 +12,12 @@ import { type EditorPreset, presetMap } from "@/components/editor/presets"; import { escapeMdxExpressions } from "@/components/editor/utils/escape-mdx"; import { Editor, EditorContainer } from "@/components/ui/editor"; +/** Live editor instance returned by `usePlateEditor`. Exposed via the + * `onEditorReady` prop so callers (e.g. `EditorPanelContent`) can drive + * plugin options imperatively — most notably setting + * `FindReplacePlugin`'s `search` option for citation-jump highlights. */ +export type PlateEditorInstance = ReturnType; + export interface PlateEditorProps { /** Markdown string to load as initial content */ markdown?: string; @@ -62,6 +68,15 @@ export interface PlateEditorProps { * without modifying the core editor component. */ extraPlugins?: AnyPluginConfig[]; + /** + * Called whenever the live editor instance (re)mounts, with `null` on + * unmount. Used by callers that need to drive plugin options imperatively + * — e.g. `EditorPanelContent` setting `FindReplacePlugin`'s `search` + * option for citation-jump highlights. The callback is invoked exactly + * once per editor lifetime (the parent's `key` prop forces a fresh + * editor when needed, e.g. on edit-mode toggle). + */ + onEditorReady?: (editor: PlateEditorInstance | null) => void; } function PlateEditorContent({ @@ -100,6 +115,7 @@ export function PlateEditor({ defaultEditing = false, preset = "full", extraPlugins = [], + onEditorReady, }: PlateEditorProps) { const lastMarkdownRef = useRef(markdown); const lastHtmlRef = useRef(html); @@ -156,6 +172,21 @@ export function PlateEditor({ : undefined, }); + // Expose the live editor instance to imperative callers (e.g. citation + // jump highlights). We deliberately don't depend on `onEditorReady` + // itself in the cleanup closure — callers commonly pass an arrow that + // closes over a stable ref setter, but if they pass a freshly-bound + // callback per render, the `onEditorReady?.(editor)` re-fires which is + // idempotent for ref-style setters. + const onEditorReadyRef = useRef(onEditorReady); + useEffect(() => { + onEditorReadyRef.current = onEditorReady; + }, [onEditorReady]); + useEffect(() => { + onEditorReadyRef.current?.(editor); + return () => onEditorReadyRef.current?.(null); + }, [editor]); + // Update editor content when html prop changes externally useEffect(() => { if (html !== undefined && html !== lastHtmlRef.current) { diff --git a/surfsense_web/components/editor/presets.ts b/surfsense_web/components/editor/presets.ts index c207b5e56..49f53ecf1 100644 --- a/surfsense_web/components/editor/presets.ts +++ b/surfsense_web/components/editor/presets.ts @@ -1,5 +1,6 @@ "use client"; +import { FindReplacePlugin } from "@platejs/find-replace"; import type { AnyPluginConfig } from "platejs"; import { TrailingBlockPlugin } from "platejs"; @@ -17,6 +18,30 @@ import { SelectionKit } from "@/components/editor/plugins/selection-kit"; import { SlashCommandKit } from "@/components/editor/plugins/slash-command-kit"; import { TableKit } from "@/components/editor/plugins/table-kit"; import { ToggleKit } from "@/components/editor/plugins/toggle-kit"; +import { SearchHighlightLeaf } from "@/components/ui/search-highlight-node"; + +/** + * Citation-jump highlighter. Re-uses Plate's built-in `FindReplacePlugin` + * (decorate-only, no editing surface) to drive the "scroll-to-cited-text" + * UX in `EditorPanelContent`. We register it in every preset because: + * - Decorate is a no-op when `search` is empty (single getOptions() check + * per block), so cost is effectively zero for non-citation viewers. + * - Keeping it preset-agnostic means citations work whether the doc is + * opened in editable (`full`) or pure-viewer (`readonly`) modes. + * + * The parent component drives `setOption(FindReplacePlugin, 'search', ...)` + * + `editor.api.redecorate()` to trigger highlights, then queries the + * editor DOM for `.citation-highlight-leaf` to scroll the first match + * into view. (We can't use a `data-*` attribute here — Plate's + * `PlateLeaf` runs props through `useNodeAttributes`, which only forwards + * `attributes`, `className`, `ref`, `style`; arbitrary `data-*` props are + * silently dropped.) See `components/ui/search-highlight-node.tsx` for + * the leaf component and `CITATION_HIGHLIGHT_CLASS` constant. + */ +const CitationFindReplacePlugin = FindReplacePlugin.configure({ + options: { search: "" }, + render: { node: SearchHighlightLeaf }, +}); /** * Full preset – every plugin kit enabled. @@ -38,6 +63,7 @@ export const fullPreset: AnyPluginConfig[] = [ ...AutoformatKit, ...DndKit, TrailingBlockPlugin, + CitationFindReplacePlugin, ]; /** @@ -52,6 +78,7 @@ export const minimalPreset: AnyPluginConfig[] = [ ...LinkKit, ...AutoformatKit, TrailingBlockPlugin, + CitationFindReplacePlugin, ]; /** @@ -68,6 +95,7 @@ export const readonlyPreset: AnyPluginConfig[] = [ ...CalloutKit, ...ToggleKit, ...MathKit, + CitationFindReplacePlugin, ]; /** All available preset names */ diff --git a/surfsense_web/components/new-chat/source-detail-panel.tsx b/surfsense_web/components/new-chat/source-detail-panel.tsx deleted file mode 100644 index aded206c7..000000000 --- a/surfsense_web/components/new-chat/source-detail-panel.tsx +++ /dev/null @@ -1,719 +0,0 @@ -"use client"; - -import { useQuery } from "@tanstack/react-query"; -import { - BookOpen, - ChevronDown, - ChevronUp, - ExternalLink, - FileQuestionMark, - FileText, - Hash, - Loader2, - Sparkles, - X, -} from "lucide-react"; -import { AnimatePresence, motion, useReducedMotion } from "motion/react"; -import { useTranslations } from "next-intl"; -import type React from "react"; -import { forwardRef, memo, type ReactNode, useCallback, useEffect, useRef, useState } from "react"; -import { createPortal } from "react-dom"; -import { MarkdownViewer } from "@/components/markdown-viewer"; -import { Badge } from "@/components/ui/badge"; -import { Button } from "@/components/ui/button"; -import { ScrollArea } from "@/components/ui/scroll-area"; -import { Spinner } from "@/components/ui/spinner"; -import type { - GetDocumentByChunkResponse, - GetSurfsenseDocsByChunkResponse, -} from "@/contracts/types/document.types"; -import { documentsApiService } from "@/lib/apis/documents-api.service"; -import { cacheKeys } from "@/lib/query-client/cache-keys"; -import { cn } from "@/lib/utils"; - -type DocumentData = GetDocumentByChunkResponse | GetSurfsenseDocsByChunkResponse; - -interface SourceDetailPanelProps { - open: boolean; - onOpenChange: (open: boolean) => void; - chunkId: number; - sourceType: string; - title: string; - description?: string; - url?: string; - children?: ReactNode; - isDocsChunk?: boolean; -} - -const formatDocumentType = (type: string) => { - if (!type) return ""; - return type - .split("_") - .map((word) => word.charAt(0) + word.slice(1).toLowerCase()) - .join(" "); -}; - -// Chunk card component -// For large documents (>30 chunks), we disable animation to prevent layout shifts -// which break auto-scroll functionality -interface ChunkCardProps { - chunk: { id: number; content: string }; - localIndex: number; - chunkNumber: number; - totalChunks: number; - isCited: boolean; - isActive: boolean; - disableLayoutAnimation?: boolean; -} - -const ChunkCard = memo( - forwardRef( - ({ chunk, localIndex, chunkNumber, totalChunks, isCited }, ref) => { - return ( -
- {isCited &&
} - -
-
-
- {chunkNumber} -
- - Chunk {chunkNumber} of {totalChunks} - -
- {isCited && ( - - - Cited Source - - )} -
- -
- -
-
- ); - } - ) -); -ChunkCard.displayName = "ChunkCard"; - -export function SourceDetailPanel({ - open, - onOpenChange, - chunkId, - sourceType, - title, - description, - url, - children, - isDocsChunk = false, -}: SourceDetailPanelProps) { - const t = useTranslations("dashboard"); - const scrollAreaRef = useRef(null); - const hasScrolledRef = useRef(false); // Use ref to avoid stale closures - const scrollTimersRef = useRef[]>([]); - const [activeChunkIndex, setActiveChunkIndex] = useState(null); - const [mounted, setMounted] = useState(false); - const shouldReduceMotion = useReducedMotion(); - - useEffect(() => { - setMounted(true); - }, []); - - const { - data: documentData, - isLoading: isDocumentByChunkFetching, - error: documentByChunkFetchingError, - } = useQuery({ - queryKey: isDocsChunk - ? cacheKeys.documents.byChunk(`doc-${chunkId}`) - : cacheKeys.documents.byChunk(chunkId.toString()), - queryFn: async () => { - if (isDocsChunk) { - return documentsApiService.getSurfsenseDocByChunk(chunkId); - } - return documentsApiService.getDocumentByChunk({ chunk_id: chunkId, chunk_window: 5 }); - }, - enabled: !!chunkId && open, - staleTime: 5 * 60 * 1000, - }); - - const totalChunks = - documentData && "total_chunks" in documentData - ? (documentData.total_chunks ?? documentData.chunks.length) - : (documentData?.chunks?.length ?? 0); - const [beforeChunks, setBeforeChunks] = useState< - Array<{ id: number; content: string; created_at: string }> - >([]); - const [afterChunks, setAfterChunks] = useState< - Array<{ id: number; content: string; created_at: string }> - >([]); - const [loadingBefore, setLoadingBefore] = useState(false); - const [loadingAfter, setLoadingAfter] = useState(false); - - useEffect(() => { - setBeforeChunks([]); - setAfterChunks([]); - }, [chunkId, open]); - - const chunkStartIndex = - documentData && "chunk_start_index" in documentData ? (documentData.chunk_start_index ?? 0) : 0; - const initialChunks = documentData?.chunks ?? []; - const allChunks = [...beforeChunks, ...initialChunks, ...afterChunks]; - const absoluteStart = chunkStartIndex - beforeChunks.length; - const absoluteEnd = chunkStartIndex + initialChunks.length + afterChunks.length; - const canLoadBefore = absoluteStart > 0; - const canLoadAfter = absoluteEnd < totalChunks; - - const EXPAND_SIZE = 10; - - const loadBefore = useCallback(async () => { - if (!documentData || !("search_space_id" in documentData) || !canLoadBefore) return; - setLoadingBefore(true); - try { - const count = Math.min(EXPAND_SIZE, absoluteStart); - const result = await documentsApiService.getDocumentChunks({ - document_id: documentData.id, - page: 0, - page_size: count, - start_offset: absoluteStart - count, - }); - const existingIds = new Set(allChunks.map((c) => c.id)); - const newChunks = result.items - .filter((c) => !existingIds.has(c.id)) - .map((c) => ({ id: c.id, content: c.content, created_at: c.created_at })); - setBeforeChunks((prev) => [...newChunks, ...prev]); - } catch (err) { - console.error("Failed to load earlier chunks:", err); - } finally { - setLoadingBefore(false); - } - }, [documentData, absoluteStart, canLoadBefore, allChunks]); - - const loadAfter = useCallback(async () => { - if (!documentData || !("search_space_id" in documentData) || !canLoadAfter) return; - setLoadingAfter(true); - try { - const result = await documentsApiService.getDocumentChunks({ - document_id: documentData.id, - page: 0, - page_size: EXPAND_SIZE, - start_offset: absoluteEnd, - }); - const existingIds = new Set(allChunks.map((c) => c.id)); - const newChunks = result.items - .filter((c) => !existingIds.has(c.id)) - .map((c) => ({ id: c.id, content: c.content, created_at: c.created_at })); - setAfterChunks((prev) => [...prev, ...newChunks]); - } catch (err) { - console.error("Failed to load later chunks:", err); - } finally { - setLoadingAfter(false); - } - }, [documentData, absoluteEnd, canLoadAfter, allChunks]); - - const isDirectRenderSource = - sourceType === "TAVILY_API" || - sourceType === "LINKUP_API" || - sourceType === "SEARXNG_API" || - sourceType === "BAIDU_SEARCH_API"; - - const citedChunkIndex = allChunks.findIndex((chunk) => chunk.id === chunkId); - - // Simple scroll function that scrolls to a chunk by index - const scrollToChunkByIndex = useCallback( - (chunkIndex: number, smooth = true) => { - const scrollContainer = scrollAreaRef.current; - if (!scrollContainer) return; - - const viewport = scrollContainer.querySelector( - "[data-radix-scroll-area-viewport]" - ) as HTMLElement | null; - if (!viewport) return; - - const chunkElement = scrollContainer.querySelector( - `[data-chunk-index="${chunkIndex}"]` - ) as HTMLElement | null; - if (!chunkElement) return; - - // Get positions using getBoundingClientRect for accuracy - const viewportRect = viewport.getBoundingClientRect(); - const chunkRect = chunkElement.getBoundingClientRect(); - - // Calculate where to scroll to center the chunk - const currentScrollTop = viewport.scrollTop; - const chunkTopRelativeToViewport = chunkRect.top - viewportRect.top + currentScrollTop; - const scrollTarget = - chunkTopRelativeToViewport - viewportRect.height / 2 + chunkRect.height / 2; - - viewport.scrollTo({ - top: Math.max(0, scrollTarget), - behavior: smooth && !shouldReduceMotion ? "smooth" : "auto", - }); - - setActiveChunkIndex(chunkIndex); - }, - [shouldReduceMotion] - ); - - // Callback ref for the cited chunk - scrolls when the element mounts - const citedChunkRefCallback = useCallback( - (node: HTMLDivElement | null) => { - if (node && !hasScrolledRef.current && open) { - hasScrolledRef.current = true; // Mark immediately to prevent duplicate scrolls - - // Store the node reference for the delayed scroll - const scrollToCitedChunk = () => { - const scrollContainer = scrollAreaRef.current; - if (!scrollContainer || !node.isConnected) return false; - - const viewport = scrollContainer.querySelector( - "[data-radix-scroll-area-viewport]" - ) as HTMLElement | null; - if (!viewport) return false; - - // Get positions - const viewportRect = viewport.getBoundingClientRect(); - const chunkRect = node.getBoundingClientRect(); - - // Calculate scroll position to center the chunk - const currentScrollTop = viewport.scrollTop; - const chunkTopRelativeToViewport = chunkRect.top - viewportRect.top + currentScrollTop; - const scrollTarget = - chunkTopRelativeToViewport - viewportRect.height / 2 + chunkRect.height / 2; - - viewport.scrollTo({ - top: Math.max(0, scrollTarget), - behavior: "auto", // Instant scroll for initial positioning - }); - - return true; - }; - - // Scroll multiple times with delays to handle progressive content rendering - // Each subsequent scroll will correct for any layout shifts - const scrollAttempts = [50, 150, 300, 600, 1000]; - - scrollAttempts.forEach((delay) => { - scrollTimersRef.current.push( - setTimeout(() => { - scrollToCitedChunk(); - }, delay) - ); - }); - - // After final attempt, mark the cited chunk as active - scrollTimersRef.current.push( - setTimeout( - () => { - setActiveChunkIndex(citedChunkIndex); - }, - scrollAttempts[scrollAttempts.length - 1] + 50 - ) - ); - } - }, - [open, citedChunkIndex] - ); - - // Reset scroll state when panel closes - useEffect(() => { - if (!open) { - scrollTimersRef.current.forEach(clearTimeout); - scrollTimersRef.current = []; - hasScrolledRef.current = false; - setActiveChunkIndex(null); - } - return () => { - scrollTimersRef.current.forEach(clearTimeout); - scrollTimersRef.current = []; - }; - }, [open]); - - // Handle escape key - useEffect(() => { - const handleEscape = (e: KeyboardEvent) => { - if (e.key === "Escape" && open) { - onOpenChange(false); - } - }; - window.addEventListener("keydown", handleEscape); - return () => window.removeEventListener("keydown", handleEscape); - }, [open, onOpenChange]); - - // Prevent body scroll when open - useEffect(() => { - if (open) { - document.body.style.overflow = "hidden"; - } else { - document.body.style.overflow = ""; - } - return () => { - document.body.style.overflow = ""; - }; - }, [open]); - - const handleUrlClick = (e: React.MouseEvent, clickUrl: string) => { - e.preventDefault(); - e.stopPropagation(); - window.open(clickUrl, "_blank", "noopener,noreferrer"); - }; - - const scrollToChunk = useCallback( - (index: number) => { - scrollToChunkByIndex(index, true); - }, - [scrollToChunkByIndex] - ); - - const panelContent = ( - - {open && ( - <> - {/* Backdrop */} - onOpenChange(false)} - /> - - {/* Panel */} - - {/* Header */} - -
-

- {documentData?.title || title || "Source Document"} -

-

- {documentData && "document_type" in documentData - ? formatDocumentType(documentData.document_type) - : sourceType && formatDocumentType(sourceType)} - {totalChunks > 0 && ( - - • {totalChunks} chunk{totalChunks !== 1 ? "s" : ""} - {allChunks.length < totalChunks && ` (showing ${allChunks.length})`} - - )} -

-
-
- {url && ( - - )} - -
- - - {/* Loading State */} - {!isDirectRenderSource && isDocumentByChunkFetching && ( -
- - -

- {t("loading_document")} -

- -
- )} - - {/* Error State */} - {!isDirectRenderSource && documentByChunkFetchingError && ( -
- -
- -
-
-

Document unavailable

-

- {documentByChunkFetchingError.message || - "An unexpected error occurred. Please try again."} -

-
- - -
- )} - - {/* Direct render for web search providers */} - {isDirectRenderSource && ( - -
- {url && ( - - )} - -

- - Source Information -

-
- {title || "Untitled"} -
-
- {description || "No content available"} -
- -
- - )} - - {/* API-fetched document content */} - {!isDirectRenderSource && documentData && ( -
- {/* Chunk Navigation Sidebar */} - {allChunks.length > 1 && ( - - -
- {allChunks.map((chunk, idx) => { - const absNum = absoluteStart + idx + 1; - const isCited = chunk.id === chunkId; - const isActive = activeChunkIndex === idx; - return ( - scrollToChunk(idx)} - initial={{ opacity: 0, scale: 0.8 }} - animate={{ opacity: 1, scale: 1 }} - transition={{ delay: Math.min(idx * 0.02, 0.2) }} - className={cn( - "relative w-11 h-9 mx-auto rounded-lg text-xs font-semibold transition-all duration-200 flex items-center justify-center", - isCited - ? "bg-primary text-primary-foreground shadow-md" - : isActive - ? "bg-muted text-foreground" - : "bg-muted/50 text-muted-foreground hover:bg-muted hover:text-foreground" - )} - title={isCited ? `Chunk ${absNum} (Cited)` : `Chunk ${absNum}`} - > - {absNum} - {isCited && ( - - - - )} - - ); - })} -
- - - )} - - {/* Main Content */} - -
- {/* Document Metadata */} - {"document_metadata" in documentData && - documentData.document_metadata && - Object.keys(documentData.document_metadata).length > 0 && ( - -

- - Document Information -

-
- {Object.entries(documentData.document_metadata).map(([key, value]) => ( -
-
- {key.replace(/_/g, " ")} -
-
{String(value)}
-
- ))} -
- - )} - - {/* Chunks Header */} -
-

- - Chunks {absoluteStart + 1}–{absoluteEnd} of {totalChunks} -

- {citedChunkIndex !== -1 && ( - - )} -
- - {/* Load Earlier */} - {canLoadBefore && ( -
- -
- )} - - {/* Chunks */} -
- {allChunks.map((chunk, idx) => { - const isCited = chunk.id === chunkId; - const chunkNumber = absoluteStart + idx + 1; - return ( - 30} - /> - ); - })} -
- - {/* Load Later */} - {canLoadAfter && ( -
- -
- )} -
- -
- )} - - - )} - - ); - - if (!mounted) return <>{children}; - - return ( - <> - {children} - {createPortal(panelContent, globalThis.document.body)} - - ); -} diff --git a/surfsense_web/components/settings/user-settings-dialog.tsx b/surfsense_web/components/settings/user-settings-dialog.tsx index 7352a82ee..a04ce16dd 100644 --- a/surfsense_web/components/settings/user-settings-dialog.tsx +++ b/surfsense_web/components/settings/user-settings-dialog.tsx @@ -67,9 +67,6 @@ const DesktopShortcutsContent = dynamic( import( "@/app/dashboard/[search_space_id]/user-settings/components/DesktopShortcutsContent" ).then((m) => ({ default: m.DesktopShortcutsContent })), - import( - "@/app/dashboard/[search_space_id]/user-settings/components/DesktopShortcutsContent" - ).then((m) => ({ default: m.DesktopShortcutsContent })), { ssr: false } ); const MemoryContent = dynamic( diff --git a/surfsense_web/components/ui/search-highlight-node.tsx b/surfsense_web/components/ui/search-highlight-node.tsx new file mode 100644 index 000000000..e3f316cce --- /dev/null +++ b/surfsense_web/components/ui/search-highlight-node.tsx @@ -0,0 +1,45 @@ +"use client"; + +import type { PlateLeafProps } from "platejs/react"; +import { PlateLeaf } from "platejs/react"; + +/** + * Stable class name used to identify Plate-rendered citation highlight + * leaves in the DOM. We can't use a `data-*` attribute here — Plate's + * `PlateLeaf` runs its props through `useNodeAttributes`, which only + * forwards `attributes`, `className`, `ref`, and `style` to the rendered + * element; arbitrary `data-*` props are silently dropped (verified + * against `@platejs/core/dist/react/index.js` v52). So `className` is + * the only escape hatch that's guaranteed to survive into the DOM. + */ +export const CITATION_HIGHLIGHT_CLASS = "citation-highlight-leaf"; + +/** + * Leaf rendered for ranges decorated by `@platejs/find-replace`'s + * `FindReplacePlugin`. We re-purpose that plugin to drive the citation-jump + * highlight: when a citation is staged, the parent sets the plugin's `search` + * option to a snippet of the chunk text and Plate decorates every match with + * `searchHighlight: true`. This component renders those decorations as a + * `` tagged with `CITATION_HIGHLIGHT_CLASS` so the parent can: + * 1. Query the first match in DOM order to scroll it into view. + * 2. Detect the active-highlight state without a separate React ref. + * + * The highlight is **persistent** — it does not auto-fade. The parent in + * `EditorPanelContent` clears it by setting the plugin's `search` option + * back to "" when one of: (a) the user clicks anywhere inside the editor, + * (b) the panel switches to a different document, (c) the user toggles + * into edit mode, (d) another citation jump is staged, (e) the panel + * unmounts. We use a brief entrance pulse (`citation-flash-in`, see + * `globals.css`) purely to draw the eye after `scrollIntoView` lands. + */ +export function SearchHighlightLeaf(props: PlateLeafProps) { + return ( + + {props.children} + + ); +} diff --git a/surfsense_web/lib/citation-search.ts b/surfsense_web/lib/citation-search.ts new file mode 100644 index 000000000..f80f13076 --- /dev/null +++ b/surfsense_web/lib/citation-search.ts @@ -0,0 +1,125 @@ +/** + * Snippet generation for the citation-jump highlight, driven by Plate's + * `FindReplacePlugin`. The plugin runs `decorate` per-block and only matches + * within blocks whose children are all `Text` nodes (so it crosses inline + * marks like bold/italic but **not** block boundaries, and a block that + * contains even one inline element such as a link is silently skipped). + * That means a full chunk that spans heading + paragraph won't match as a + * single string — we have to pick a shorter snippet that fits inside one + * rendered block. + * + * `buildCitationSearchCandidates` returns search strings ordered from + * "most-specific anchor" to "broadest fallback": + * 1. First sentence of the chunk (capped at `FIRST_SENTENCE_MAX`). + * 2. First `FIRST_PHRASE_WORDS` words. + * 3. Each non-trivial line of the chunk, in source order — gives us a + * separate attempt for each rendered block, so a heading line with + * an inline link doesn't doom the whole jump. + * 4. Full chunk (only if it's already short enough to plausibly fit + * inside one block). + * + * The caller tries each candidate in turn — set the plugin's `search` + * option, `editor.api.redecorate()`, then check the editor DOM for a + * `.citation-highlight-leaf` element. First candidate that produces one + * wins; subsequent candidates are skipped. + */ + +const FIRST_SENTENCE_MAX = 120; +const FIRST_PHRASE_WORDS = 8; +const MIN_SNIPPET_LENGTH = 6; +const FULL_CHUNK_MAX = FIRST_SENTENCE_MAX * 2; +const MAX_LINE_CANDIDATES = 6; +const LINE_CANDIDATE_MAX = FIRST_SENTENCE_MAX; + +function normalizeWhitespace(input: string): string { + return input.replace(/\s+/g, " ").trim(); +} + +/** + * Strip the markdown syntax that won't survive into the rendered editor's + * plain text, so the chunk text (which comes back from the indexer as raw + * source markdown) can be matched against the literal text values stored + * in Plate's Slate tree. + * + * Order matters: handle multi-char and "container" syntax before single- + * char emphasis, otherwise `**text**` collapses to `*text*` first. + * + * Heuristic only — we don't aim to be a full markdown parser, just to + * remove the common markers (`**bold**`, `[text](url)`, `# headings`, + * `- list`, etc.) that show up in connector-doc chunks and would break + * literal substring search. + */ +export function stripMarkdownForMatch(input: string): string { + let s = input; + s = s.replace(/```[a-z0-9_+-]*\n?([\s\S]*?)```/gi, (_, body: string) => body); + s = s.replace(//g, " "); + s = s.replace(/!\[([^\]]*)\]\([^)]*\)/g, "$1"); + s = s.replace(/!\[([^\]]*)\]\[[^\]]*\]/g, "$1"); + s = s.replace(/\[([^\]]+)\]\([^)]*\)/g, "$1"); + s = s.replace(/\[([^\]]+)\]\[[^\]]*\]/g, "$1"); + s = s.replace(/<((?:https?|mailto):[^>\s]+)>/g, "$1"); + s = s.replace(/`+([^`\n]+?)`+/g, "$1"); + s = s.replace(/(\*\*|__)([\s\S]+?)\1/g, "$2"); + s = s.replace(/(?+[ \t]?/gm, ""); + s = s.replace(/^[ \t]*[-*+][ \t]+/gm, ""); + s = s.replace(/^[ \t]*\d+\.[ \t]+/gm, ""); + s = s.replace(/^[ \t]{0,3}(?:[-*_])(?:[ \t]*[-*_]){2,}[ \t]*$/gm, ""); + s = s.replace(/^[ \t]*\|?(?:[ \t]*:?-+:?[ \t]*\|)+[ \t]*:?-+:?[ \t]*\|?[ \t]*$/gm, ""); + s = s.replace(/\\([\\`*_{}[\]()#+\-.!~>])/g, "$1"); + return s; +} + +export function buildCitationSearchCandidates(rawText: string): string[] { + if (!rawText) return []; + const stripped = stripMarkdownForMatch(rawText); + const normalized = normalizeWhitespace(stripped); + if (normalized.length < MIN_SNIPPET_LENGTH) return []; + + const out: string[] = []; + const seen = new Set(); + const push = (s: string) => { + const t = normalizeWhitespace(s); + if (t.length >= MIN_SNIPPET_LENGTH && !seen.has(t)) { + out.push(t); + seen.add(t); + } + }; + + const sentenceMatch = normalized.match(/^[^.!?]+[.!?]/); + if (sentenceMatch) { + const sentence = sentenceMatch[0]; + push(sentence.length > FIRST_SENTENCE_MAX ? sentence.slice(0, FIRST_SENTENCE_MAX) : sentence); + } else if (normalized.length > FIRST_SENTENCE_MAX) { + push(normalized.slice(0, FIRST_SENTENCE_MAX)); + } + + const words = normalized.split(" ").filter(Boolean); + if (words.length > FIRST_PHRASE_WORDS) { + push(words.slice(0, FIRST_PHRASE_WORDS).join(" ")); + } + + // Per-line candidates: each chunk line is roughly one block in the + // rendered editor. Trying them in order gives us a separate decorate + // attempt for each block, which matters when the first line is a + // heading containing a link (Plate's `FindReplacePlugin` will skip + // any block whose children aren't all text nodes). + const rawLines = stripped.split(/\r?\n/); + let lineCount = 0; + for (const line of rawLines) { + if (lineCount >= MAX_LINE_CANDIDATES) break; + const trimmed = normalizeWhitespace(line); + if (trimmed.length < MIN_SNIPPET_LENGTH) continue; + push(trimmed.length > LINE_CANDIDATE_MAX ? trimmed.slice(0, LINE_CANDIDATE_MAX) : trimmed); + lineCount++; + } + + if (normalized.length <= FULL_CHUNK_MAX) { + push(normalized); + } + + return out; +} diff --git a/surfsense_web/package.json b/surfsense_web/package.json index 41175daeb..665490e4f 100644 --- a/surfsense_web/package.json +++ b/surfsense_web/package.json @@ -36,6 +36,7 @@ "@platejs/code-block": "^52.0.11", "@platejs/combobox": "^52.0.15", "@platejs/dnd": "^52.0.11", + "@platejs/find-replace": "^52.3.10", "@platejs/floating": "^52.0.11", "@platejs/indent": "^52.0.11", "@platejs/link": "^52.0.11", diff --git a/surfsense_web/pnpm-lock.yaml b/surfsense_web/pnpm-lock.yaml index b1730e842..a1a7bea12 100644 --- a/surfsense_web/pnpm-lock.yaml +++ b/surfsense_web/pnpm-lock.yaml @@ -53,6 +53,9 @@ importers: '@platejs/dnd': specifier: ^52.0.11 version: 52.0.11(platejs@52.0.17(@types/react@19.2.14)(immer@10.2.0)(react-dom@19.2.4(react@19.2.4))(react@19.2.4)(scheduler@0.27.0)(use-sync-external-store@1.6.0(react@19.2.4)))(react-dnd-html5-backend@16.0.1)(react-dnd@16.0.1(@types/node@20.19.33)(@types/react@19.2.14)(react@19.2.4))(react-dom@19.2.4(react@19.2.4))(react@19.2.4) + '@platejs/find-replace': + specifier: ^52.3.10 + version: 52.3.10(platejs@52.0.17(@types/react@19.2.14)(immer@10.2.0)(react-dom@19.2.4(react@19.2.4))(react@19.2.4)(scheduler@0.27.0)(use-sync-external-store@1.6.0(react@19.2.4)))(react-dom@19.2.4(react@19.2.4))(react@19.2.4) '@platejs/floating': specifier: ^52.0.11 version: 52.0.11(platejs@52.0.17(@types/react@19.2.14)(immer@10.2.0)(react-dom@19.2.4(react@19.2.4))(react@19.2.4)(scheduler@0.27.0)(use-sync-external-store@1.6.0(react@19.2.4)))(react-dom@19.2.4(react@19.2.4))(react@19.2.4) @@ -2827,6 +2830,13 @@ packages: react-dnd-html5-backend: '>=14.0.0' react-dom: '>=18.0.0' + '@platejs/find-replace@52.3.10': + resolution: {integrity: sha512-V/MOMMUYxHfEn/skd2+YO213xSATFDVsl8FzVzVRV/XaxwwVefH2EPD1lAVIvmYjennTVTTsHHtEI9K9iOsEaA==} + peerDependencies: + platejs: '>=52.0.11' + react: '>=18.0.0' + react-dom: '>=18.0.0' + '@platejs/floating@52.0.11': resolution: {integrity: sha512-ApNpw4KWml+kuK+XTTpji+f/7GxTR4nRzlnfJMvGBrJpLPQ4elS5MABm3oUi81DZn+aub5HvsyH7UqCw7F76IA==} peerDependencies: @@ -11105,6 +11115,13 @@ snapshots: react-dnd-html5-backend: 16.0.1 react-dom: 19.2.4(react@19.2.4) + '@platejs/find-replace@52.3.10(platejs@52.0.17(@types/react@19.2.14)(immer@10.2.0)(react-dom@19.2.4(react@19.2.4))(react@19.2.4)(scheduler@0.27.0)(use-sync-external-store@1.6.0(react@19.2.4)))(react-dom@19.2.4(react@19.2.4))(react@19.2.4)': + dependencies: + platejs: 52.0.17(@types/react@19.2.14)(immer@10.2.0)(react-dom@19.2.4(react@19.2.4))(react@19.2.4)(scheduler@0.27.0)(use-sync-external-store@1.6.0(react@19.2.4)) + react: 19.2.4 + react-compiler-runtime: 1.0.0(react@19.2.4) + react-dom: 19.2.4(react@19.2.4) + '@platejs/floating@52.0.11(platejs@52.0.17(@types/react@19.2.14)(immer@10.2.0)(react-dom@19.2.4(react@19.2.4))(react@19.2.4)(scheduler@0.27.0)(use-sync-external-store@1.6.0(react@19.2.4)))(react-dom@19.2.4(react@19.2.4))(react@19.2.4)': dependencies: '@floating-ui/core': 1.7.4 From 4845b96209834badced8e201e45a4c89acc12ca4 Mon Sep 17 00:00:00 2001 From: yeranyang Date: Tue, 28 Apr 2026 12:16:27 +0800 Subject: [PATCH 12/17] perf(blog): derive search results with useMemo instead of useState+useEffect Fixes #1246 Replace the useState/useEffect pattern that synced fuzzy search results into local state on every search or searcher change with a single useMemo that derives results directly during render. Before: const [results, setResults] = useState(allBlogs); useEffect(() => { setResults(searcher.search(search)); }, [search, searcher]); After: const gridItems = useMemo(() => { const results = search.trim() ? searcher.search(search) : allBlogs; ... }, [search, searcher, allBlogs, featuredSlug]); This removes an extra re-render per keystroke and eliminates the stale intermediate state that occurred between the search input change and the effect firing. --- surfsense_web/app/(home)/blog/blog-magazine.tsx | 10 +++------- 1 file changed, 3 insertions(+), 7 deletions(-) diff --git a/surfsense_web/app/(home)/blog/blog-magazine.tsx b/surfsense_web/app/(home)/blog/blog-magazine.tsx index 96c7f6789..02e5045a9 100644 --- a/surfsense_web/app/(home)/blog/blog-magazine.tsx +++ b/surfsense_web/app/(home)/blog/blog-magazine.tsx @@ -3,7 +3,7 @@ import { format } from "date-fns"; import FuzzySearch from "fuzzy-search"; import Link from "next/link"; -import { useEffect, useMemo, useState } from "react"; +import { useMemo, useState } from "react"; import { Container } from "@/components/container"; import type { BlogEntry } from "./page"; @@ -127,17 +127,13 @@ function MagazineSearchGrid({ [allBlogs] ); - const [results, setResults] = useState(allBlogs); - useEffect(() => { - setResults(searcher.search(search)); - }, [search, searcher]); - const gridItems = useMemo(() => { + const results = search.trim() ? searcher.search(search) : allBlogs; if (search.trim()) { return results; } return results.filter((b) => b.slug !== featuredSlug); - }, [results, search, featuredSlug]); + }, [search, searcher, allBlogs, featuredSlug]); return (
From dcafa364ffad6337003108c992bf7253efda2cfa Mon Sep 17 00:00:00 2001 From: guangyang1206 Date: Wed, 29 Apr 2026 12:12:30 +0800 Subject: [PATCH 13/17] feat(perf): add loading.tsx skeletons for async marketing routes MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Fixes #1243 Add sibling loading.tsx files for all 6 async route segments that were missing instant loading UI, causing blank screens during navigation on slow networks or cold caches. Routes covered: - /docs/[[...slug]] — awaits getDocPage + MDX body - /blog — awaits source.getPages() - /blog/[slug] — awaits params + MDX body - /changelog — awaits source.getPages() - /free — awaits getModels() fetch - /free/[model_slug] — awaits Promise.all([getModel, getAllModels]) Each loading.tsx is a Server Component returning an animate-pulse skeleton that matches its route's layout (header, content area, grid/table/timeline as appropriate). Uses the Skeleton component and Tailwind classes already present in the project. Follows the pattern established in: - app/dashboard/[search_space_id]/logs/loading.tsx - app/dashboard/[search_space_id]/new-chat/loading.tsx --- .../app/(home)/blog/[slug]/loading.tsx | 66 +++++++++++++++++++ surfsense_web/app/(home)/blog/loading.tsx | 50 ++++++++++++++ .../app/(home)/changelog/loading.tsx | 63 ++++++++++++++++++ .../app/(home)/free/[model_slug]/loading.tsx | 65 ++++++++++++++++++ surfsense_web/app/(home)/free/loading.tsx | 60 +++++++++++++++++ .../app/docs/[[...slug]]/loading.tsx | 55 ++++++++++++++++ 6 files changed, 359 insertions(+) create mode 100644 surfsense_web/app/(home)/blog/[slug]/loading.tsx create mode 100644 surfsense_web/app/(home)/blog/loading.tsx create mode 100644 surfsense_web/app/(home)/changelog/loading.tsx create mode 100644 surfsense_web/app/(home)/free/[model_slug]/loading.tsx create mode 100644 surfsense_web/app/(home)/free/loading.tsx create mode 100644 surfsense_web/app/docs/[[...slug]]/loading.tsx diff --git a/surfsense_web/app/(home)/blog/[slug]/loading.tsx b/surfsense_web/app/(home)/blog/[slug]/loading.tsx new file mode 100644 index 000000000..0cce7f80b --- /dev/null +++ b/surfsense_web/app/(home)/blog/[slug]/loading.tsx @@ -0,0 +1,66 @@ +import { Skeleton } from "@/components/ui/skeleton"; + +export default function BlogPostLoading() { + return ( +
+
+ {/* Breadcrumb */} +
+ + + + + +
+ + {/* Tags */} +
+ + +
+ + {/* Title */} +
+ + +
+ + {/* Description */} + + + + {/* Author + date */} +
+ +
+ + +
+
+ + {/* Cover image */} + + + {/* Article body paragraphs */} + {Array.from({ length: 5 }).map((_, i) => ( +
+ + + +
+ ))} + + {/* Sub-heading */} + + + {Array.from({ length: 3 }).map((_, i) => ( +
+ + + +
+ ))} +
+
+ ); +} diff --git a/surfsense_web/app/(home)/blog/loading.tsx b/surfsense_web/app/(home)/blog/loading.tsx new file mode 100644 index 000000000..ddaf345f6 --- /dev/null +++ b/surfsense_web/app/(home)/blog/loading.tsx @@ -0,0 +1,50 @@ +import { Skeleton } from "@/components/ui/skeleton"; + +export default function BlogIndexLoading() { + return ( +
+
+ {/* Header */} +
+ +
+ + {/* Featured post skeleton */} +
+ +
+ + + +
+ + + +
+
+
+ + {/* Search bar skeleton */} +
+ +
+ + {/* Grid of article cards */} +
+ {Array.from({ length: 6 }).map((_, i) => ( +
+ + + + +
+ + +
+
+ ))} +
+
+
+ ); +} diff --git a/surfsense_web/app/(home)/changelog/loading.tsx b/surfsense_web/app/(home)/changelog/loading.tsx new file mode 100644 index 000000000..648f5a5e6 --- /dev/null +++ b/surfsense_web/app/(home)/changelog/loading.tsx @@ -0,0 +1,63 @@ +import { Skeleton } from "@/components/ui/skeleton"; + +export default function ChangelogLoading() { + return ( +
+ {/* Header */} +
+
+
+
+ {/* Breadcrumb */} +
+ + + +
+ + +
+
+
+
+ + {/* Timeline */} +
+
+ {Array.from({ length: 3 }).map((_, i) => ( +
+ {/* Left: date + version */} +
+ + +
+ + {/* Right: content */} +
+
+ {/* Title */} + + {/* Tags */} +
+ + +
+ {/* Body paragraphs */} +
+ + + +
+
+ + +
+
+
+
+ ))} +
+
+
+ ); +} diff --git a/surfsense_web/app/(home)/free/[model_slug]/loading.tsx b/surfsense_web/app/(home)/free/[model_slug]/loading.tsx new file mode 100644 index 000000000..97660188d --- /dev/null +++ b/surfsense_web/app/(home)/free/[model_slug]/loading.tsx @@ -0,0 +1,65 @@ +import { Skeleton } from "@/components/ui/skeleton"; + +export default function FreeModelLoading() { + return ( + <> + {/* Chat area skeleton - fills viewport */} +
+ {/* Chat header */} +
+ + +
+ + {/* Chat messages area */} +
+
+ +
+
+ + + +
+
+ + {/* Input bar */} +
+ +
+
+ + {/* SEO section skeleton */} +
+
+ {/* Breadcrumb */} +
+ + + + + +
+ + + + + +
+ + {/* FAQ skeleton */} + +
+ {Array.from({ length: 4 }).map((_, i) => ( +
+ + + +
+ ))} +
+
+
+ + ); +} diff --git a/surfsense_web/app/(home)/free/loading.tsx b/surfsense_web/app/(home)/free/loading.tsx new file mode 100644 index 000000000..08a4ed6b6 --- /dev/null +++ b/surfsense_web/app/(home)/free/loading.tsx @@ -0,0 +1,60 @@ +import { Skeleton } from "@/components/ui/skeleton"; + +export default function FreeChatLoading() { + return ( +
+
+ {/* Breadcrumb */} +
+ + + +
+ + {/* Hero section */} +
+ + + + +
+ {Array.from({ length: 4 }).map((_, i) => ( + + ))} +
+
+ +
+ + {/* Model table */} +
+ + + +
+ {/* Table header */} +
+ + + + +
+ + {/* Table rows */} + {Array.from({ length: 8 }).map((_, i) => ( +
+
+ + +
+ + + +
+ ))} +
+
+
+
+ ); +} diff --git a/surfsense_web/app/docs/[[...slug]]/loading.tsx b/surfsense_web/app/docs/[[...slug]]/loading.tsx new file mode 100644 index 000000000..6bedcfc40 --- /dev/null +++ b/surfsense_web/app/docs/[[...slug]]/loading.tsx @@ -0,0 +1,55 @@ +import { Skeleton } from "@/components/ui/skeleton"; + +export default function DocsLoading() { + return ( +
+ {/* Title */} + + + {/* Description */} + + +
+ {/* Paragraph block 1 */} +
+ + + +
+ + {/* Sub-heading */} + + + {/* Paragraph block 2 */} +
+ + + + +
+ + {/* Code block placeholder */} + + + {/* Sub-heading */} + + + {/* List items */} +
+ {Array.from({ length: 4 }).map((_, i) => ( +
+ + +
+ ))} +
+ + {/* Paragraph block 3 */} +
+ + +
+
+
+ ); +} From 942077c7736c758bfcfca379e6eaf5902e1320c6 Mon Sep 17 00:00:00 2001 From: yeranyang Date: Tue, 28 Apr 2026 12:17:44 +0800 Subject: [PATCH 14/17] perf(docs): replace full lucide barrel import with explicit icon whitelist Fixes #1241 The docs bundle was importing `{ icons }` from lucide-react, which pulls the entire Lucide icon library (~1 400 SVGs, ~500 kB of JS) into the Next.js docs bundle even though only nine icons are used in docs frontmatter and meta.json files. Replace with a hand-maintained DOCS_ICONS whitelist that imports only the icons that are actually referenced (BookOpen, ClipboardCheck, Compass, Container, Download, FlaskConical, Heart, Unplug, Wrench). To add a new docs icon: import it from lucide-react and add it to the DOCS_ICONS record. The icon() callback remains the same for callers. --- surfsense_web/lib/source.ts | 31 +++++++++++++++++++++++++++++-- 1 file changed, 29 insertions(+), 2 deletions(-) diff --git a/surfsense_web/lib/source.ts b/surfsense_web/lib/source.ts index 162cca57a..b94f990ab 100644 --- a/surfsense_web/lib/source.ts +++ b/surfsense_web/lib/source.ts @@ -1,12 +1,39 @@ import { loader } from "fumadocs-core/source"; -import { icons } from "lucide-react"; +import { + BookOpen, + ClipboardCheck, + Compass, + Container, + Download, + FlaskConical, + Heart, + Unplug, + Wrench, +} from "lucide-react"; import { createElement } from "react"; import { docs } from "@/.source/server"; +/** Explicit whitelist of Lucide icons used in docs frontmatter / meta.json. + * Importing the full `icons` barrel would pull every Lucide icon (~1 400 SVGs) + * into the docs bundle even though only a handful are referenced. Add new icons + * here as docs pages are added. + */ +const DOCS_ICONS: Record = { + BookOpen, + ClipboardCheck, + Compass, + Container, + Download, + FlaskConical, + Heart, + Unplug, + Wrench, +}; + export const source = loader({ baseUrl: "/docs", source: docs.toFumadocsSource(), icon(icon) { - if (icon && icon in icons) return createElement(icons[icon as keyof typeof icons]); + if (icon && icon in DOCS_ICONS) return createElement(DOCS_ICONS[icon]); }, }); From ca9bbee06dbd2e9e50be27f54a3967e20dfc0e7d Mon Sep 17 00:00:00 2001 From: "DESKTOP-RTLN3BA\\$punk" Date: Tue, 28 Apr 2026 21:37:51 -0700 Subject: [PATCH 15/17] chore: linting --- .../versions/130_add_agent_action_log.py | 4 +- .../133_drop_documents_content_hash_unique.py | 4 +- .../app/agents/new_chat/chat_deepagent.py | 19 ++++--- .../app/agents/new_chat/feature_flags.py | 20 +++++-- .../agents/new_chat/middleware/action_log.py | 4 +- .../agents/new_chat/middleware/compaction.py | 4 +- .../new_chat/middleware/context_editing.py | 7 ++- .../agents/new_chat/middleware/doom_loop.py | 21 ++++--- .../new_chat/middleware/knowledge_search.py | 11 ++-- .../new_chat/middleware/noop_injection.py | 12 ++-- .../agents/new_chat/middleware/otel_span.py | 14 ++--- .../agents/new_chat/middleware/permission.py | 41 ++++++++------ .../agents/new_chat/middleware/retry_after.py | 12 +++- .../new_chat/middleware/skills_backends.py | 17 ++++-- .../new_chat/middleware/tool_call_repair.py | 10 ++-- .../new_chat/plugins/year_substituter.py | 29 +++++----- .../app/agents/new_chat/prompts/composer.py | 8 +-- .../app/agents/new_chat/subagents/config.py | 4 +- .../app/agents/new_chat/tools/registry.py | 2 + surfsense_backend/app/observability/otel.py | 9 +-- .../app/routes/agent_flags_route.py | 2 +- .../app/routes/agent_permissions_route.py | 8 +-- .../app/routes/agent_revert_route.py | 6 +- .../app/routes/new_chat_routes.py | 4 +- .../app/services/revert_service.py | 4 +- .../app/utils/user_message_multimodal.py | 4 +- .../agents/new_chat/prompts/test_composer.py | 17 +++--- .../unit/agents/new_chat/test_action_log.py | 56 ++++++++++--------- .../unit/agents/new_chat/test_compaction.py | 20 +++++-- .../agents/new_chat/test_context_editing.py | 3 +- .../agents/new_chat/test_dedup_tool_calls.py | 18 +++++- .../test_default_permissions_layering.py | 8 +-- .../unit/agents/new_chat/test_doom_loop.py | 15 ++--- .../agents/new_chat/test_noop_injection.py | 8 ++- .../new_chat/test_permission_middleware.py | 4 +- .../agents/new_chat/test_plugin_loader.py | 12 ++-- .../unit/agents/new_chat/test_retry_after.py | 10 ++-- .../new_chat/test_specialized_subagents.py | 31 +++++----- .../agents/new_chat/test_tool_call_repair.py | 54 ++++++++++++------ .../test_kb_persistence_filesystem_parity.py | 2 +- .../unit/services/test_revert_service.py | 20 ++----- 41 files changed, 314 insertions(+), 244 deletions(-) diff --git a/surfsense_backend/alembic/versions/130_add_agent_action_log.py b/surfsense_backend/alembic/versions/130_add_agent_action_log.py index 5793988cb..2f06b8ddd 100644 --- a/surfsense_backend/alembic/versions/130_add_agent_action_log.py +++ b/surfsense_backend/alembic/versions/130_add_agent_action_log.py @@ -88,7 +88,5 @@ def upgrade() -> None: def downgrade() -> None: - op.drop_index( - "ix_agent_action_log_thread_created", table_name="agent_action_log" - ) + op.drop_index("ix_agent_action_log_thread_created", table_name="agent_action_log") op.drop_table("agent_action_log") diff --git a/surfsense_backend/alembic/versions/133_drop_documents_content_hash_unique.py b/surfsense_backend/alembic/versions/133_drop_documents_content_hash_unique.py index 88c3e203f..eec53ecb6 100644 --- a/surfsense_backend/alembic/versions/133_drop_documents_content_hash_unique.py +++ b/surfsense_backend/alembic/versions/133_drop_documents_content_hash_unique.py @@ -51,9 +51,7 @@ def upgrade() -> None: # implicit-unique-index variant SQLAlchemy may emit need draining. constraints = _existing_constraint_names(bind, "documents") if "uq_documents_content_hash" in constraints: - op.drop_constraint( - "uq_documents_content_hash", "documents", type_="unique" - ) + op.drop_constraint("uq_documents_content_hash", "documents", type_="unique") indexes = _existing_index_names(bind, "documents") # Some Postgres versions surface the unique constraint via a unique diff --git a/surfsense_backend/app/agents/new_chat/chat_deepagent.py b/surfsense_backend/app/agents/new_chat/chat_deepagent.py index 672570696..3ca44dd4f 100644 --- a/surfsense_backend/app/agents/new_chat/chat_deepagent.py +++ b/surfsense_backend/app/agents/new_chat/chat_deepagent.py @@ -416,10 +416,10 @@ async def create_surfsense_deep_agent( # cheap to build. ``SubAgentMiddleware.__init__`` calls ``create_agent`` # synchronously to compile the general-purpose subagent's full state graph # (every tool + every middleware → pydantic schemas + langgraph compile). - # On gpt-5.x agents that's roughly 1.5–2s of pure CPU work. If we run it + # On gpt-5.x agents that's roughly 1.5-2s of pure CPU work. If we run it # directly here it blocks the asyncio event loop for the whole streaming # task (and any other coroutine sharing this loop), which is why - # "agent creation" wall-clock time used to stretch to ~3–4s. Move the + # "agent creation" wall-clock time used to stretch to ~3-4s. Move the # entire middleware build + main-graph compile into a single # ``asyncio.to_thread`` so the heavy CPU work runs off-loop and the # event loop stays responsive. @@ -587,10 +587,7 @@ def _build_compiled_agent_blocking( # by name. Off by default until the flag flips so existing deployments # don't see new agent types in the task tool description. specialized_subagents: list[SubAgent] = [] - if ( - flags.enable_specialized_subagents - and not flags.disable_new_agent_stack - ): + if flags.enable_specialized_subagents and not flags.disable_new_agent_stack: try: # Specialized subagents share the parent's filesystem + # todo view so their system prompts (which promise @@ -696,7 +693,9 @@ def _build_compiled_agent_blocking( else None ) tool_call_limit_mw = ( - ToolCallLimitMiddleware(thread_limit=300, run_limit=80, exit_behavior="continue") + ToolCallLimitMiddleware( + thread_limit=300, run_limit=80, exit_behavior="continue" + ) if flags.enable_tool_call_limit and not flags.disable_new_agent_stack else None ) @@ -879,7 +878,11 @@ def _build_compiled_agent_blocking( max_tools=12, always_include=[ name - for name in ("update_memory", "get_connected_accounts", "scrape_webpage") + for name in ( + "update_memory", + "get_connected_accounts", + "scrape_webpage", + ) if name in {t.name for t in tools} ], ) diff --git a/surfsense_backend/app/agents/new_chat/feature_flags.py b/surfsense_backend/app/agents/new_chat/feature_flags.py index ce0a3b3fa..89c4fb14f 100644 --- a/surfsense_backend/app/agents/new_chat/feature_flags.py +++ b/surfsense_backend/app/agents/new_chat/feature_flags.py @@ -65,7 +65,9 @@ class AgentFeatureFlags: enable_model_call_limit: bool = False enable_tool_call_limit: bool = False enable_tool_call_repair: bool = False - enable_doom_loop: bool = False # Default OFF until UI handles permission='doom_loop' + enable_doom_loop: bool = ( + False # Default OFF until UI handles permission='doom_loop' + ) # Tier 2 — Safety enable_permission: bool = False # Default OFF for first deploy @@ -79,7 +81,9 @@ class AgentFeatureFlags: # Tier 5 — Snapshot / revert enable_action_log: bool = False - enable_revert_route: bool = False # Backend ships before UI; route returns 503 until this flips + enable_revert_route: bool = ( + False # Backend ships before UI; route returns 503 until this flips + ) # Tier 6 — Plugins enable_plugin_loader: bool = False @@ -109,14 +113,20 @@ class AgentFeatureFlags: enable_compaction_v2=_env_bool("SURFSENSE_ENABLE_COMPACTION_V2", False), enable_retry_after=_env_bool("SURFSENSE_ENABLE_RETRY_AFTER", False), enable_model_fallback=_env_bool("SURFSENSE_ENABLE_MODEL_FALLBACK", False), - enable_model_call_limit=_env_bool("SURFSENSE_ENABLE_MODEL_CALL_LIMIT", False), + enable_model_call_limit=_env_bool( + "SURFSENSE_ENABLE_MODEL_CALL_LIMIT", False + ), enable_tool_call_limit=_env_bool("SURFSENSE_ENABLE_TOOL_CALL_LIMIT", False), - enable_tool_call_repair=_env_bool("SURFSENSE_ENABLE_TOOL_CALL_REPAIR", False), + enable_tool_call_repair=_env_bool( + "SURFSENSE_ENABLE_TOOL_CALL_REPAIR", False + ), enable_doom_loop=_env_bool("SURFSENSE_ENABLE_DOOM_LOOP", False), # Tier 2 enable_permission=_env_bool("SURFSENSE_ENABLE_PERMISSION", False), enable_busy_mutex=_env_bool("SURFSENSE_ENABLE_BUSY_MUTEX", False), - enable_llm_tool_selector=_env_bool("SURFSENSE_ENABLE_LLM_TOOL_SELECTOR", False), + enable_llm_tool_selector=_env_bool( + "SURFSENSE_ENABLE_LLM_TOOL_SELECTOR", False + ), # Tier 4 enable_skills=_env_bool("SURFSENSE_ENABLE_SKILLS", False), enable_specialized_subagents=_env_bool( diff --git a/surfsense_backend/app/agents/new_chat/middleware/action_log.py b/surfsense_backend/app/agents/new_chat/middleware/action_log.py index cf0b57fd4..3675064e8 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/action_log.py +++ b/surfsense_backend/app/agents/new_chat/middleware/action_log.py @@ -101,9 +101,7 @@ class ActionLogMiddleware(AgentMiddleware): async def awrap_tool_call( self, request: ToolCallRequest, - handler: Callable[ - [ToolCallRequest], Awaitable[ToolMessage | Command[Any]] - ], + handler: Callable[[ToolCallRequest], Awaitable[ToolMessage | Command[Any]]], ) -> ToolMessage | Command[Any]: if not self._enabled(): return await handler(request) diff --git a/surfsense_backend/app/agents/new_chat/middleware/compaction.py b/surfsense_backend/app/agents/new_chat/middleware/compaction.py index 8b02089c9..b0a1a7ec5 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/compaction.py +++ b/surfsense_backend/app/agents/new_chat/middleware/compaction.py @@ -177,8 +177,8 @@ class SurfSenseCompactionMiddleware(SummarizationMiddleware): messages_in=len(conversation_messages), extra={"compaction.cutoff_index": int(cutoff_index)}, ): - messages_to_summarize, preserved_messages = ( - super()._partition_messages(conversation_messages, cutoff_index) + messages_to_summarize, preserved_messages = super()._partition_messages( + conversation_messages, cutoff_index ) protected: list[AnyMessage] = [] diff --git a/surfsense_backend/app/agents/new_chat/middleware/context_editing.py b/surfsense_backend/app/agents/new_chat/middleware/context_editing.py index 93ceab8ee..360e3e28f 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/context_editing.py +++ b/surfsense_backend/app/agents/new_chat/middleware/context_editing.py @@ -58,8 +58,7 @@ DEFAULT_SPILL_PREFIX = "/tool_outputs" def _build_spill_placeholder(spill_path: str) -> str: """Build the user-facing placeholder text shown to the model.""" return ( - f"[cleared — full output at {spill_path}; " - f"ask the explore subagent to read it]" + f"[cleared — full output at {spill_path}; ask the explore subagent to read it]" ) @@ -131,7 +130,9 @@ class SpillToBackendEdit(ContextEdit): return candidates = [ - (idx, msg) for idx, msg in enumerate(messages) if isinstance(msg, ToolMessage) + (idx, msg) + for idx, msg in enumerate(messages) + if isinstance(msg, ToolMessage) ] if self.keep >= len(candidates): return diff --git a/surfsense_backend/app/agents/new_chat/middleware/doom_loop.py b/surfsense_backend/app/agents/new_chat/middleware/doom_loop.py index 49ac7dfa8..1dde87752 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/doom_loop.py +++ b/surfsense_backend/app/agents/new_chat/middleware/doom_loop.py @@ -137,16 +137,21 @@ class DoomLoopMiddleware(AgentMiddleware[AgentState[ResponseT], ContextT, Respon triggered_call: dict[str, Any] | None = None for call in message.tool_calls: - name = call.get("name") if isinstance(call, dict) else getattr(call, "name", None) - args = call.get("args") if isinstance(call, dict) else getattr(call, "args", {}) + name = ( + call.get("name") + if isinstance(call, dict) + else getattr(call, "name", None) + ) + args = ( + call.get("args") + if isinstance(call, dict) + else getattr(call, "args", {}) + ) if not isinstance(name, str): continue sig = _signature(name, args) window.append(sig) - if ( - len(window) >= self._threshold - and len(set(window)) == 1 - ): + if len(window) >= self._threshold and len(set(window)) == 1: triggered_call = {"name": name, "params": args or {}} break @@ -209,7 +214,9 @@ class DoomLoopMiddleware(AgentMiddleware[AgentState[ResponseT], ContextT, Respon # tool call proceeds. The frontend's exact reply names may differ — # we tolerate any shape that contains a string with "reject"/"cancel". if isinstance(decision, dict): - kind = str(decision.get("decision_type") or decision.get("type") or "").lower() + kind = str( + decision.get("decision_type") or decision.get("type") or "" + ).lower() if "reject" in kind or "cancel" in kind: return {"jump_to": "end"} return None diff --git a/surfsense_backend/app/agents/new_chat/middleware/knowledge_search.py b/surfsense_backend/app/agents/new_chat/middleware/knowledge_search.py index f39870df6..08ca8e18b 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/knowledge_search.py +++ b/surfsense_backend/app/agents/new_chat/middleware/knowledge_search.py @@ -552,7 +552,7 @@ def _render_priority_message(priority: list[dict[str, Any]]) -> SystemMessage: for entry in priority: score = entry.get("score") mentioned = entry.get("mentioned") - score_str = f"{score:.3f}" if isinstance(score, (int, float)) else "n/a" + score_str = f"{score:.3f}" if isinstance(score, int | float) else "n/a" mark = " [USER-MENTIONED]" if mentioned else "" lines.append(f"- {entry.get('path', '')} (score={score_str}){mark}") body = "\n".join(lines) @@ -593,7 +593,7 @@ class KnowledgePriorityMiddleware(AgentMiddleware): # type: ignore[type-arg] self.top_k = top_k self.mentioned_document_ids = mentioned_document_ids or [] # Tier 4.2: build the kb-planner private Runnable ONCE here so we - # don't pay the create_agent compile cost (50–200ms) on every turn. + # don't pay the create_agent compile cost (50-200ms) on every turn. # Disabled by default behind ``enable_kb_planner_runnable``; when off # the planner falls back to the legacy ``self.llm.ainvoke`` path. self._planner: Runnable | None = None @@ -617,10 +617,7 @@ class KnowledgePriorityMiddleware(AgentMiddleware): # type: ignore[type-arg] if self.llm is None: return None flags = get_flags() - if ( - not flags.enable_kb_planner_runnable - or flags.disable_new_agent_stack - ): + if not flags.enable_kb_planner_runnable or flags.disable_new_agent_stack: return None from app.agents.new_chat.middleware.retry_after import RetryAfterMiddleware @@ -920,7 +917,7 @@ class KnowledgePriorityMiddleware(AgentMiddleware): # type: ignore[type-arg] chunk_ids = doc.get("matched_chunk_ids") or [] if chunk_ids: matched_chunk_ids[doc_id] = [ - int(cid) for cid in chunk_ids if isinstance(cid, (int, str)) + int(cid) for cid in chunk_ids if isinstance(cid, int | str) ] return priority, matched_chunk_ids diff --git a/surfsense_backend/app/agents/new_chat/middleware/noop_injection.py b/surfsense_backend/app/agents/new_chat/middleware/noop_injection.py index f16084892..8628479c7 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/noop_injection.py +++ b/surfsense_backend/app/agents/new_chat/middleware/noop_injection.py @@ -35,9 +35,7 @@ from langchain_core.tools import tool logger = logging.getLogger(__name__) NOOP_TOOL_NAME = "_noop" -NOOP_TOOL_DESCRIPTION = ( - "Do not call this tool. It exists only for API compatibility." -) +NOOP_TOOL_DESCRIPTION = "Do not call this tool. It exists only for API compatibility." @tool(name_or_callable=NOOP_TOOL_NAME, description=NOOP_TOOL_DESCRIPTION) @@ -78,7 +76,9 @@ def _last_ai_has_tool_calls(messages: list[Any]) -> bool: return False -class NoopInjectionMiddleware(AgentMiddleware[AgentState[ResponseT], ContextT, ResponseT]): +class NoopInjectionMiddleware( + AgentMiddleware[AgentState[ResponseT], ContextT, ResponseT] +): """Inject the ``_noop`` tool only when the provider would otherwise 400. The check fires per model call, not at agent build time, because the @@ -116,7 +116,9 @@ class NoopInjectionMiddleware(AgentMiddleware[AgentState[ResponseT], ContextT, R async def awrap_model_call( # type: ignore[override] self, request: ModelRequest[ContextT], - handler: Callable[[ModelRequest[ContextT]], Awaitable[ModelResponse[ResponseT]]], + handler: Callable[ + [ModelRequest[ContextT]], Awaitable[ModelResponse[ResponseT]] + ], ) -> Any: if self._should_inject(request): logger.debug("Injecting _noop tool for provider compatibility") diff --git a/surfsense_backend/app/agents/new_chat/middleware/otel_span.py b/surfsense_backend/app/agents/new_chat/middleware/otel_span.py index 5585cf7a2..f51d2f7bb 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/otel_span.py +++ b/surfsense_backend/app/agents/new_chat/middleware/otel_span.py @@ -56,9 +56,7 @@ class OtelSpanMiddleware(AgentMiddleware): async def awrap_model_call( self, request: ModelRequest, - handler: Callable[ - [ModelRequest], Awaitable[ModelResponse | AIMessage | Any] - ], + handler: Callable[[ModelRequest], Awaitable[ModelResponse | AIMessage | Any]], ) -> ModelResponse | AIMessage | Any: if not ot.is_enabled(): return await handler(request) @@ -81,9 +79,7 @@ class OtelSpanMiddleware(AgentMiddleware): async def awrap_tool_call( self, request: ToolCallRequest, - handler: Callable[ - [ToolCallRequest], Awaitable[ToolMessage | Command[Any]] - ], + handler: Callable[[ToolCallRequest], Awaitable[ToolMessage | Command[Any]]], ) -> ToolMessage | Command[Any]: if not ot.is_enabled(): return await handler(request) @@ -187,7 +183,11 @@ def _annotate_model_response(span: Any, result: Any) -> None: def _annotate_tool_result(span: Any, result: Any) -> None: try: if isinstance(result, ToolMessage): - content = result.content if isinstance(result.content, str) else repr(result.content) + content = ( + result.content + if isinstance(result.content, str) + else repr(result.content) + ) span.set_attribute("tool.output.size", len(content)) status = getattr(result, "status", None) if isinstance(status, str): diff --git a/surfsense_backend/app/agents/new_chat/middleware/permission.py b/surfsense_backend/app/agents/new_chat/middleware/permission.py index f59e70bc0..6e1f42baf 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/permission.py +++ b/surfsense_backend/app/agents/new_chat/middleware/permission.py @@ -145,7 +145,9 @@ class PermissionMiddleware(AgentMiddleware): # type: ignore[type-arg] try: patterns = resolver(args or {}) except Exception: - logger.exception("Pattern resolver for %s raised; using bare name", tool_name) + logger.exception( + "Pattern resolver for %s raised; using bare name", tool_name + ) patterns = [tool_name] if not patterns: patterns = [tool_name] @@ -198,11 +200,14 @@ class PermissionMiddleware(AgentMiddleware): # type: ignore[type-arg] # Tier 3b: permission.asked + interrupt.raised spans (no-op when # OTel is disabled). Both fire here so dashboards can correlate # "we asked X" with "interrupt was actually delivered". - with ot.permission_asked_span( - permission=tool_name, - pattern=patterns[0] if patterns else None, - extra={"permission.patterns": list(patterns)}, - ), ot.interrupt_span(interrupt_type="permission_ask"): + with ( + ot.permission_asked_span( + permission=tool_name, + pattern=patterns[0] if patterns else None, + extra={"permission.patterns": list(patterns)}, + ), + ot.interrupt_span(interrupt_type="permission_ask"), + ): decision = interrupt(payload) if isinstance(decision, dict): return decision @@ -211,9 +216,7 @@ class PermissionMiddleware(AgentMiddleware): # type: ignore[type-arg] return {"decision_type": decision} return {"decision_type": "reject"} - def _persist_always( - self, tool_name: str, patterns: list[str] - ) -> None: + def _persist_always(self, tool_name: str, patterns: list[str]) -> None: """Promote ``always`` reply into runtime allow rules. Persistence to ``agent_permission_rules`` is done by the @@ -276,12 +279,16 @@ class PermissionMiddleware(AgentMiddleware): # type: ignore[type-arg] any_change = False for raw in last.tool_calls: - call = dict(raw) if isinstance(raw, dict) else { - "name": getattr(raw, "name", None), - "args": getattr(raw, "args", {}), - "id": getattr(raw, "id", None), - "type": "tool_call", - } + call = ( + dict(raw) + if isinstance(raw, dict) + else { + "name": getattr(raw, "name", None), + "args": getattr(raw, "args", {}), + "id": getattr(raw, "id", None), + "type": "tool_call", + } + ) name = call.get("name") or "" args = call.get("args") or {} action, patterns, rules = self._evaluate(name, args) @@ -307,7 +314,9 @@ class PermissionMiddleware(AgentMiddleware): # type: ignore[type-arg] feedback = decision.get("feedback") if isinstance(feedback, str) and feedback.strip(): raise CorrectedError(feedback, tool=name) - raise RejectedError(tool=name, pattern=patterns[0] if patterns else None) + raise RejectedError( + tool=name, pattern=patterns[0] if patterns else None + ) else: logger.warning( "Unknown permission decision %r; treating as reject", kind diff --git a/surfsense_backend/app/agents/new_chat/middleware/retry_after.py b/surfsense_backend/app/agents/new_chat/middleware/retry_after.py index 82da6a97c..394bb0371 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/retry_after.py +++ b/surfsense_backend/app/agents/new_chat/middleware/retry_after.py @@ -113,7 +113,9 @@ def _exponential_delay( jitter: bool, ) -> float: """Compute an exponential-backoff delay with optional ±25% jitter.""" - delay = initial_delay * (backoff_factor**attempt) if backoff_factor else initial_delay + delay = ( + initial_delay * (backoff_factor**attempt) if backoff_factor else initial_delay + ) delay = min(delay, max_delay) if jitter and delay > 0: delay *= 1 + random.uniform(-0.25, 0.25) @@ -201,7 +203,9 @@ class RetryAfterMiddleware(AgentMiddleware[AgentState[ResponseT], ContextT, Resp }, ) except Exception: - logger.debug("dispatch_custom_event failed; suppressed", exc_info=True) + logger.debug( + "dispatch_custom_event failed; suppressed", exc_info=True + ) if delay > 0: time.sleep(delay) # Unreachable @@ -210,7 +214,9 @@ class RetryAfterMiddleware(AgentMiddleware[AgentState[ResponseT], ContextT, Resp async def awrap_model_call( # type: ignore[override] self, request: ModelRequest[ContextT], - handler: Callable[[ModelRequest[ContextT]], Awaitable[ModelResponse[ResponseT]]], + handler: Callable[ + [ModelRequest[ContextT]], Awaitable[ModelResponse[ResponseT]] + ], ) -> ModelResponse[ResponseT] | AIMessage: for attempt in range(self.max_retries + 1): try: diff --git a/surfsense_backend/app/agents/new_chat/middleware/skills_backends.py b/surfsense_backend/app/agents/new_chat/middleware/skills_backends.py index 4c3791c87..072d73401 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/skills_backends.py +++ b/surfsense_backend/app/agents/new_chat/middleware/skills_backends.py @@ -29,6 +29,7 @@ gives a clean failure mode if anything tries. from __future__ import annotations +import contextlib import logging from collections.abc import Callable from dataclasses import replace @@ -114,8 +115,10 @@ class BuiltinSkillsBackend(BackendProtocol): infos: list[FileInfo] = [] # Build virtual paths anchored at "/" because CompositeBackend already # stripped the route prefix before calling us. - target_virtual = "/" if target == self.root else ( - "/" + str(target.relative_to(self.root)).replace("\\", "/") + target_virtual = ( + "/" + if target == self.root + else ("/" + str(target.relative_to(self.root)).replace("\\", "/")) ) for child in sorted(target.iterdir()): child_virtual = ( @@ -128,10 +131,8 @@ class BuiltinSkillsBackend(BackendProtocol): "is_dir": child.is_dir(), } if child.is_file(): - try: + with contextlib.suppress(OSError): # pragma: no cover - defensive info["size"] = child.stat().st_size - except OSError: # pragma: no cover - defensive - pass infos.append(info) return infos @@ -163,7 +164,9 @@ class BuiltinSkillsBackend(BackendProtocol): else: content = target.read_bytes() except PermissionError: - responses.append(FileDownloadResponse(path=p, error="permission_denied")) + responses.append( + FileDownloadResponse(path=p, error="permission_denied") + ) continue except OSError as exc: # pragma: no cover - defensive logger.warning("Builtin skill read failed %s: %s", target, exc) @@ -286,6 +289,7 @@ def build_skills_backend_factory( builtin = BuiltinSkillsBackend(builtin_root) if search_space_id is None: + def _factory_builtin_only(runtime: ToolRuntime) -> BackendProtocol: # Default StateBackend is intentionally inert: any path outside the # ``/skills/builtin/`` route resolves to an empty per-runtime state @@ -294,6 +298,7 @@ def build_skills_backend_factory( default=StateBackend(runtime), routes={SKILLS_BUILTIN_PREFIX: builtin}, ) + return _factory_builtin_only def _factory_with_space(runtime: ToolRuntime) -> BackendProtocol: diff --git a/surfsense_backend/app/agents/new_chat/middleware/tool_call_repair.py b/surfsense_backend/app/agents/new_chat/middleware/tool_call_repair.py index 6c3bc674d..54df0cc60 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/tool_call_repair.py +++ b/surfsense_backend/app/agents/new_chat/middleware/tool_call_repair.py @@ -51,13 +51,15 @@ def _coerce_existing_tool_call(call: Any) -> dict[str, Any]: } -class ToolCallNameRepairMiddleware(AgentMiddleware[AgentState[ResponseT], ContextT, ResponseT]): +class ToolCallNameRepairMiddleware( + AgentMiddleware[AgentState[ResponseT], ContextT, ResponseT] +): """Two-stage tool-name repair on the most recent ``AIMessage``. Args: registered_tool_names: Set of canonically-registered tool names. ``invalid`` should be in this set so the fallback dispatches. - fuzzy_match_threshold: Optional ``difflib`` ratio (0–1) for the + fuzzy_match_threshold: Optional ``difflib`` ratio (0-1) for the fuzzy-match step that runs *between* lowercase and invalid. Set to ``None`` to disable fuzzy matching (opencode parity). """ @@ -77,9 +79,9 @@ class ToolCallNameRepairMiddleware(AgentMiddleware[AgentState[ResponseT], Contex def _registered_for_runtime(self, runtime: Runtime[ContextT]) -> set[str]: """Allow runtime overrides to expand the set (e.g. dynamic MCP tools).""" ctx_tools = getattr(runtime.context, "registered_tool_names", None) - if isinstance(ctx_tools, (set, frozenset)): + if isinstance(ctx_tools, set | frozenset): return self._registered | set(ctx_tools) - if isinstance(ctx_tools, (list, tuple)): + if isinstance(ctx_tools, list | tuple): return self._registered | set(ctx_tools) return self._registered diff --git a/surfsense_backend/app/agents/new_chat/plugins/year_substituter.py b/surfsense_backend/app/agents/new_chat/plugins/year_substituter.py index 927d533d5..3e2e631d2 100644 --- a/surfsense_backend/app/agents/new_chat/plugins/year_substituter.py +++ b/surfsense_backend/app/agents/new_chat/plugins/year_substituter.py @@ -52,25 +52,26 @@ class _YearSubstituterMiddleware(AgentMiddleware): async def awrap_tool_call( self, request: ToolCallRequest, - handler: Callable[ - [ToolCallRequest], Awaitable[ToolMessage | Command[Any]] - ], + handler: Callable[[ToolCallRequest], Awaitable[ToolMessage | Command[Any]]], ) -> ToolMessage | Command[Any]: result = await handler(request) try: from langchain_core.messages import ToolMessage - if isinstance(result, ToolMessage) and isinstance(result.content, str): - if "{{year}}" in result.content: - new_text = result.content.replace("{{year}}", self._year) - result = ToolMessage( - content=new_text, - tool_call_id=result.tool_call_id, - id=result.id, - name=result.name, - status=result.status, - artifact=result.artifact, - ) + if ( + isinstance(result, ToolMessage) + and isinstance(result.content, str) + and "{{year}}" in result.content + ): + new_text = result.content.replace("{{year}}", self._year) + result = ToolMessage( + content=new_text, + tool_call_id=result.tool_call_id, + id=result.id, + name=result.name, + status=result.status, + artifact=result.artifact, + ) except Exception: # pragma: no cover - defensive logger.exception("year_substituter plugin failed; passing original result") return result diff --git a/surfsense_backend/app/agents/new_chat/prompts/composer.py b/surfsense_backend/app/agents/new_chat/prompts/composer.py index bad033490..77b86aeef 100644 --- a/surfsense_backend/app/agents/new_chat/prompts/composer.py +++ b/surfsense_backend/app/agents/new_chat/prompts/composer.py @@ -62,7 +62,9 @@ ProviderVariant = str # More specific patterns must come first (e.g. ``codex`` before # ``openai_reasoning`` because codex model ids contain ``gpt``). -_OPENAI_CODEX_RE = re.compile(r"\b(gpt-codex|codex-mini|gpt-[\d.]+-codex)\b", re.IGNORECASE) +_OPENAI_CODEX_RE = re.compile( + r"\b(gpt-codex|codex-mini|gpt-[\d.]+-codex)\b", re.IGNORECASE +) _OPENAI_REASONING_RE = re.compile(r"\b(gpt-5|o\d|o-)", re.IGNORECASE) _OPENAI_CLASSIC_RE = re.compile(r"\bgpt-4", re.IGNORECASE) _ANTHROPIC_RE = re.compile(r"\bclaude\b", re.IGNORECASE) @@ -257,9 +259,7 @@ def _build_tools_section( ) if known_disabled: disabled_list = ", ".join( - _format_tool_label(n) - for n in ALL_TOOL_NAMES_ORDERED - if n in known_disabled + _format_tool_label(n) for n in ALL_TOOL_NAMES_ORDERED if n in known_disabled ) parts.append( "\n" diff --git a/surfsense_backend/app/agents/new_chat/subagents/config.py b/surfsense_backend/app/agents/new_chat/subagents/config.py index e20bc06bf..b36d35fa0 100644 --- a/surfsense_backend/app/agents/new_chat/subagents/config.py +++ b/surfsense_backend/app/agents/new_chat/subagents/config.py @@ -279,9 +279,7 @@ def build_explore_subagent( selected_tools = _filter_tools(tools, EXPLORE_READ_TOOLS) deny_rules = _read_only_deny_rules() - permission_mw = _build_permission_middleware( - deny_rules, origin="subagent_explore" - ) + permission_mw = _build_permission_middleware(deny_rules, origin="subagent_explore") spec: dict = { "name": "explore", diff --git a/surfsense_backend/app/agents/new_chat/tools/registry.py b/surfsense_backend/app/agents/new_chat/tools/registry.py index f5ee1a61d..fce1bf872 100644 --- a/surfsense_backend/app/agents/new_chat/tools/registry.py +++ b/surfsense_backend/app/agents/new_chat/tools/registry.py @@ -111,6 +111,8 @@ from .update_memory import create_update_memory_tool, create_update_team_memory_ from .video_presentation import create_generate_video_presentation_tool from .web_search import create_web_search_tool +logger = logging.getLogger(__name__) + # ============================================================================= # Tool Definition # ============================================================================= diff --git a/surfsense_backend/app/observability/otel.py b/surfsense_backend/app/observability/otel.py index 0229524f2..4f2257ab7 100644 --- a/surfsense_backend/app/observability/otel.py +++ b/surfsense_backend/app/observability/otel.py @@ -22,6 +22,7 @@ Goals from __future__ import annotations +import contextlib import logging import os from collections.abc import Iterator @@ -154,18 +155,14 @@ def span( with tracer.start_as_current_span(name) as sp: if attributes: - try: + with contextlib.suppress(Exception): # pragma: no cover — defensive sp.set_attributes(attributes) - except Exception: # pragma: no cover — defensive - pass try: yield sp except BaseException as exc: - try: + with contextlib.suppress(Exception): # pragma: no cover — defensive sp.record_exception(exc) sp.set_status(_OtStatus(_OtStatusCode.ERROR, str(exc))) - except Exception: # pragma: no cover — defensive - pass raise diff --git a/surfsense_backend/app/routes/agent_flags_route.py b/surfsense_backend/app/routes/agent_flags_route.py index d3c90a58d..5732a8dfb 100644 --- a/surfsense_backend/app/routes/agent_flags_route.py +++ b/surfsense_backend/app/routes/agent_flags_route.py @@ -59,7 +59,7 @@ class AgentFeatureFlagsRead(BaseModel): enable_otel: bool @classmethod - def from_flags(cls, flags: AgentFeatureFlags) -> "AgentFeatureFlagsRead": + def from_flags(cls, flags: AgentFeatureFlags) -> AgentFeatureFlagsRead: # asdict() avoids missing-field bugs when AgentFeatureFlags grows. return cls(**asdict(flags)) diff --git a/surfsense_backend/app/routes/agent_permissions_route.py b/surfsense_backend/app/routes/agent_permissions_route.py index e87af29c7..1c76e00e6 100644 --- a/surfsense_backend/app/routes/agent_permissions_route.py +++ b/surfsense_backend/app/routes/agent_permissions_route.py @@ -210,7 +210,7 @@ async def create_rule( session.add(row) try: await session.commit() - except IntegrityError: + except IntegrityError as err: await session.rollback() raise HTTPException( status_code=409, @@ -218,7 +218,7 @@ async def create_rule( "An identical rule already exists for this scope. Update the " "existing rule instead." ), - ) + ) from err await session.refresh(row) return _to_read(row) @@ -248,12 +248,12 @@ async def update_rule( try: await session.commit() - except IntegrityError: + except IntegrityError as err: await session.rollback() raise HTTPException( status_code=409, detail="Update would create a duplicate rule for this scope.", - ) + ) from err await session.refresh(row) return _to_read(row) diff --git a/surfsense_backend/app/routes/agent_revert_route.py b/surfsense_backend/app/routes/agent_revert_route.py index 2f6fe6a32..cbe4e7417 100644 --- a/surfsense_backend/app/routes/agent_revert_route.py +++ b/surfsense_backend/app/routes/agent_revert_route.py @@ -97,10 +97,12 @@ async def revert_agent_action( action=action, requester_user_id=str(user.id) if user is not None else None, ) - except Exception: + except Exception as err: logger.exception("Revert dispatch raised for action_id=%s", action_id) await session.rollback() - raise HTTPException(status_code=500, detail="Internal error during revert.") + raise HTTPException( + status_code=500, detail="Internal error during revert." + ) from err if outcome.status == "ok": await session.commit() diff --git a/surfsense_backend/app/routes/new_chat_routes.py b/surfsense_backend/app/routes/new_chat_routes.py index cbc660222..b5560d90d 100644 --- a/surfsense_backend/app/routes/new_chat_routes.py +++ b/surfsense_backend/app/routes/new_chat_routes.py @@ -1242,7 +1242,9 @@ async def handle_new_chat( await session.close() image_urls = ( - [p.as_data_url() for p in request.user_images] if request.user_images else None + [p.as_data_url() for p in request.user_images] + if request.user_images + else None ) return StreamingResponse( diff --git a/surfsense_backend/app/services/revert_service.py b/surfsense_backend/app/services/revert_service.py index e072f90c6..f3630e0b4 100644 --- a/surfsense_backend/app/services/revert_service.py +++ b/surfsense_backend/app/services/revert_service.py @@ -79,9 +79,7 @@ async def load_action( return result.scalars().first() -async def load_thread( - session: AsyncSession, *, thread_id: int -) -> NewChatThread | None: +async def load_thread(session: AsyncSession, *, thread_id: int) -> NewChatThread | None: stmt = select(NewChatThread).where(NewChatThread.id == thread_id) result = await session.execute(stmt) return result.scalars().first() diff --git a/surfsense_backend/app/utils/user_message_multimodal.py b/surfsense_backend/app/utils/user_message_multimodal.py index 1d0691697..dc9a6fe76 100644 --- a/surfsense_backend/app/utils/user_message_multimodal.py +++ b/surfsense_backend/app/utils/user_message_multimodal.py @@ -7,7 +7,9 @@ import binascii from typing import Any -def build_human_message_content(final_query: str, image_data_urls: list[str]) -> str | list[dict[str, Any]]: +def build_human_message_content( + final_query: str, image_data_urls: list[str] +) -> str | list[dict[str, Any]]: if not image_data_urls: return final_query parts: list[dict[str, Any]] = [{"type": "text", "text": final_query}] diff --git a/surfsense_backend/tests/unit/agents/new_chat/prompts/test_composer.py b/surfsense_backend/tests/unit/agents/new_chat/prompts/test_composer.py index d08bbc8cf..aa0c215b9 100644 --- a/surfsense_backend/tests/unit/agents/new_chat/prompts/test_composer.py +++ b/surfsense_backend/tests/unit/agents/new_chat/prompts/test_composer.py @@ -90,9 +90,7 @@ class TestCompose: assert "" in prompt assert "[citation:chunk_id]" in prompt - def test_team_visibility_uses_team_variants( - self, fixed_today: datetime - ) -> None: + def test_team_visibility_uses_team_variants(self, fixed_today: datetime) -> None: prompt = compose_system_prompt( today=fixed_today, thread_visibility=ChatVisibility.SEARCH_SPACE, @@ -145,9 +143,7 @@ class TestCompose: assert "Generate Image" in prompt assert "Generate Podcast" in prompt - def test_mcp_routing_block_emits_when_provided( - self, fixed_today: datetime - ) -> None: + def test_mcp_routing_block_emits_when_provided(self, fixed_today: datetime) -> None: prompt = compose_system_prompt( today=fixed_today, mcp_connector_tools={"My GitLab": ["gitlab_search", "gitlab_create_mr"]}, @@ -162,9 +158,7 @@ class TestCompose: prompt = compose_system_prompt(today=fixed_today, mcp_connector_tools={}) assert "" not in prompt - def test_provider_block_renders_when_anthropic( - self, fixed_today: datetime - ) -> None: + def test_provider_block_renders_when_anthropic(self, fixed_today: datetime) -> None: prompt = compose_system_prompt( today=fixed_today, model_name="anthropic:claude-3-5-sonnet" ) @@ -267,7 +261,10 @@ class TestStableOrderingForCacheStability: ) b = compose_system_prompt( today=fixed_today, - enabled_tool_names={"scrape_webpage", "web_search"}, # set order shouldn't matter + enabled_tool_names={ + "scrape_webpage", + "web_search", + }, # set order shouldn't matter mcp_connector_tools={"X": ["x_a", "x_b"]}, ) assert a == b diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_action_log.py b/surfsense_backend/tests/unit/agents/new_chat/test_action_log.py index 6834b5be7..aad1524c9 100644 --- a/surfsense_backend/tests/unit/agents/new_chat/test_action_log.py +++ b/surfsense_backend/tests/unit/agents/new_chat/test_action_log.py @@ -83,7 +83,11 @@ class TestActionLogMiddlewareDisabled: async def test_no_op_when_flag_off(self, patch_get_flags) -> None: mw = ActionLogMiddleware(thread_id=1, search_space_id=1, user_id=None) request = _FakeRequest( - tool_call={"name": "make_widget", "args": {"color": "red", "size": 1}, "id": "tc1"} + tool_call={ + "name": "make_widget", + "args": {"color": "red", "size": 1}, + "id": "tc1", + } ) handler = AsyncMock(return_value=ToolMessage(content="ok", tool_call_id="tc1")) with patch_get_flags(_disabled_flags()): @@ -117,13 +121,12 @@ class TestActionLogMiddlewarePersistence: "id": "tc-abc", }, ) - result_msg = ToolMessage( - content="ok", tool_call_id="tc-abc", id="msg-1" - ) + result_msg = ToolMessage(content="ok", tool_call_id="tc-abc", id="msg-1") handler = AsyncMock(return_value=result_msg) - with patch_get_flags(_enabled_flags()), patch( - "app.db.shielded_async_session", side_effect=lambda: factory() + with ( + patch_get_flags(_enabled_flags()), + patch("app.db.shielded_async_session", side_effect=lambda: factory()), ): result = await mw.awrap_tool_call(request, handler) @@ -151,9 +154,11 @@ class TestActionLogMiddlewarePersistence: ) handler = AsyncMock(side_effect=ValueError("boom")) - with patch_get_flags(_enabled_flags()), patch( - "app.db.shielded_async_session", side_effect=lambda: factory() - ), pytest.raises(ValueError, match="boom"): + with ( + patch_get_flags(_enabled_flags()), + patch("app.db.shielded_async_session", side_effect=lambda: factory()), + pytest.raises(ValueError, match="boom"), + ): await mw.awrap_tool_call(request, handler) assert len(captured["rows"]) == 1 @@ -177,8 +182,9 @@ class TestActionLogMiddlewarePersistence: def _exploding_session(): raise RuntimeError("DB is down") - with patch_get_flags(_enabled_flags()), patch( - "app.db.shielded_async_session", side_effect=_exploding_session + with ( + patch_get_flags(_enabled_flags()), + patch("app.db.shielded_async_session", side_effect=_exploding_session), ): result = await mw.awrap_tool_call(request, handler) assert result is result_msg @@ -218,8 +224,9 @@ class TestReverseDescriptor: ) handler = AsyncMock(return_value=result_msg) - with patch_get_flags(_enabled_flags()), patch( - "app.db.shielded_async_session", side_effect=lambda: factory() + with ( + patch_get_flags(_enabled_flags()), + patch("app.db.shielded_async_session", side_effect=lambda: factory()), ): await mw.awrap_tool_call(request, handler) @@ -257,8 +264,9 @@ class TestReverseDescriptor: result_msg = ToolMessage(content="ok", tool_call_id="tc1") handler = AsyncMock(return_value=result_msg) - with patch_get_flags(_enabled_flags()), patch( - "app.db.shielded_async_session", side_effect=lambda: factory() + with ( + patch_get_flags(_enabled_flags()), + patch("app.db.shielded_async_session", side_effect=lambda: factory()), ): await mw.awrap_tool_call(request, handler) @@ -275,11 +283,10 @@ class TestReverseDescriptor: request = _FakeRequest( tool_call={"name": "unknown_tool", "args": {}, "id": "tc1"} ) - handler = AsyncMock( - return_value=ToolMessage(content="ok", tool_call_id="tc1") - ) - with patch_get_flags(_enabled_flags()), patch( - "app.db.shielded_async_session", side_effect=lambda: factory() + handler = AsyncMock(return_value=ToolMessage(content="ok", tool_call_id="tc1")) + with ( + patch_get_flags(_enabled_flags()), + patch("app.db.shielded_async_session", side_effect=lambda: factory()), ): await mw.awrap_tool_call(request, handler) row = captured["rows"][0] @@ -298,11 +305,10 @@ class TestArgsTruncation: request = _FakeRequest( tool_call={"name": "make_widget", "args": {"blob": huge}, "id": "tc1"}, ) - handler = AsyncMock( - return_value=ToolMessage(content="ok", tool_call_id="tc1") - ) - with patch_get_flags(_enabled_flags()), patch( - "app.db.shielded_async_session", side_effect=lambda: factory() + handler = AsyncMock(return_value=ToolMessage(content="ok", tool_call_id="tc1")) + with ( + patch_get_flags(_enabled_flags()), + patch("app.db.shielded_async_session", side_effect=lambda: factory()), ): await mw.awrap_tool_call(request, handler) row = captured["rows"][0] diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_compaction.py b/surfsense_backend/tests/unit/agents/new_chat/test_compaction.py index 4d8d6805c..c6d4cc452 100644 --- a/surfsense_backend/tests/unit/agents/new_chat/test_compaction.py +++ b/surfsense_backend/tests/unit/agents/new_chat/test_compaction.py @@ -26,10 +26,16 @@ class TestIsProtectedSystemMessage: assert _is_protected_system_message(msg) is True def test_unprotected_system_message(self) -> None: - assert _is_protected_system_message(SystemMessage(content="random instructions")) is False + assert ( + _is_protected_system_message(SystemMessage(content="random instructions")) + is False + ) def test_human_message_never_protected(self) -> None: - assert _is_protected_system_message(HumanMessage(content="...")) is False + assert ( + _is_protected_system_message(HumanMessage(content="...")) + is False + ) def test_tolerates_leading_whitespace(self) -> None: msg = SystemMessage(content=" \n\n...") @@ -97,11 +103,17 @@ class TestPartitionMessages: assert protected not in to_summary assert protected in preserved # The non-protected old messages remain in to_summary - assert any(isinstance(m, HumanMessage) and m.content == "old human" for m in to_summary) + assert any( + isinstance(m, HumanMessage) and m.content == "old human" for m in to_summary + ) def test_unprotected_messages_unaffected(self) -> None: partitioner = self._build_partitioner() - msgs = [HumanMessage(content="a"), HumanMessage(content="b"), HumanMessage(content="c")] + msgs = [ + HumanMessage(content="a"), + HumanMessage(content="b"), + HumanMessage(content="c"), + ] to_summary, preserved = partitioner._partition_messages(msgs, 2) assert [m.content for m in to_summary] == ["a", "b"] assert [m.content for m in preserved] == ["c"] diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_context_editing.py b/surfsense_backend/tests/unit/agents/new_chat/test_context_editing.py index 3c31155d4..ba2246413 100644 --- a/surfsense_backend/tests/unit/agents/new_chat/test_context_editing.py +++ b/surfsense_backend/tests/unit/agents/new_chat/test_context_editing.py @@ -70,7 +70,8 @@ class TestSpillEdit: # Earlier ToolMessages should now contain the placeholder text cleared = [ - m for m in tool_messages + m + for m in tool_messages if isinstance(m.content, str) and m.content.startswith("[cleared") ] assert len(cleared) >= 1 diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_dedup_tool_calls.py b/surfsense_backend/tests/unit/agents/new_chat/test_dedup_tool_calls.py index 95017d744..e04f50815 100644 --- a/surfsense_backend/tests/unit/agents/new_chat/test_dedup_tool_calls.py +++ b/surfsense_backend/tests/unit/agents/new_chat/test_dedup_tool_calls.py @@ -46,9 +46,21 @@ def test_callable_dedup_key_takes_priority() -> None: state = { "messages": [ _msg( - {"name": "create_doc", "args": {"parent_id": "x", "title": "y"}, "id": "1"}, - {"name": "create_doc", "args": {"parent_id": "x", "title": "y"}, "id": "2"}, - {"name": "create_doc", "args": {"parent_id": "x", "title": "z"}, "id": "3"}, + { + "name": "create_doc", + "args": {"parent_id": "x", "title": "y"}, + "id": "1", + }, + { + "name": "create_doc", + "args": {"parent_id": "x", "title": "y"}, + "id": "2", + }, + { + "name": "create_doc", + "args": {"parent_id": "x", "title": "z"}, + "id": "3", + }, ) ] } diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_default_permissions_layering.py b/surfsense_backend/tests/unit/agents/new_chat/test_default_permissions_layering.py index d49edbfec..ac6b5d95c 100644 --- a/surfsense_backend/tests/unit/agents/new_chat/test_default_permissions_layering.py +++ b/surfsense_backend/tests/unit/agents/new_chat/test_default_permissions_layering.py @@ -84,9 +84,7 @@ class TestConnectorDenyOverridesDefaultAllow: Rule(permission="linear_create_issue", pattern="*", action="deny") ] ) - rules = evaluate_many( - "linear_create_issue", ["linear_create_issue"], *rulesets - ) + rules = evaluate_many("linear_create_issue", ["linear_create_issue"], *rulesets) assert aggregate_action(rules) == "deny" def test_default_allow_still_applies_to_other_tools(self) -> None: @@ -124,5 +122,7 @@ class TestUserRuleOverridesDefault: rules=[Rule(permission="send_*", pattern="*", action="deny")], origin="user", ) - rules = evaluate_many("send_gmail_email", ["send_gmail_email"], defaults, user_ruleset) + rules = evaluate_many( + "send_gmail_email", ["send_gmail_email"], defaults, user_ruleset + ) assert aggregate_action(rules) == "deny" diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_doom_loop.py b/surfsense_backend/tests/unit/agents/new_chat/test_doom_loop.py index c54163dc3..802129bf6 100644 --- a/surfsense_backend/tests/unit/agents/new_chat/test_doom_loop.py +++ b/surfsense_backend/tests/unit/agents/new_chat/test_doom_loop.py @@ -64,22 +64,17 @@ def test_threshold_triggers_after_n_identical_calls() -> None: runtime, ) name = type(excinfo.value).__name__.lower() - assert ( - "interrupt" in name - or "runtimeerror" in name - ), f"Expected an interrupt-style exception, got {name}" + assert "interrupt" in name or "runtimeerror" in name, ( + f"Expected an interrupt-style exception, got {name}" + ) def test_does_not_trigger_when_args_differ() -> None: mw = DoomLoopMiddleware(threshold=2) runtime = _FakeRuntime() - out = mw.after_model( - {"messages": [_msg_calling("repeat", {"x": 1}, "1")]}, runtime - ) + out = mw.after_model({"messages": [_msg_calling("repeat", {"x": 1}, "1")]}, runtime) assert out is None - out = mw.after_model( - {"messages": [_msg_calling("repeat", {"x": 2}, "2")]}, runtime - ) + out = mw.after_model({"messages": [_msg_calling("repeat", {"x": 2}, "2")]}, runtime) assert out is None diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_noop_injection.py b/surfsense_backend/tests/unit/agents/new_chat/test_noop_injection.py index 8555eea76..346271f4b 100644 --- a/surfsense_backend/tests/unit/agents/new_chat/test_noop_injection.py +++ b/surfsense_backend/tests/unit/agents/new_chat/test_noop_injection.py @@ -91,7 +91,9 @@ class TestShouldInject: mw = NoopInjectionMiddleware() req = _FakeRequest( tools=[object()], - messages=[AIMessage(content="", tool_calls=[{"name": "x", "args": {}, "id": "1"}])], + messages=[ + AIMessage(content="", tool_calls=[{"name": "x", "args": {}, "id": "1"}]) + ], model=_LiteLLMModel(), ) assert mw._should_inject(req) is False @@ -109,7 +111,9 @@ class TestShouldInject: mw = NoopInjectionMiddleware() req = _FakeRequest( tools=[], - messages=[AIMessage(content="", tool_calls=[{"name": "x", "args": {}, "id": "1"}])], + messages=[ + AIMessage(content="", tool_calls=[{"name": "x", "args": {}, "id": "1"}]) + ], model=_OpenAIModel(), ) assert mw._should_inject(req) is False diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_permission_middleware.py b/surfsense_backend/tests/unit/agents/new_chat/test_permission_middleware.py index 194a6eb27..a997c8d61 100644 --- a/surfsense_backend/tests/unit/agents/new_chat/test_permission_middleware.py +++ b/surfsense_backend/tests/unit/agents/new_chat/test_permission_middleware.py @@ -111,6 +111,4 @@ class TestAsk: assert out is None # call kept # Runtime ruleset got the always-allow rule new_rules = [r for r in mw._runtime_ruleset.rules if r.action == "allow"] - assert any( - r.permission == "send_email" for r in new_rules - ) + assert any(r.permission == "send_email" for r in new_rules) diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_plugin_loader.py b/surfsense_backend/tests/unit/agents/new_chat/test_plugin_loader.py index 8d98e1328..c2118c697 100644 --- a/surfsense_backend/tests/unit/agents/new_chat/test_plugin_loader.py +++ b/surfsense_backend/tests/unit/agents/new_chat/test_plugin_loader.py @@ -69,7 +69,9 @@ class TestPluginLoaderBasics: "app.agents.new_chat.plugin_loader.entry_points", return_value=[ep], ): - result = load_plugin_middlewares(_ctx(), allowed_plugin_names=["allowed_only"]) + result = load_plugin_middlewares( + _ctx(), allowed_plugin_names=["allowed_only"] + ) assert result == [] assert not called @@ -135,9 +137,7 @@ class TestPluginLoaderIsolation: _FakeEntryPoint("crashing", crashing_factory), _FakeEntryPoint("ok", year_substituter_factory), ] - with patch( - "app.agents.new_chat.plugin_loader.entry_points", return_value=eps - ): + with patch("app.agents.new_chat.plugin_loader.entry_points", return_value=eps): result = load_plugin_middlewares( _ctx(), allowed_plugin_names={"crashing", "ok"} ) @@ -151,9 +151,7 @@ class TestAllowlistEnv: assert load_allowed_plugin_names_from_env() == set() def test_parses_comma_separated_value(self, monkeypatch) -> None: - monkeypatch.setenv( - "SURFSENSE_ALLOWED_PLUGINS", " year_substituter , noisy , " - ) + monkeypatch.setenv("SURFSENSE_ALLOWED_PLUGINS", " year_substituter , noisy , ") assert load_allowed_plugin_names_from_env() == { "year_substituter", "noisy", diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_retry_after.py b/surfsense_backend/tests/unit/agents/new_chat/test_retry_after.py index 39dd9bf00..d23fd693b 100644 --- a/surfsense_backend/tests/unit/agents/new_chat/test_retry_after.py +++ b/surfsense_backend/tests/unit/agents/new_chat/test_retry_after.py @@ -18,7 +18,7 @@ class _FakeResponse: self.headers = headers -class _FakeRateLimit(Exception): +class _FakeRateLimitError(Exception): def __init__(self, msg: str, headers: dict[str, str] | None = None) -> None: super().__init__(msg) if headers is not None: @@ -27,15 +27,15 @@ class _FakeRateLimit(Exception): class TestExtractRetryAfter: def test_seconds_header(self) -> None: - exc = _FakeRateLimit("rate", {"Retry-After": "30"}) + exc = _FakeRateLimitError("rate", {"Retry-After": "30"}) assert _extract_retry_after_seconds(exc) == 30.0 def test_milliseconds_header_overrides_seconds(self) -> None: - exc = _FakeRateLimit("rate", {"retry-after-ms": "1500"}) + exc = _FakeRateLimitError("rate", {"retry-after-ms": "1500"}) assert _extract_retry_after_seconds(exc) == 1.5 def test_case_insensitive(self) -> None: - exc = _FakeRateLimit("rate", {"RETRY-AFTER": "12"}) + exc = _FakeRateLimitError("rate", {"RETRY-AFTER": "12"}) assert _extract_retry_after_seconds(exc) == 12.0 def test_falls_back_to_message_regex(self) -> None: @@ -67,7 +67,7 @@ class TestIsNonRetryable: class TestDelayCalculation: def test_takes_max_of_backoff_and_header(self) -> None: mw = RetryAfterMiddleware(max_retries=3, initial_delay=1.0, jitter=False) - exc = _FakeRateLimit("rl", {"retry-after": "10"}) + exc = _FakeRateLimitError("rl", {"retry-after": "10"}) delay = mw._delay_for_attempt(0, exc) assert delay == pytest.approx(10.0) diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_specialized_subagents.py b/surfsense_backend/tests/unit/agents/new_chat/test_specialized_subagents.py index 3819b4605..0adb578ce 100644 --- a/surfsense_backend/tests/unit/agents/new_chat/test_specialized_subagents.py +++ b/surfsense_backend/tests/unit/agents/new_chat/test_specialized_subagents.py @@ -122,7 +122,9 @@ class TestExploreSubagent: def test_includes_permission_middleware_with_deny_rules(self) -> None: spec = build_explore_subagent(tools=ALL_TOOLS) permission_mws = [ - m for m in spec["middleware"] if isinstance(m, PermissionMiddleware) # type: ignore[index] + m + for m in spec["middleware"] + if isinstance(m, PermissionMiddleware) # type: ignore[index] ] assert len(permission_mws) == 1 ruleset = permission_mws[0]._static_rulesets[0] @@ -164,7 +166,9 @@ class TestReportWriterSubagent: def test_deny_rules_block_writes_but_allow_generate_report(self) -> None: spec = build_report_writer_subagent(tools=ALL_TOOLS) permission_mws = [ - m for m in spec["middleware"] if isinstance(m, PermissionMiddleware) # type: ignore[index] + m + for m in spec["middleware"] + if isinstance(m, PermissionMiddleware) # type: ignore[index] ] ruleset = permission_mws[0]._static_rulesets[0] deny_patterns = {r.permission for r in ruleset.rules if r.action == "deny"} @@ -194,17 +198,15 @@ class TestConnectorNegotiatorSubagent: def test_deny_ruleset_blocks_mutating_connector_tools(self) -> None: spec = build_connector_negotiator_subagent(tools=ALL_TOOLS) permission_mws = [ - m for m in spec["middleware"] if isinstance(m, PermissionMiddleware) # type: ignore[index] + m + for m in spec["middleware"] + if isinstance(m, PermissionMiddleware) # type: ignore[index] ] ruleset = permission_mws[0]._static_rulesets[0] deny_patterns = {r.permission for r in ruleset.rules if r.action == "deny"} # `linear_create_issue` matches the `*_create` deny pattern. - assert any( - _wildcard_matches(p, "linear_create_issue") for p in deny_patterns - ) - assert any( - _wildcard_matches(p, "slack_send_message") for p in deny_patterns - ) + assert any(_wildcard_matches(p, "linear_create_issue") for p in deny_patterns) + assert any(_wildcard_matches(p, "slack_send_message") for p in deny_patterns) class TestBuildSpecializedSubagents: @@ -242,8 +244,7 @@ class TestBuildSpecializedSubagents: # order: extra → custom → patch → dedup. sentinel_idx = mws.index(sentinel) perm_idx = next( - (i for i, m in enumerate(mws) - if isinstance(m, PermissionMiddleware)), + (i for i, m in enumerate(mws) if isinstance(m, PermissionMiddleware)), None, ) assert perm_idx is not None @@ -259,7 +260,9 @@ class TestFilterToolsWarningSuppression: from app.agents.new_chat.subagents.config import _filter_tools - with caplog.at_level(logging.INFO, logger="app.agents.new_chat.subagents.config"): + with caplog.at_level( + logging.INFO, logger="app.agents.new_chat.subagents.config" + ): # Allowed set asks for two registry tools (one present, one # not) plus a bunch of middleware-provided names. _filter_tools( @@ -275,9 +278,7 @@ class TestFilterToolsWarningSuppression: }, ) - warnings = [ - r.message for r in caplog.records if r.levelno >= logging.INFO - ] + warnings = [r.message for r in caplog.records if r.levelno >= logging.INFO] # Exactly one warning, and it should mention scrape_webpage but not # any middleware-provided name. Inspect the rendered "missing" # list (between the brackets) so we don't false-match substrings diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_tool_call_repair.py b/surfsense_backend/tests/unit/agents/new_chat/test_tool_call_repair.py index f792aef60..e02a04774 100644 --- a/surfsense_backend/tests/unit/agents/new_chat/test_tool_call_repair.py +++ b/surfsense_backend/tests/unit/agents/new_chat/test_tool_call_repair.py @@ -27,9 +27,12 @@ class TestRepair: mw = ToolCallNameRepairMiddleware( registered_tool_names={"echo"}, fuzzy_match_threshold=None ) - msg = AIMessage(content="", tool_calls=[ - {"name": "echo", "args": {}, "id": "1"}, - ]) + msg = AIMessage( + content="", + tool_calls=[ + {"name": "echo", "args": {}, "id": "1"}, + ], + ) out = mw.after_model(_make_state(msg), _FakeRuntime()) assert out is None # no change @@ -37,9 +40,12 @@ class TestRepair: mw = ToolCallNameRepairMiddleware( registered_tool_names={"echo"}, fuzzy_match_threshold=None ) - msg = AIMessage(content="", tool_calls=[ - {"name": "Echo", "args": {"x": 1}, "id": "1"}, - ]) + msg = AIMessage( + content="", + tool_calls=[ + {"name": "Echo", "args": {"x": 1}, "id": "1"}, + ], + ) out = mw.after_model(_make_state(msg), _FakeRuntime()) assert out is not None repaired = out["messages"][0] @@ -50,9 +56,12 @@ class TestRepair: registered_tool_names={"echo", INVALID_TOOL_NAME}, fuzzy_match_threshold=None, ) - msg = AIMessage(content="", tool_calls=[ - {"name": "totally_different_name", "args": {"k": "v"}, "id": "1"}, - ]) + msg = AIMessage( + content="", + tool_calls=[ + {"name": "totally_different_name", "args": {"k": "v"}, "id": "1"}, + ], + ) out = mw.after_model(_make_state(msg), _FakeRuntime()) assert out is not None repaired_call = out["messages"][0].tool_calls[0] @@ -64,9 +73,12 @@ class TestRepair: mw = ToolCallNameRepairMiddleware( registered_tool_names={"echo"}, fuzzy_match_threshold=None ) - msg = AIMessage(content="", tool_calls=[ - {"name": "unknown", "args": {}, "id": "1"}, - ]) + msg = AIMessage( + content="", + tool_calls=[ + {"name": "unknown", "args": {}, "id": "1"}, + ], + ) out = mw.after_model(_make_state(msg), _FakeRuntime()) # No repair available; original returned unchanged (no update) assert out is None @@ -76,9 +88,12 @@ class TestRepair: registered_tool_names={"search_documents"}, fuzzy_match_threshold=0.7, ) - msg = AIMessage(content="", tool_calls=[ - {"name": "search_docments", "args": {}, "id": "1"}, - ]) + msg = AIMessage( + content="", + tool_calls=[ + {"name": "search_docments", "args": {}, "id": "1"}, + ], + ) out = mw.after_model(_make_state(msg), _FakeRuntime()) assert out is not None assert out["messages"][0].tool_calls[0]["name"] == "search_documents" @@ -94,9 +109,12 @@ class TestRepair: mw = ToolCallNameRepairMiddleware( registered_tool_names={"echo"}, fuzzy_match_threshold=None ) - msg = AIMessage(content="", tool_calls=[ - {"name": "DynamicTool", "args": {}, "id": "1"}, - ]) + msg = AIMessage( + content="", + tool_calls=[ + {"name": "DynamicTool", "args": {}, "id": "1"}, + ], + ) runtime = _FakeRuntime(SimpleNamespace(registered_tool_names=["dynamictool"])) out = mw.after_model(_make_state(msg), runtime) assert out is not None diff --git a/surfsense_backend/tests/unit/middleware/test_kb_persistence_filesystem_parity.py b/surfsense_backend/tests/unit/middleware/test_kb_persistence_filesystem_parity.py index 8b464d48d..ef95434bf 100644 --- a/surfsense_backend/tests/unit/middleware/test_kb_persistence_filesystem_parity.py +++ b/surfsense_backend/tests/unit/middleware/test_kb_persistence_filesystem_parity.py @@ -10,7 +10,7 @@ through :class:`KnowledgeBasePersistenceMiddleware` without losing the copy. from __future__ import annotations from typing import Any -from unittest.mock import AsyncMock, MagicMock +from unittest.mock import AsyncMock import numpy as np import pytest diff --git a/surfsense_backend/tests/unit/services/test_revert_service.py b/surfsense_backend/tests/unit/services/test_revert_service.py index cb8443291..e2cbe383a 100644 --- a/surfsense_backend/tests/unit/services/test_revert_service.py +++ b/surfsense_backend/tests/unit/services/test_revert_service.py @@ -16,9 +16,7 @@ class _FakeAction: class TestCanRevert: def test_owner_can_revert_their_own_action(self) -> None: action = _FakeAction(user_id="user-123") - assert can_revert( - requester_user_id="user-123", action=action, is_admin=False - ) + assert can_revert(requester_user_id="user-123", action=action, is_admin=False) def test_other_user_cannot_revert(self) -> None: action = _FakeAction(user_id="user-123") @@ -28,21 +26,15 @@ class TestCanRevert: def test_admin_always_allowed(self) -> None: action = _FakeAction(user_id="user-123") - assert can_revert( - requester_user_id="anybody", action=action, is_admin=True - ) + assert can_revert(requester_user_id="anybody", action=action, is_admin=True) def test_admin_can_revert_anonymous_action(self) -> None: action = _FakeAction(user_id=None) - assert can_revert( - requester_user_id="admin", action=action, is_admin=True - ) + assert can_revert(requester_user_id="admin", action=action, is_admin=True) def test_anonymous_action_blocks_non_admin(self) -> None: action = _FakeAction(user_id=None) - assert not can_revert( - requester_user_id="user-1", action=action, is_admin=False - ) + assert not can_revert(requester_user_id="user-1", action=action, is_admin=False) def test_uuid_string_normalization(self) -> None: """``user_id`` may be a UUID object; comparison should still work.""" @@ -51,6 +43,4 @@ class TestCanRevert: u = uuid.uuid4() action = _FakeAction(user_id=u) # Same UUID, passed as string from the requesting side. - assert can_revert( - requester_user_id=str(u), action=action, is_admin=False - ) + assert can_revert(requester_user_id=str(u), action=action, is_admin=False) From f23be16b351da11130c01d6157cc0fc817cb51b9 Mon Sep 17 00:00:00 2001 From: "DESKTOP-RTLN3BA\\$punk" Date: Tue, 28 Apr 2026 23:25:26 -0700 Subject: [PATCH 16/17] refactor: citation viewer --- surfsense_web/app/globals.css | 21 - .../atoms/citation/citation-panel.atom.ts | 40 ++ .../pending-chunk-highlight.atom.ts | 19 - .../atoms/layout/right-panel.atom.ts | 2 +- .../assistant-ui/inline-citation.tsx | 80 +--- .../citation-panel/citation-panel.tsx | 230 ++++++++++ .../components/editor-panel/editor-panel.tsx | 407 +----------------- .../components/editor/plate-editor.tsx | 30 +- surfsense_web/components/editor/presets.ts | 28 -- .../layout/ui/right-panel/RightPanel.tsx | 72 +++- .../components/ui/search-highlight-node.tsx | 45 -- surfsense_web/lib/citation-search.ts | 125 ------ surfsense_web/package.json | 1 - surfsense_web/pnpm-lock.yaml | 17 - 14 files changed, 362 insertions(+), 755 deletions(-) create mode 100644 surfsense_web/atoms/citation/citation-panel.atom.ts delete mode 100644 surfsense_web/atoms/document-viewer/pending-chunk-highlight.atom.ts create mode 100644 surfsense_web/components/citation-panel/citation-panel.tsx delete mode 100644 surfsense_web/components/ui/search-highlight-node.tsx delete mode 100644 surfsense_web/lib/citation-search.ts diff --git a/surfsense_web/app/globals.css b/surfsense_web/app/globals.css index f54bc2197..a37ddb8f3 100644 --- a/surfsense_web/app/globals.css +++ b/surfsense_web/app/globals.css @@ -210,27 +210,6 @@ button { } } -/* Citation-jump highlight — entrance pulse only. The `SearchHighlightLeaf` - (see components/ui/search-highlight-node.tsx) is otherwise statically - tinted; this animation runs once on mount to draw the eye to the cited - text after `scrollIntoView` lands. The highlight itself is permanent - until the user clicks inside the editor (or another dismissal trigger - fires in `EditorPanelContent`). */ -@keyframes citation-flash-in { - 0% { - background-color: transparent; - box-shadow: 0 0 0 0 transparent; - } - 40% { - background-color: color-mix(in oklab, var(--primary) 30%, transparent); - box-shadow: 0 0 0 3px color-mix(in oklab, var(--primary) 25%, transparent); - } - 100% { - background-color: color-mix(in oklab, var(--primary) 15%, transparent); - box-shadow: 0 0 0 1px color-mix(in oklab, var(--primary) 40%, transparent); - } -} - /* Human-in-the-loop approval card animations */ @keyframes pulse-subtle { 0%, diff --git a/surfsense_web/atoms/citation/citation-panel.atom.ts b/surfsense_web/atoms/citation/citation-panel.atom.ts new file mode 100644 index 000000000..ca7312857 --- /dev/null +++ b/surfsense_web/atoms/citation/citation-panel.atom.ts @@ -0,0 +1,40 @@ +import { atom } from "jotai"; +import { rightPanelCollapsedAtom, rightPanelTabAtom } from "@/atoms/layout/right-panel.atom"; + +interface CitationPanelState { + isOpen: boolean; + chunkId: number | null; +} + +const initialState: CitationPanelState = { + isOpen: false, + chunkId: null, +}; + +export const citationPanelAtom = atom(initialState); + +export const citationPanelOpenAtom = atom((get) => get(citationPanelAtom).isOpen); + +const preCitationCollapsedAtom = atom(null); + +export const openCitationPanelAtom = atom(null, (get, set, payload: { chunkId: number }) => { + if (!get(citationPanelAtom).isOpen) { + set(preCitationCollapsedAtom, get(rightPanelCollapsedAtom)); + } + set(citationPanelAtom, { + isOpen: true, + chunkId: payload.chunkId, + }); + set(rightPanelTabAtom, "citation"); + set(rightPanelCollapsedAtom, false); +}); + +export const closeCitationPanelAtom = atom(null, (get, set) => { + set(citationPanelAtom, initialState); + set(rightPanelTabAtom, "sources"); + const prev = get(preCitationCollapsedAtom); + if (prev !== null) { + set(rightPanelCollapsedAtom, prev); + set(preCitationCollapsedAtom, null); + } +}); diff --git a/surfsense_web/atoms/document-viewer/pending-chunk-highlight.atom.ts b/surfsense_web/atoms/document-viewer/pending-chunk-highlight.atom.ts deleted file mode 100644 index a3f8357e8..000000000 --- a/surfsense_web/atoms/document-viewer/pending-chunk-highlight.atom.ts +++ /dev/null @@ -1,19 +0,0 @@ -import { atom } from "jotai"; - -/** - * Cross-component handoff for citation jumps. Set by `InlineCitation` when a - * numeric chunk badge is clicked (after the document has been resolved); read - * by `DocumentTabContent` once the matching document tab mounts so it can - * scroll to and softly highlight the cited chunk inside the rendered markdown. - * - * Cleared by `DocumentTabContent` only after a terminal state — exact / - * approximate / miss — has been reached, so that an escalation refetch (2MB - * preview → 16MB) keeps the pending intent alive across the re-render. - */ -export interface PendingChunkHighlight { - documentId: number; - chunkId: number; - chunkText: string; -} - -export const pendingChunkHighlightAtom = atom(null); diff --git a/surfsense_web/atoms/layout/right-panel.atom.ts b/surfsense_web/atoms/layout/right-panel.atom.ts index e06500113..d296587ed 100644 --- a/surfsense_web/atoms/layout/right-panel.atom.ts +++ b/surfsense_web/atoms/layout/right-panel.atom.ts @@ -1,6 +1,6 @@ import { atom } from "jotai"; -export type RightPanelTab = "sources" | "report" | "editor" | "hitl-edit"; +export type RightPanelTab = "sources" | "report" | "editor" | "hitl-edit" | "citation"; export const rightPanelTabAtom = atom("sources"); diff --git a/surfsense_web/components/assistant-ui/inline-citation.tsx b/surfsense_web/components/assistant-ui/inline-citation.tsx index ae8d434a8..2aeba89ca 100644 --- a/surfsense_web/components/assistant-ui/inline-citation.tsx +++ b/surfsense_web/components/assistant-ui/inline-citation.tsx @@ -1,13 +1,11 @@ "use client"; -import { useQuery, useQueryClient } from "@tanstack/react-query"; +import { useQuery } from "@tanstack/react-query"; import { useSetAtom } from "jotai"; import { ExternalLink, FileText } from "lucide-react"; import type { FC } from "react"; import { useCallback, useEffect, useRef, useState } from "react"; -import { toast } from "sonner"; -import { pendingChunkHighlightAtom } from "@/atoms/document-viewer/pending-chunk-highlight.atom"; -import { openEditorPanelAtom } from "@/atoms/editor/editor-panel.atom"; +import { openCitationPanelAtom } from "@/atoms/citation/citation-panel.atom"; import { useCitationMetadata } from "@/components/assistant-ui/citation-metadata-context"; import { MarkdownViewer } from "@/components/markdown-viewer"; import { Citation } from "@/components/tool-ui/citation"; @@ -29,11 +27,11 @@ const POPOVER_HOVER_CLOSE_DELAY_MS = 150; * Surfsense documentation chunks (`isDocsChunk`). Negative chunk IDs render as * a static "doc" pill (anonymous/synthetic uploads). * - * Numeric KB chunks: clicking resolves the parent document via - * `getDocumentByChunk`, opens the document in the right side panel (alongside - * the chat — does not replace it), and stages the cited chunk text in - * `pendingChunkHighlightAtom` so `EditorPanelContent` can scroll to and softly - * highlight it inside the rendered markdown. + * Numeric KB chunks: clicking opens the citation panel in the right + * sidebar (alongside the chat — does not replace it). The panel shows + * the cited chunk surrounded by adjacent chunks (via the API's + * `chunk_window`), with the cited one highlighted and an option to + * expand the window or jump into the full document via the editor panel. * * Surfsense docs chunks: rendered as a hover-controlled shadcn Popover that * lazily fetches and previews the cited chunk inline, since those docs aren't @@ -65,71 +63,17 @@ export const InlineCitation: FC = ({ chunkId, isDocsChunk = }; const NumericChunkCitation: FC<{ chunkId: number }> = ({ chunkId }) => { - const queryClient = useQueryClient(); - const setPendingHighlight = useSetAtom(pendingChunkHighlightAtom); - const openEditorPanel = useSetAtom(openEditorPanelAtom); - const [resolving, setResolving] = useState(false); - - const handleClick = useCallback(async () => { - if (resolving) return; - setResolving(true); - console.log("[citation:click] start", { chunkId }); - try { - const data = await queryClient.fetchQuery({ - // Local key with explicit window. The shared `cacheKeys.documents.byChunk` - // is window-agnostic (latent footgun); namespace the call to avoid - // reusing a different-window cached result. - queryKey: ["documents", "by-chunk", chunkId, "w0"] as const, - queryFn: () => - documentsApiService.getDocumentByChunk({ chunk_id: chunkId, chunk_window: 0 }), - staleTime: 5 * 60 * 1000, - }); - const cited = data.chunks.find((c) => c.id === chunkId) ?? data.chunks[0]; - console.log("[citation:click] fetched doc-by-chunk", { - docId: data.id, - docTitle: data.title, - chunksReturned: data.chunks.length, - citedChunkId: cited?.id, - citedChunkContentLen: cited?.content?.length ?? 0, - citedChunkPreview: - cited?.content && cited.content.length > 120 - ? `${cited.content.slice(0, 120)}…(+${cited.content.length - 120})` - : (cited?.content ?? ""), - }); - // Stage the highlight BEFORE opening the panel so `EditorPanelContent` - // already sees the pending intent on its very first render — avoids a - // "fetch → render → no-pending → next-tick render with pending" race. - setPendingHighlight({ - documentId: data.id, - chunkId, - chunkText: cited?.content ?? "", - }); - openEditorPanel({ - documentId: data.id, - searchSpaceId: data.search_space_id, - title: data.title, - }); - console.log("[citation:click] staged highlight + opened editor panel", { - documentId: data.id, - }); - } catch (err) { - console.warn("[citation:click] failed", err); - toast.error(err instanceof Error ? err.message : "Couldn't open cited document"); - } finally { - setResolving(false); - } - }, [chunkId, openEditorPanel, queryClient, resolving, setPendingHighlight]); + const openCitationPanel = useSetAtom(openCitationPanelAtom); return ( ); }; diff --git a/surfsense_web/components/citation-panel/citation-panel.tsx b/surfsense_web/components/citation-panel/citation-panel.tsx new file mode 100644 index 000000000..cec07b9cf --- /dev/null +++ b/surfsense_web/components/citation-panel/citation-panel.tsx @@ -0,0 +1,230 @@ +"use client"; + +import { useQuery } from "@tanstack/react-query"; +import { useSetAtom } from "jotai"; +import { ChevronDown, ChevronUp, ExternalLink, XIcon } from "lucide-react"; +import type { FC } from "react"; +import { useEffect, useMemo, useRef, useState } from "react"; +import { openEditorPanelAtom } from "@/atoms/editor/editor-panel.atom"; +import { MarkdownViewer } from "@/components/markdown-viewer"; +import { Button } from "@/components/ui/button"; +import { Spinner } from "@/components/ui/spinner"; +import { documentsApiService } from "@/lib/apis/documents-api.service"; + +const DEFAULT_CHUNK_WINDOW = 5; +const EXPANDED_CHUNK_WINDOW = 50; + +interface CitationPanelContentProps { + chunkId: number; + onClose?: () => void; +} + +/** + * Right-panel citation viewer. Shows the cited chunk surrounded by + * adjacent chunks (±N chunks via the API's `chunk_window` parameter), + * with the cited one visually highlighted and auto-scrolled into view. + * The window can be expanded to a wider range, or the user can jump to + * the full document via the editor panel. + */ +export const CitationPanelContent: FC = ({ chunkId, onClose }) => { + const openEditorPanel = useSetAtom(openEditorPanelAtom); + const [expanded, setExpanded] = useState(false); + + useEffect(() => { + setExpanded(false); + }, []); + + const chunkWindow = expanded ? EXPANDED_CHUNK_WINDOW : DEFAULT_CHUNK_WINDOW; + + const { data, isLoading, error } = useQuery({ + queryKey: ["citation-panel", chunkId, chunkWindow] as const, + queryFn: () => + documentsApiService.getDocumentByChunk({ + chunk_id: chunkId, + chunk_window: chunkWindow, + }), + staleTime: 5 * 60 * 1000, + }); + + const cited = useMemo(() => data?.chunks.find((c) => c.id === chunkId) ?? null, [data, chunkId]); + + const totalChunks = data?.total_chunks ?? data?.chunks.length ?? 0; + const startIndex = data?.chunk_start_index ?? 0; + const citedIndexInWindow = data + ? Math.max( + 0, + data.chunks.findIndex((c) => c.id === chunkId) + ) + : 0; + const shownAbove = citedIndexInWindow; + const shownBelow = data ? Math.max(0, data.chunks.length - 1 - citedIndexInWindow) : 0; + const hasMoreAbove = startIndex > 0; + const hasMoreBelow = data ? startIndex + data.chunks.length < totalChunks : false; + + // Scroll the cited chunk into view inside the panel's scroll container + // (not the page). We anchor the scroll to the panel's scroll element + // so opening the citation doesn't yank the chat scroll on the left. + const scrollContainerRef = useRef(null); + const citedRef = useRef(null); + useEffect(() => { + if (!cited) return; + const id = requestAnimationFrame(() => { + const container = scrollContainerRef.current; + const target = citedRef.current; + if (!container || !target) return; + const containerRect = container.getBoundingClientRect(); + const targetRect = target.getBoundingClientRect(); + const offset = targetRect.top - containerRect.top + container.scrollTop; + container.scrollTo({ + top: Math.max(0, offset - 16), + behavior: "smooth", + }); + }); + return () => cancelAnimationFrame(id); + }, [cited]); + + const handleOpenFullDocument = () => { + if (!data) return; + openEditorPanel({ + documentId: data.id, + searchSpaceId: data.search_space_id, + title: data.title, + }); + }; + + return ( + <> +
+
+

Citation

+
+ {onClose && ( + + )} +
+
+
+
+

+ {data?.title ?? (isLoading ? "Loading…" : `Chunk #${chunkId}`)} +

+
+
+ Chunk #{chunkId} + {totalChunks > 0 && · {totalChunks} chunks} +
+
+
+ +
+ {isLoading && ( +
+ + Loading citation… +
+ )} + + {error && ( +

+ {error instanceof Error ? error.message : "Failed to load citation"} +

+ )} + + {!isLoading && !error && data && ( + <> + {hasMoreAbove && ( +

+ … {startIndex} earlier chunk{startIndex === 1 ? "" : "s"} not shown +

+ )} +
+ {data.chunks.map((chunk) => { + const isCited = chunk.id === chunkId; + return ( +
+
+ + {isCited ? "Cited chunk" : `Chunk #${chunk.id}`} + + {isCited && ( + #{chunk.id} + )} +
+
+ +
+
+ ); + })} +
+ {hasMoreBelow && ( +

+ … {totalChunks - (startIndex + data.chunks.length)} later chunk + {totalChunks - (startIndex + data.chunks.length) === 1 ? "" : "s"} not shown +

+ )} + + )} +
+ + {!isLoading && !error && data && ( +
+
+ Showing {shownAbove} above · cited · {shownBelow} below +
+
+ {(hasMoreAbove || hasMoreBelow) && !expanded && ( + + )} + {expanded && ( + + )} + +
+
+ )} + + ); +}; diff --git a/surfsense_web/components/editor-panel/editor-panel.tsx b/surfsense_web/components/editor-panel/editor-panel.tsx index 0c4e9485b..df138e97e 100644 --- a/surfsense_web/components/editor-panel/editor-panel.tsx +++ b/surfsense_web/components/editor-panel/editor-panel.tsx @@ -1,6 +1,5 @@ "use client"; -import { FindReplacePlugin } from "@platejs/find-replace"; import { useAtomValue, useSetAtom } from "jotai"; import { Check, @@ -15,21 +14,17 @@ import { import dynamic from "next/dynamic"; import { useCallback, useEffect, useRef, useState } from "react"; import { toast } from "sonner"; -import { pendingChunkHighlightAtom } from "@/atoms/document-viewer/pending-chunk-highlight.atom"; import { closeEditorPanelAtom, editorPanelAtom } from "@/atoms/editor/editor-panel.atom"; import { VersionHistoryButton } from "@/components/documents/version-history"; -import type { PlateEditorInstance } from "@/components/editor/plate-editor"; import { SourceCodeEditor } from "@/components/editor/source-code-editor"; import { MarkdownViewer } from "@/components/markdown-viewer"; import { Alert, AlertDescription } from "@/components/ui/alert"; import { Button } from "@/components/ui/button"; import { Drawer, DrawerContent, DrawerHandle, DrawerTitle } from "@/components/ui/drawer"; -import { CITATION_HIGHLIGHT_CLASS } from "@/components/ui/search-highlight-node"; import { Spinner } from "@/components/ui/spinner"; import { useMediaQuery } from "@/hooks/use-media-query"; import { useElectronAPI } from "@/hooks/use-platform"; import { authenticatedFetch, getBearerToken, redirectToLogin } from "@/lib/auth-utils"; -import { buildCitationSearchCandidates } from "@/lib/citation-search"; import { inferMonacoLanguageFromPath } from "@/lib/editor-language"; const PlateEditor = dynamic( @@ -37,10 +32,7 @@ const PlateEditor = dynamic( { ssr: false, loading: () => } ); -type CitationHighlightStatus = "exact" | "miss"; - const LARGE_DOCUMENT_THRESHOLD = 2 * 1024 * 1024; // 2MB -const CITATION_MAX_LENGTH = 16 * 1024 * 1024; // 16MB on-demand cap for citation jumps interface EditorContent { document_id: number; @@ -145,60 +137,6 @@ export function EditorPanelContent({ const isLocalFileMode = kind === "local_file"; const editorRenderMode: EditorRenderMode = isLocalFileMode ? "source_code" : "rich_markdown"; - // --- Citation-jump highlight wiring ---------------------------------- - // `EditorPanelContent` is the consumer of `pendingChunkHighlightAtom`: when - // a citation badge is clicked, the badge stages `{documentId, chunkId, - // chunkText}` and opens this panel. We drive Plate's `FindReplacePlugin` - // (registered in every preset) to highlight the cited text natively via - // Slate decorations — no DOM walking, no Range gymnastics. The state - // machine below escalates the document fetch from 2MB → 16MB once if no - // candidate snippet matched in the preview, and surfaces miss outcomes - // via an inline alert. - const pending = useAtomValue(pendingChunkHighlightAtom); - const setPendingHighlight = useSetAtom(pendingChunkHighlightAtom); - const [fetchKey, setFetchKey] = useState(0); - const [maxLengthOverride, setMaxLengthOverride] = useState(null); - const [highlightResult, setHighlightResult] = useState(null); - const editorRef = useRef(null); - const escalatedForRef = useRef(null); - const lastAppliedChunkIdRef = useRef(null); - // Tracks whether a citation highlight is currently decorated in the - // editor. We use a ref (not state) because the click-to-dismiss handler - // runs in a stable callback that would otherwise close over stale state. - const isHighlightActiveRef = useRef(false); - // Once a citation jump targets this doc we have to keep `PlateEditor` - // mounted for the *rest of the doc session* — even after the highlight - // effect clears `pendingChunkHighlightAtom` (which it does as soon as - // the decoration is applied, so a follow-up citation on the same chunk - // can re-trigger). Without this latch, non-editable docs would re-render - // back into `MarkdownViewer` the instant `pending` is released, tearing - // down the Plate decorations and dropping the highlight after a frame. - const [stickyPlateMode, setStickyPlateMode] = useState(false); - - const clearCitationSearch = useCallback(() => { - isHighlightActiveRef.current = false; - const editor = editorRef.current; - if (!editor) return; - try { - editor.setOption(FindReplacePlugin, "search", ""); - editor.api.redecorate(); - } catch (err) { - console.warn("[EditorPanelContent] clearCitationSearch failed:", err); - } - }, []); - - // Dismiss the highlight when the user interacts with the editor surface. - // `onPointerDown` fires before focus / selection changes so the click - // itself feels responsive — the highlight clears in the same event tick - // that places the cursor. No-op when nothing is highlighted, so we don't - // thrash `redecorate` on every click in normal editing. - const handleEditorPointerDown = useCallback(() => { - if (!isHighlightActiveRef.current) return; - clearCitationSearch(); - setHighlightResult(null); - }, [clearCitationSearch]); - - const isCitationTarget = !!pending && !isLocalFileMode && pending.documentId === documentId; const resolveLocalVirtualPath = useCallback( async (candidatePath: string): Promise => { if (!electronAPI?.getAgentFilesystemMounts) { @@ -218,8 +156,6 @@ export function EditorPanelContent({ const isLargeDocument = (editorDoc?.content_size_bytes ?? 0) > LARGE_DOCUMENT_THRESHOLD; - // `fetchKey` is an explicit re-fetch trigger (escalation bumps it to force - // a new request even when documentId/searchSpaceId haven't changed). useEffect(() => { const controller = new AbortController(); setIsLoading(true); @@ -231,12 +167,6 @@ export function EditorPanelContent({ setIsEditing(false); initialLoadDone.current = false; changeCountRef.current = 0; - // Clear any in-flight FindReplacePlugin search before the editor - // re-mounts on new content (a fresh editor key is generated below - // from documentId + isEditing, so the previous editor + its - // decorations are about to be discarded anyway, but we belt-and- - // brace here for the case where only `fetchKey` changed). - clearCitationSearch(); const doFetch = async () => { try { @@ -281,11 +211,7 @@ export function EditorPanelContent({ const url = new URL( `${process.env.NEXT_PUBLIC_FASTAPI_BACKEND_URL}/api/v1/search-spaces/${searchSpaceId}/documents/${documentId}/editor-content` ); - url.searchParams.set("max_length", String(maxLengthOverride ?? LARGE_DOCUMENT_THRESHOLD)); - // `fetchKey` participates here so biome's noUnusedVariables sees it - // as consumed; bumping it forces a fresh request even when the URL - // is otherwise identical. - if (fetchKey > 0) url.searchParams.set("_n", String(fetchKey)); + url.searchParams.set("max_length", String(LARGE_DOCUMENT_THRESHOLD)); const response = await authenticatedFetch(url.toString(), { method: "GET" }); @@ -331,259 +257,8 @@ export function EditorPanelContent({ resolveLocalVirtualPath, searchSpaceId, title, - fetchKey, - maxLengthOverride, - clearCitationSearch, ]); - // Reset citation-jump bookkeeping whenever the panel switches to a different - // document (or local file). Body only writes setters — the deps are the - // real triggers we want to react to. - // biome-ignore lint/correctness/useExhaustiveDependencies: documentId/localFilePath are intentional triggers. - useEffect(() => { - clearCitationSearch(); - escalatedForRef.current = null; - lastAppliedChunkIdRef.current = null; - setHighlightResult(null); - setMaxLengthOverride(null); - setFetchKey(0); - // Drop sticky Plate mode when the panel moves to a different doc - // — the next doc starts in its preferred render mode (Plate for - // editable, MarkdownViewer for everything else) until/unless a - // citation jump targets it. - setStickyPlateMode(false); - }, [documentId, localFilePath, clearCitationSearch]); - - // Latch sticky Plate mode the first time a citation jump targets this - // doc. We keep it sticky for the remainder of this doc session so the - // highlight effect's `setPendingHighlight(null)` doesn't unmount the - // editor mid-flight (see comment on `stickyPlateMode` declaration). - useEffect(() => { - if (isCitationTarget) setStickyPlateMode(true); - }, [isCitationTarget]); - - // `isEditorReady` is what `useEffect` actually depends on — `editorRef` - // is a ref so changes don't trigger re-runs. We flip this to `true` once - // `PlateEditor` calls back with its live editor instance (its - // `usePlateEditor` value-init runs synchronously, so by the time this - // flips true the markdown is already deserialized into the Slate tree). - const [isEditorReady, setIsEditorReady] = useState(false); - const handleEditorReady = useCallback((editor: PlateEditorInstance | null) => { - console.log("[citation:editor] handleEditorReady", { ready: !!editor }); - editorRef.current = editor; - setIsEditorReady(!!editor); - }, []); - - // --- Citation jump highlight effect ----------------------------------- - // Drives Plate's FindReplacePlugin to highlight the cited chunk: - // 1. Build candidate snippets from the chunk text (first sentence, - // first 8 words, full chunk if short). Plate's decorate runs per- - // block and won't cross block boundaries, so the shorter - // candidates exist to give us something that fits in one - // paragraph / heading. - // 2. For each candidate: setOption('search', ...) → redecorate → - // wait two animation frames for React to flush → query the editor - // DOM for `.${CITATION_HIGHLIGHT_CLASS}`. First hit wins. - // - // Why a className and not a `data-*` attribute? Plate's - // `PlateLeaf` runs its props through `useNodeAttributes`, which - // only forwards `attributes`, `className`, `ref`, and `style` — - // arbitrary `data-*` attributes are silently dropped. `className` - // is the only escape hatch guaranteed to survive into the DOM. - // 3. On hit: smooth-scroll the first match into view, mark the - // highlight active (so a click inside the editor can dismiss it), - // release the pending atom. - // 4. On terminal miss: if the doc was truncated and we haven't - // escalated yet, bump the fetch's `max_length` to the citation - // cap and re-fetch — the post-refetch render will re-run this - // effect against the larger preview. Otherwise, release the - // atom and show the miss alert. - useEffect(() => { - console.log("[citation:effect] fired", { - isCitationTarget, - pendingDocId: pending?.documentId, - pendingChunkId: pending?.chunkId, - pendingChunkTextLen: pending?.chunkText?.length, - documentId, - isLocalFileMode, - isEditing, - hasMarkdown: !!editorDoc?.source_markdown, - markdownLen: editorDoc?.source_markdown?.length, - truncated: editorDoc?.truncated, - isEditorReady, - editorRefSet: !!editorRef.current, - maxLengthOverride, - }); - if (!isCitationTarget || !pending) { - console.log("[citation:effect] guard ✗ no citation target / no pending"); - return; - } - if (isLocalFileMode || isEditing) { - console.log("[citation:effect] guard ✗ localFileMode/editing"); - return; - } - if (!editorDoc?.source_markdown) { - console.log("[citation:effect] guard ✗ source_markdown not ready"); - return; - } - if (!isEditorReady) { - console.log("[citation:effect] guard ✗ editor not ready yet"); - return; - } - const editor = editorRef.current; - if (!editor) { - console.log("[citation:effect] guard ✗ editorRef.current is null"); - return; - } - - if (lastAppliedChunkIdRef.current !== pending.chunkId) { - lastAppliedChunkIdRef.current = pending.chunkId; - } - - let cancelled = false; - - const finishMiss = () => { - console.log("[citation:effect] terminal miss — no candidate matched"); - try { - editor.setOption(FindReplacePlugin, "search", ""); - editor.api.redecorate(); - } catch (err) { - console.warn("[EditorPanelContent] reset search after miss failed:", err); - } - const canEscalate = - editorDoc.truncated === true && - (maxLengthOverride ?? LARGE_DOCUMENT_THRESHOLD) < CITATION_MAX_LENGTH && - escalatedForRef.current !== pending.chunkId; - console.log("[citation:effect] miss decision", { - truncated: editorDoc.truncated, - currentMaxLength: maxLengthOverride ?? LARGE_DOCUMENT_THRESHOLD, - canEscalate, - }); - if (canEscalate) { - escalatedForRef.current = pending.chunkId; - setMaxLengthOverride(CITATION_MAX_LENGTH); - setFetchKey((k) => k + 1); - // Keep the atom set so the post-refetch render re-runs. - return; - } - setHighlightResult("miss"); - setPendingHighlight(null); - }; - - const tryCandidates = async () => { - const candidates = buildCitationSearchCandidates(pending.chunkText); - console.log("[citation:effect] candidates built", { - count: candidates.length, - previews: candidates.map((c) => c.slice(0, 60)), - }); - if (candidates.length === 0) { - if (!cancelled) finishMiss(); - return; - } - // Resolve the editor's rendered DOM root via Slate's stable - // `[data-slate-editor="true"]` attribute (set by slate-react's - // ``). Scoping queries to this root prevents - // `` elements rendered elsewhere on the page (e.g. chat - // search-highlight leaves in another mounted PlateEditor) from - // being mistaken for citation hits. - const editorRoot = document.querySelector('[data-slate-editor="true"]'); - console.log("[citation:effect] editor root", { - hasRoot: !!editorRoot, - }); - const root: ParentNode = editorRoot ?? document; - - for (let i = 0; i < candidates.length; i++) { - const candidate = candidates[i]; - if (cancelled) return; - try { - editor.setOption(FindReplacePlugin, "search", candidate); - editor.api.redecorate(); - console.log(`[citation:effect] try #${i} setOption + redecorate`, { - len: candidate.length, - preview: candidate.slice(0, 80), - }); - } catch (err) { - console.warn("[EditorPanelContent] setOption/redecorate failed:", err); - continue; - } - // Two rAFs: first lets Slate flush its onChange, second lets - // React commit the decoration leaves into the DOM. - await new Promise((resolve) => - requestAnimationFrame(() => requestAnimationFrame(() => resolve())) - ); - if (cancelled) return; - // Primary probe: by our stable class on the rendered . - let el = root.querySelector(`.${CITATION_HIGHLIGHT_CLASS}`); - const classMarkCount = root.querySelectorAll(`.${CITATION_HIGHLIGHT_CLASS}`).length; - // Diagnostic fallback: any inside the editor root. - // If we ever see allMarks > 0 but classMarkCount === 0, - // the className was stripped again and we need to revisit - // `useNodeAttributes` filtering. - const allMarkCount = root.querySelectorAll("mark").length; - if (!el && allMarkCount > 0) { - el = root.querySelector("mark"); - } - console.log(`[citation:effect] try #${i} DOM probe`, { - foundEl: !!el, - classMarkCount, - allMarkCount, - usedFallback: !!el && classMarkCount === 0, - }); - if (el) { - try { - el.scrollIntoView({ block: "center", behavior: "smooth" }); - } catch { - el.scrollIntoView(); - } - isHighlightActiveRef.current = true; - setHighlightResult("exact"); - console.log(`[citation:effect] ✓ exact via candidate #${i} — atom released`); - // No auto-clear timer — the highlight is intentionally - // permanent until the user clicks inside the editor (see - // `handleEditorPointerDown`) or another dismissal trigger - // fires (doc switch, edit-mode toggle, panel unmount, - // next citation jump). Sticky Plate mode keeps the - // editor mounted after the atom clears. - setPendingHighlight(null); - return; - } - } - if (!cancelled) finishMiss(); - }; - - void tryCandidates(); - - return () => { - cancelled = true; - }; - }, [ - isCitationTarget, - pending, - documentId, - editorDoc?.source_markdown, - editorDoc?.truncated, - isLocalFileMode, - isEditing, - isEditorReady, - maxLengthOverride, - clearCitationSearch, - setPendingHighlight, - ]); - - // Cleanup any active highlight on unmount. - useEffect(() => { - return () => clearCitationSearch(); - }, [clearCitationSearch]); - - // Toggling into edit mode swaps Plate out of readOnly. Clear the citation - // search so stale leaves don't linger in the editing surface. - useEffect(() => { - if (isEditing) { - clearCitationSearch(); - setHighlightResult(null); - } - }, [isEditing, clearCitationSearch]); - useEffect(() => { return () => { if (copyResetTimeoutRef.current) { @@ -617,7 +292,7 @@ export function EditorPanelContent({ }, [editorDoc?.source_markdown]); const handleSave = useCallback( - async (_options?: { silent?: boolean }) => { + async (options?: { silent?: boolean }) => { setSaving(true); try { if (isLocalFileMode) { @@ -668,11 +343,15 @@ export function EditorPanelContent({ setEditorDoc((prev) => (prev ? { ...prev, source_markdown: markdownRef.current } : prev)); setEditedMarkdown(null); - toast.success("Document saved! Reindexing in background..."); + if (!options?.silent) { + toast.success("Document saved! Reindexing in background..."); + } return true; } catch (err) { console.error("Error saving document:", err); - toast.error(err instanceof Error ? err.message : "Failed to save document"); + if (!options?.silent) { + toast.error(err instanceof Error ? err.message : "Failed to save document"); + } return false; } finally { setSaving(false); @@ -693,15 +372,11 @@ export function EditorPanelContent({ EDITABLE_DOCUMENT_TYPES.has(editorDoc.document_type ?? "")) && !isLargeDocument : false; - // Use PlateEditor for any of: - // - Editable doc types (FILE/NOTE) — existing editing UX. - // - Active citation jump in flight (`isCitationTarget`) — covers the - // mount in the very first render where the atom is set but the - // sticky effect hasn't fired yet. - // - Sticky Plate mode latched on a previous citation jump — keeps - // the editor mounted (with its decorations) after the highlight - // effect clears the atom. Resets when the doc changes. - const renderInPlateEditor = isEditableType || isCitationTarget || stickyPlateMode; + // Render through PlateEditor for editable doc types (FILE/NOTE). + // Everything else (large docs, non-editable types) falls back to the + // lightweight `MarkdownViewer` — Plate is heavy on multi-MB docs and + // non-editable types don't benefit from its editing UX. + const renderInPlateEditor = isEditableType; const hasUnsavedChanges = editedMarkdown !== null; const showDesktopHeader = !!onClose; const showEditingActions = isEditableType && isEditing; @@ -744,36 +419,6 @@ export function EditorPanelContent({ } }, [documentId, editorDoc?.title, searchSpaceId]); - // We no longer surface an "approximate" status — Plate's FindReplacePlugin - // either decorates an exact match or it doesn't, and the candidate snippet - // strategy (first sentence → first 8 words → full chunk) means we either - // land on the citation start or fall through to the miss alert. - const showMissAlert = isCitationTarget && highlightResult === "miss"; - - const citationAlerts = showMissAlert && ( - - - - Cited section couldn't be located in this view. - {editorDoc?.truncated && ( - - )} - - - ); - const largeDocAlert = isLargeDocument && !isLocalFileMode && editorDoc && ( @@ -1002,30 +647,17 @@ export function EditorPanelContent({ }} />
- ) : isLargeDocument && !isLocalFileMode && !isCitationTarget ? ( - // Large doc, no active citation — fast Streamdown preview - // + download CTA. We only fall back to MarkdownViewer here - // because Plate is heavy on multi-MB docs and the user - // isn't waiting on a specific citation to render. + ) : isLargeDocument && !isLocalFileMode ? ( + // Large doc — fast Streamdown preview + download CTA. + // Plate is heavy on multi-MB docs.
{largeDocAlert}
) : renderInPlateEditor ? ( - // Editable doc (FILE/NOTE) OR active citation jump (any - // doc type). The citation path uses Plate's - // FindReplacePlugin for native, decoration-based - // highlighting — see the citation-jump highlight effect - // above for how `editorRef` and `handleEditorReady` are - // wired. + // Editable doc (FILE/NOTE) — Plate editing UX.
- {(citationAlerts || (isLargeDocument && isCitationTarget && !isLocalFileMode)) && ( -
- {isLargeDocument && isCitationTarget && largeDocAlert} - {citationAlerts} -
- )} -
+
diff --git a/surfsense_web/components/editor/plate-editor.tsx b/surfsense_web/components/editor/plate-editor.tsx index eef18ef6a..7f12d3cae 100644 --- a/surfsense_web/components/editor/plate-editor.tsx +++ b/surfsense_web/components/editor/plate-editor.tsx @@ -12,10 +12,7 @@ import { type EditorPreset, presetMap } from "@/components/editor/presets"; import { escapeMdxExpressions } from "@/components/editor/utils/escape-mdx"; import { Editor, EditorContainer } from "@/components/ui/editor"; -/** Live editor instance returned by `usePlateEditor`. Exposed via the - * `onEditorReady` prop so callers (e.g. `EditorPanelContent`) can drive - * plugin options imperatively — most notably setting - * `FindReplacePlugin`'s `search` option for citation-jump highlights. */ +/** Live editor instance returned by `usePlateEditor`. */ export type PlateEditorInstance = ReturnType; export interface PlateEditorProps { @@ -68,15 +65,6 @@ export interface PlateEditorProps { * without modifying the core editor component. */ extraPlugins?: AnyPluginConfig[]; - /** - * Called whenever the live editor instance (re)mounts, with `null` on - * unmount. Used by callers that need to drive plugin options imperatively - * — e.g. `EditorPanelContent` setting `FindReplacePlugin`'s `search` - * option for citation-jump highlights. The callback is invoked exactly - * once per editor lifetime (the parent's `key` prop forces a fresh - * editor when needed, e.g. on edit-mode toggle). - */ - onEditorReady?: (editor: PlateEditorInstance | null) => void; } function PlateEditorContent({ @@ -115,7 +103,6 @@ export function PlateEditor({ defaultEditing = false, preset = "full", extraPlugins = [], - onEditorReady, }: PlateEditorProps) { const lastMarkdownRef = useRef(markdown); const lastHtmlRef = useRef(html); @@ -172,21 +159,6 @@ export function PlateEditor({ : undefined, }); - // Expose the live editor instance to imperative callers (e.g. citation - // jump highlights). We deliberately don't depend on `onEditorReady` - // itself in the cleanup closure — callers commonly pass an arrow that - // closes over a stable ref setter, but if they pass a freshly-bound - // callback per render, the `onEditorReady?.(editor)` re-fires which is - // idempotent for ref-style setters. - const onEditorReadyRef = useRef(onEditorReady); - useEffect(() => { - onEditorReadyRef.current = onEditorReady; - }, [onEditorReady]); - useEffect(() => { - onEditorReadyRef.current?.(editor); - return () => onEditorReadyRef.current?.(null); - }, [editor]); - // Update editor content when html prop changes externally useEffect(() => { if (html !== undefined && html !== lastHtmlRef.current) { diff --git a/surfsense_web/components/editor/presets.ts b/surfsense_web/components/editor/presets.ts index 49f53ecf1..c207b5e56 100644 --- a/surfsense_web/components/editor/presets.ts +++ b/surfsense_web/components/editor/presets.ts @@ -1,6 +1,5 @@ "use client"; -import { FindReplacePlugin } from "@platejs/find-replace"; import type { AnyPluginConfig } from "platejs"; import { TrailingBlockPlugin } from "platejs"; @@ -18,30 +17,6 @@ import { SelectionKit } from "@/components/editor/plugins/selection-kit"; import { SlashCommandKit } from "@/components/editor/plugins/slash-command-kit"; import { TableKit } from "@/components/editor/plugins/table-kit"; import { ToggleKit } from "@/components/editor/plugins/toggle-kit"; -import { SearchHighlightLeaf } from "@/components/ui/search-highlight-node"; - -/** - * Citation-jump highlighter. Re-uses Plate's built-in `FindReplacePlugin` - * (decorate-only, no editing surface) to drive the "scroll-to-cited-text" - * UX in `EditorPanelContent`. We register it in every preset because: - * - Decorate is a no-op when `search` is empty (single getOptions() check - * per block), so cost is effectively zero for non-citation viewers. - * - Keeping it preset-agnostic means citations work whether the doc is - * opened in editable (`full`) or pure-viewer (`readonly`) modes. - * - * The parent component drives `setOption(FindReplacePlugin, 'search', ...)` - * + `editor.api.redecorate()` to trigger highlights, then queries the - * editor DOM for `.citation-highlight-leaf` to scroll the first match - * into view. (We can't use a `data-*` attribute here — Plate's - * `PlateLeaf` runs props through `useNodeAttributes`, which only forwards - * `attributes`, `className`, `ref`, `style`; arbitrary `data-*` props are - * silently dropped.) See `components/ui/search-highlight-node.tsx` for - * the leaf component and `CITATION_HIGHLIGHT_CLASS` constant. - */ -const CitationFindReplacePlugin = FindReplacePlugin.configure({ - options: { search: "" }, - render: { node: SearchHighlightLeaf }, -}); /** * Full preset – every plugin kit enabled. @@ -63,7 +38,6 @@ export const fullPreset: AnyPluginConfig[] = [ ...AutoformatKit, ...DndKit, TrailingBlockPlugin, - CitationFindReplacePlugin, ]; /** @@ -78,7 +52,6 @@ export const minimalPreset: AnyPluginConfig[] = [ ...LinkKit, ...AutoformatKit, TrailingBlockPlugin, - CitationFindReplacePlugin, ]; /** @@ -95,7 +68,6 @@ export const readonlyPreset: AnyPluginConfig[] = [ ...CalloutKit, ...ToggleKit, ...MathKit, - CitationFindReplacePlugin, ]; /** All available preset names */ diff --git a/surfsense_web/components/layout/ui/right-panel/RightPanel.tsx b/surfsense_web/components/layout/ui/right-panel/RightPanel.tsx index 04bae010c..3481eec28 100644 --- a/surfsense_web/components/layout/ui/right-panel/RightPanel.tsx +++ b/surfsense_web/components/layout/ui/right-panel/RightPanel.tsx @@ -6,6 +6,7 @@ import dynamic from "next/dynamic"; import { startTransition, useEffect } from "react"; import { closeHitlEditPanelAtom, hitlEditPanelAtom } from "@/atoms/chat/hitl-edit-panel.atom"; import { closeReportPanelAtom, reportPanelAtom } from "@/atoms/chat/report-panel.atom"; +import { citationPanelAtom, closeCitationPanelAtom } from "@/atoms/citation/citation-panel.atom"; import { documentsSidebarOpenAtom } from "@/atoms/documents/ui.atoms"; import { closeEditorPanelAtom, editorPanelAtom } from "@/atoms/editor/editor-panel.atom"; import { rightPanelCollapsedAtom, rightPanelTabAtom } from "@/atoms/layout/right-panel.atom"; @@ -21,6 +22,14 @@ const EditorPanelContent = dynamic( { ssr: false, loading: () => null } ); +const CitationPanelContent = dynamic( + () => + import("@/components/citation-panel/citation-panel").then((m) => ({ + default: m.CitationPanelContent, + })), + { ssr: false, loading: () => null } +); + const HitlEditPanelContent = dynamic( () => import("@/components/hitl-edit-panel/hitl-edit-panel").then((m) => ({ @@ -69,12 +78,14 @@ export function RightPanelExpandButton() { const reportState = useAtomValue(reportPanelAtom); const editorState = useAtomValue(editorPanelAtom); const hitlEditState = useAtomValue(hitlEditPanelAtom); + const citationState = useAtomValue(citationPanelAtom); const reportOpen = reportState.isOpen && !!reportState.reportId; const editorOpen = editorState.isOpen && (editorState.kind === "document" ? !!editorState.documentId : !!editorState.localFilePath); const hitlEditOpen = hitlEditState.isOpen && !!hitlEditState.onSave; - const hasContent = documentsOpen || reportOpen || editorOpen || hitlEditOpen; + const citationOpen = citationState.isOpen && citationState.chunkId != null; + const hasContent = documentsOpen || reportOpen || editorOpen || hitlEditOpen || citationOpen; if (!collapsed || !hasContent) return null; @@ -98,7 +109,13 @@ export function RightPanelExpandButton() { ); } -const PANEL_WIDTHS = { sources: 420, report: 640, editor: 640, "hitl-edit": 640 } as const; +const PANEL_WIDTHS = { + sources: 420, + report: 640, + editor: 640, + "hitl-edit": 640, + citation: 560, +} as const; export function RightPanel({ documentsPanel }: RightPanelProps) { const [activeTab] = useAtom(rightPanelTabAtom); @@ -108,6 +125,8 @@ export function RightPanel({ documentsPanel }: RightPanelProps) { const closeEditor = useSetAtom(closeEditorPanelAtom); const hitlEditState = useAtomValue(hitlEditPanelAtom); const closeHitlEdit = useSetAtom(closeHitlEditPanelAtom); + const citationState = useAtomValue(citationPanelAtom); + const closeCitation = useSetAtom(closeCitationPanelAtom); const [collapsed, setCollapsed] = useAtom(rightPanelCollapsedAtom); const documentsOpen = documentsPanel?.open ?? false; @@ -116,37 +135,59 @@ export function RightPanel({ documentsPanel }: RightPanelProps) { editorState.isOpen && (editorState.kind === "document" ? !!editorState.documentId : !!editorState.localFilePath); const hitlEditOpen = hitlEditState.isOpen && !!hitlEditState.onSave; + const citationOpen = citationState.isOpen && citationState.chunkId != null; useEffect(() => { - if (!reportOpen && !editorOpen && !hitlEditOpen) return; + if (!reportOpen && !editorOpen && !hitlEditOpen && !citationOpen) return; const handleKeyDown = (e: KeyboardEvent) => { if (e.key === "Escape") { if (hitlEditOpen) closeHitlEdit(); + else if (citationOpen) closeCitation(); else if (editorOpen) closeEditor(); else if (reportOpen) closeReport(); } }; document.addEventListener("keydown", handleKeyDown); return () => document.removeEventListener("keydown", handleKeyDown); - }, [reportOpen, editorOpen, hitlEditOpen, closeReport, closeEditor, closeHitlEdit]); + }, [ + reportOpen, + editorOpen, + hitlEditOpen, + citationOpen, + closeReport, + closeEditor, + closeHitlEdit, + closeCitation, + ]); - const isVisible = (documentsOpen || reportOpen || editorOpen || hitlEditOpen) && !collapsed; + const isVisible = + (documentsOpen || reportOpen || editorOpen || hitlEditOpen || citationOpen) && !collapsed; let effectiveTab = activeTab; if (effectiveTab === "hitl-edit" && !hitlEditOpen) { - effectiveTab = editorOpen ? "editor" : reportOpen ? "report" : "sources"; - } else if (effectiveTab === "editor" && !editorOpen) { - effectiveTab = reportOpen ? "report" : "sources"; - } else if (effectiveTab === "report" && !reportOpen) { - effectiveTab = editorOpen ? "editor" : "sources"; - } else if (effectiveTab === "sources" && !documentsOpen) { - effectiveTab = hitlEditOpen - ? "hitl-edit" + effectiveTab = citationOpen + ? "citation" : editorOpen ? "editor" : reportOpen ? "report" : "sources"; + } else if (effectiveTab === "citation" && !citationOpen) { + effectiveTab = editorOpen ? "editor" : reportOpen ? "report" : "sources"; + } else if (effectiveTab === "editor" && !editorOpen) { + effectiveTab = citationOpen ? "citation" : reportOpen ? "report" : "sources"; + } else if (effectiveTab === "report" && !reportOpen) { + effectiveTab = citationOpen ? "citation" : editorOpen ? "editor" : "sources"; + } else if (effectiveTab === "sources" && !documentsOpen) { + effectiveTab = hitlEditOpen + ? "hitl-edit" + : citationOpen + ? "citation" + : editorOpen + ? "editor" + : reportOpen + ? "report" + : "sources"; } const targetWidth = PANEL_WIDTHS[effectiveTab]; @@ -205,6 +246,11 @@ export function RightPanel({ documentsPanel }: RightPanelProps) { />
)} + {effectiveTab === "citation" && citationOpen && citationState.chunkId != null && ( +
+ +
+ )}
); diff --git a/surfsense_web/components/ui/search-highlight-node.tsx b/surfsense_web/components/ui/search-highlight-node.tsx deleted file mode 100644 index e3f316cce..000000000 --- a/surfsense_web/components/ui/search-highlight-node.tsx +++ /dev/null @@ -1,45 +0,0 @@ -"use client"; - -import type { PlateLeafProps } from "platejs/react"; -import { PlateLeaf } from "platejs/react"; - -/** - * Stable class name used to identify Plate-rendered citation highlight - * leaves in the DOM. We can't use a `data-*` attribute here — Plate's - * `PlateLeaf` runs its props through `useNodeAttributes`, which only - * forwards `attributes`, `className`, `ref`, and `style` to the rendered - * element; arbitrary `data-*` props are silently dropped (verified - * against `@platejs/core/dist/react/index.js` v52). So `className` is - * the only escape hatch that's guaranteed to survive into the DOM. - */ -export const CITATION_HIGHLIGHT_CLASS = "citation-highlight-leaf"; - -/** - * Leaf rendered for ranges decorated by `@platejs/find-replace`'s - * `FindReplacePlugin`. We re-purpose that plugin to drive the citation-jump - * highlight: when a citation is staged, the parent sets the plugin's `search` - * option to a snippet of the chunk text and Plate decorates every match with - * `searchHighlight: true`. This component renders those decorations as a - * `` tagged with `CITATION_HIGHLIGHT_CLASS` so the parent can: - * 1. Query the first match in DOM order to scroll it into view. - * 2. Detect the active-highlight state without a separate React ref. - * - * The highlight is **persistent** — it does not auto-fade. The parent in - * `EditorPanelContent` clears it by setting the plugin's `search` option - * back to "" when one of: (a) the user clicks anywhere inside the editor, - * (b) the panel switches to a different document, (c) the user toggles - * into edit mode, (d) another citation jump is staged, (e) the panel - * unmounts. We use a brief entrance pulse (`citation-flash-in`, see - * `globals.css`) purely to draw the eye after `scrollIntoView` lands. - */ -export function SearchHighlightLeaf(props: PlateLeafProps) { - return ( - - {props.children} - - ); -} diff --git a/surfsense_web/lib/citation-search.ts b/surfsense_web/lib/citation-search.ts deleted file mode 100644 index f80f13076..000000000 --- a/surfsense_web/lib/citation-search.ts +++ /dev/null @@ -1,125 +0,0 @@ -/** - * Snippet generation for the citation-jump highlight, driven by Plate's - * `FindReplacePlugin`. The plugin runs `decorate` per-block and only matches - * within blocks whose children are all `Text` nodes (so it crosses inline - * marks like bold/italic but **not** block boundaries, and a block that - * contains even one inline element such as a link is silently skipped). - * That means a full chunk that spans heading + paragraph won't match as a - * single string — we have to pick a shorter snippet that fits inside one - * rendered block. - * - * `buildCitationSearchCandidates` returns search strings ordered from - * "most-specific anchor" to "broadest fallback": - * 1. First sentence of the chunk (capped at `FIRST_SENTENCE_MAX`). - * 2. First `FIRST_PHRASE_WORDS` words. - * 3. Each non-trivial line of the chunk, in source order — gives us a - * separate attempt for each rendered block, so a heading line with - * an inline link doesn't doom the whole jump. - * 4. Full chunk (only if it's already short enough to plausibly fit - * inside one block). - * - * The caller tries each candidate in turn — set the plugin's `search` - * option, `editor.api.redecorate()`, then check the editor DOM for a - * `.citation-highlight-leaf` element. First candidate that produces one - * wins; subsequent candidates are skipped. - */ - -const FIRST_SENTENCE_MAX = 120; -const FIRST_PHRASE_WORDS = 8; -const MIN_SNIPPET_LENGTH = 6; -const FULL_CHUNK_MAX = FIRST_SENTENCE_MAX * 2; -const MAX_LINE_CANDIDATES = 6; -const LINE_CANDIDATE_MAX = FIRST_SENTENCE_MAX; - -function normalizeWhitespace(input: string): string { - return input.replace(/\s+/g, " ").trim(); -} - -/** - * Strip the markdown syntax that won't survive into the rendered editor's - * plain text, so the chunk text (which comes back from the indexer as raw - * source markdown) can be matched against the literal text values stored - * in Plate's Slate tree. - * - * Order matters: handle multi-char and "container" syntax before single- - * char emphasis, otherwise `**text**` collapses to `*text*` first. - * - * Heuristic only — we don't aim to be a full markdown parser, just to - * remove the common markers (`**bold**`, `[text](url)`, `# headings`, - * `- list`, etc.) that show up in connector-doc chunks and would break - * literal substring search. - */ -export function stripMarkdownForMatch(input: string): string { - let s = input; - s = s.replace(/```[a-z0-9_+-]*\n?([\s\S]*?)```/gi, (_, body: string) => body); - s = s.replace(//g, " "); - s = s.replace(/!\[([^\]]*)\]\([^)]*\)/g, "$1"); - s = s.replace(/!\[([^\]]*)\]\[[^\]]*\]/g, "$1"); - s = s.replace(/\[([^\]]+)\]\([^)]*\)/g, "$1"); - s = s.replace(/\[([^\]]+)\]\[[^\]]*\]/g, "$1"); - s = s.replace(/<((?:https?|mailto):[^>\s]+)>/g, "$1"); - s = s.replace(/`+([^`\n]+?)`+/g, "$1"); - s = s.replace(/(\*\*|__)([\s\S]+?)\1/g, "$2"); - s = s.replace(/(?+[ \t]?/gm, ""); - s = s.replace(/^[ \t]*[-*+][ \t]+/gm, ""); - s = s.replace(/^[ \t]*\d+\.[ \t]+/gm, ""); - s = s.replace(/^[ \t]{0,3}(?:[-*_])(?:[ \t]*[-*_]){2,}[ \t]*$/gm, ""); - s = s.replace(/^[ \t]*\|?(?:[ \t]*:?-+:?[ \t]*\|)+[ \t]*:?-+:?[ \t]*\|?[ \t]*$/gm, ""); - s = s.replace(/\\([\\`*_{}[\]()#+\-.!~>])/g, "$1"); - return s; -} - -export function buildCitationSearchCandidates(rawText: string): string[] { - if (!rawText) return []; - const stripped = stripMarkdownForMatch(rawText); - const normalized = normalizeWhitespace(stripped); - if (normalized.length < MIN_SNIPPET_LENGTH) return []; - - const out: string[] = []; - const seen = new Set(); - const push = (s: string) => { - const t = normalizeWhitespace(s); - if (t.length >= MIN_SNIPPET_LENGTH && !seen.has(t)) { - out.push(t); - seen.add(t); - } - }; - - const sentenceMatch = normalized.match(/^[^.!?]+[.!?]/); - if (sentenceMatch) { - const sentence = sentenceMatch[0]; - push(sentence.length > FIRST_SENTENCE_MAX ? sentence.slice(0, FIRST_SENTENCE_MAX) : sentence); - } else if (normalized.length > FIRST_SENTENCE_MAX) { - push(normalized.slice(0, FIRST_SENTENCE_MAX)); - } - - const words = normalized.split(" ").filter(Boolean); - if (words.length > FIRST_PHRASE_WORDS) { - push(words.slice(0, FIRST_PHRASE_WORDS).join(" ")); - } - - // Per-line candidates: each chunk line is roughly one block in the - // rendered editor. Trying them in order gives us a separate decorate - // attempt for each block, which matters when the first line is a - // heading containing a link (Plate's `FindReplacePlugin` will skip - // any block whose children aren't all text nodes). - const rawLines = stripped.split(/\r?\n/); - let lineCount = 0; - for (const line of rawLines) { - if (lineCount >= MAX_LINE_CANDIDATES) break; - const trimmed = normalizeWhitespace(line); - if (trimmed.length < MIN_SNIPPET_LENGTH) continue; - push(trimmed.length > LINE_CANDIDATE_MAX ? trimmed.slice(0, LINE_CANDIDATE_MAX) : trimmed); - lineCount++; - } - - if (normalized.length <= FULL_CHUNK_MAX) { - push(normalized); - } - - return out; -} diff --git a/surfsense_web/package.json b/surfsense_web/package.json index 665490e4f..41175daeb 100644 --- a/surfsense_web/package.json +++ b/surfsense_web/package.json @@ -36,7 +36,6 @@ "@platejs/code-block": "^52.0.11", "@platejs/combobox": "^52.0.15", "@platejs/dnd": "^52.0.11", - "@platejs/find-replace": "^52.3.10", "@platejs/floating": "^52.0.11", "@platejs/indent": "^52.0.11", "@platejs/link": "^52.0.11", diff --git a/surfsense_web/pnpm-lock.yaml b/surfsense_web/pnpm-lock.yaml index a1a7bea12..b1730e842 100644 --- a/surfsense_web/pnpm-lock.yaml +++ b/surfsense_web/pnpm-lock.yaml @@ -53,9 +53,6 @@ importers: '@platejs/dnd': specifier: ^52.0.11 version: 52.0.11(platejs@52.0.17(@types/react@19.2.14)(immer@10.2.0)(react-dom@19.2.4(react@19.2.4))(react@19.2.4)(scheduler@0.27.0)(use-sync-external-store@1.6.0(react@19.2.4)))(react-dnd-html5-backend@16.0.1)(react-dnd@16.0.1(@types/node@20.19.33)(@types/react@19.2.14)(react@19.2.4))(react-dom@19.2.4(react@19.2.4))(react@19.2.4) - '@platejs/find-replace': - specifier: ^52.3.10 - version: 52.3.10(platejs@52.0.17(@types/react@19.2.14)(immer@10.2.0)(react-dom@19.2.4(react@19.2.4))(react@19.2.4)(scheduler@0.27.0)(use-sync-external-store@1.6.0(react@19.2.4)))(react-dom@19.2.4(react@19.2.4))(react@19.2.4) '@platejs/floating': specifier: ^52.0.11 version: 52.0.11(platejs@52.0.17(@types/react@19.2.14)(immer@10.2.0)(react-dom@19.2.4(react@19.2.4))(react@19.2.4)(scheduler@0.27.0)(use-sync-external-store@1.6.0(react@19.2.4)))(react-dom@19.2.4(react@19.2.4))(react@19.2.4) @@ -2830,13 +2827,6 @@ packages: react-dnd-html5-backend: '>=14.0.0' react-dom: '>=18.0.0' - '@platejs/find-replace@52.3.10': - resolution: {integrity: sha512-V/MOMMUYxHfEn/skd2+YO213xSATFDVsl8FzVzVRV/XaxwwVefH2EPD1lAVIvmYjennTVTTsHHtEI9K9iOsEaA==} - peerDependencies: - platejs: '>=52.0.11' - react: '>=18.0.0' - react-dom: '>=18.0.0' - '@platejs/floating@52.0.11': resolution: {integrity: sha512-ApNpw4KWml+kuK+XTTpji+f/7GxTR4nRzlnfJMvGBrJpLPQ4elS5MABm3oUi81DZn+aub5HvsyH7UqCw7F76IA==} peerDependencies: @@ -11115,13 +11105,6 @@ snapshots: react-dnd-html5-backend: 16.0.1 react-dom: 19.2.4(react@19.2.4) - '@platejs/find-replace@52.3.10(platejs@52.0.17(@types/react@19.2.14)(immer@10.2.0)(react-dom@19.2.4(react@19.2.4))(react@19.2.4)(scheduler@0.27.0)(use-sync-external-store@1.6.0(react@19.2.4)))(react-dom@19.2.4(react@19.2.4))(react@19.2.4)': - dependencies: - platejs: 52.0.17(@types/react@19.2.14)(immer@10.2.0)(react-dom@19.2.4(react@19.2.4))(react@19.2.4)(scheduler@0.27.0)(use-sync-external-store@1.6.0(react@19.2.4)) - react: 19.2.4 - react-compiler-runtime: 1.0.0(react@19.2.4) - react-dom: 19.2.4(react@19.2.4) - '@platejs/floating@52.0.11(platejs@52.0.17(@types/react@19.2.14)(immer@10.2.0)(react-dom@19.2.4(react@19.2.4))(react@19.2.4)(scheduler@0.27.0)(use-sync-external-store@1.6.0(react@19.2.4)))(react-dom@19.2.4(react@19.2.4))(react@19.2.4)': dependencies: '@floating-ui/core': 1.7.4 From f9b5367754c5e07a586b5a318ac06245b3d10846 Mon Sep 17 00:00:00 2001 From: "DESKTOP-RTLN3BA\\$punk" Date: Tue, 28 Apr 2026 23:52:37 -0700 Subject: [PATCH 17/17] chore: cleaned comments slop --- surfsense_backend/.env.example | 14 +- .../versions/130_add_agent_action_log.py | 6 +- .../versions/131_add_document_revisions.py | 2 +- .../132_add_agent_permission_rules.py | 9 +- .../app/agents/new_chat/chat_deepagent.py | 125 +++++++++--------- .../app/agents/new_chat/errors.py | 8 +- .../app/agents/new_chat/feature_flags.py | 31 ++--- .../agents/new_chat/middleware/busy_mutex.py | 15 ++- .../agents/new_chat/middleware/compaction.py | 19 +-- .../new_chat/middleware/context_editing.py | 18 +-- .../new_chat/middleware/dedup_tool_calls.py | 10 +- .../agents/new_chat/middleware/doom_loop.py | 22 +-- .../new_chat/middleware/knowledge_search.py | 21 +-- .../new_chat/middleware/noop_injection.py | 28 ++-- .../agents/new_chat/middleware/otel_span.py | 6 +- .../agents/new_chat/middleware/permission.py | 25 ++-- .../agents/new_chat/middleware/retry_after.py | 14 +- .../new_chat/middleware/tool_call_repair.py | 19 +-- .../app/agents/new_chat/permissions.py | 9 +- .../app/agents/new_chat/plugin_loader.py | 9 +- .../new_chat/plugins/year_substituter.py | 10 +- .../app/agents/new_chat/prompts/composer.py | 21 ++- .../app/agents/new_chat/subagents/__init__.py | 15 ++- .../app/agents/new_chat/system_prompt.py | 15 ++- .../app/agents/new_chat/tools/invalid_tool.py | 5 +- .../app/agents/new_chat/tools/registry.py | 6 +- surfsense_backend/app/observability/otel.py | 4 +- .../app/routes/agent_revert_route.py | 8 +- .../agents/new_chat/prompts/test_composer.py | 2 +- .../unit/agents/new_chat/test_otel_span.py | 2 +- .../unit/agents/new_chat/test_permissions.py | 2 +- .../agents/new_chat/test_plugin_loader.py | 2 +- .../tests/unit/observability/test_otel.py | 2 +- .../unit/services/test_revert_service.py | 2 +- 34 files changed, 274 insertions(+), 232 deletions(-) diff --git a/surfsense_backend/.env.example b/surfsense_backend/.env.example index e133a2bc5..c1bfcc538 100644 --- a/surfsense_backend/.env.example +++ b/surfsense_backend/.env.example @@ -250,12 +250,12 @@ LANGSMITH_PROJECT=surfsense # ============================================================================= -# OPTIONAL: New-chat agent feature flags (OpenCode-port) +# OPTIONAL: New-chat agent feature flags # ============================================================================= # Master kill-switch — when true, every flag below is forced OFF. # SURFSENSE_DISABLE_NEW_AGENT_STACK=false -# Tier 1 — Agent quality +# Agent quality # SURFSENSE_ENABLE_CONTEXT_EDITING=false # SURFSENSE_ENABLE_COMPACTION_V2=false # SURFSENSE_ENABLE_RETRY_AFTER=false @@ -265,24 +265,24 @@ LANGSMITH_PROJECT=surfsense # SURFSENSE_ENABLE_TOOL_CALL_REPAIR=false # SURFSENSE_ENABLE_DOOM_LOOP=false # leave OFF until UI handles permission='doom_loop' -# Tier 2 — Safety +# Safety # SURFSENSE_ENABLE_PERMISSION=false # SURFSENSE_ENABLE_BUSY_MUTEX=false # SURFSENSE_ENABLE_LLM_TOOL_SELECTOR=false # adds a per-turn LLM call -# Tier 3b — Observability (also requires OTEL_EXPORTER_OTLP_ENDPOINT) +# Observability — OTel (also requires OTEL_EXPORTER_OTLP_ENDPOINT) # SURFSENSE_ENABLE_OTEL=false -# Tier 4 — Skills + subagents +# Skills + subagents # SURFSENSE_ENABLE_SKILLS=false # SURFSENSE_ENABLE_SPECIALIZED_SUBAGENTS=false # SURFSENSE_ENABLE_KB_PLANNER_RUNNABLE=false -# Tier 5 — Snapshot / revert +# Snapshot / revert # SURFSENSE_ENABLE_ACTION_LOG=false # SURFSENSE_ENABLE_REVERT_ROUTE=false # Backend-only; flip when UI ships -# Tier 6 — Plugins +# Plugins # SURFSENSE_ENABLE_PLUGIN_LOADER=false # Comma-separated allowlist of plugin entry-point names # SURFSENSE_ALLOWED_PLUGINS=year_substituter diff --git a/surfsense_backend/alembic/versions/130_add_agent_action_log.py b/surfsense_backend/alembic/versions/130_add_agent_action_log.py index 2f06b8ddd..f86a8a3b5 100644 --- a/surfsense_backend/alembic/versions/130_add_agent_action_log.py +++ b/surfsense_backend/alembic/versions/130_add_agent_action_log.py @@ -4,8 +4,10 @@ Revision ID: 130 Revises: 129 Create Date: 2026-04-28 -Tier 5.2 in the OpenCode-port plan. Adds the append-only ``agent_action_log`` -table that :class:`ActionLogMiddleware` writes to after every tool call. +Adds the append-only ``agent_action_log`` table that +:class:`ActionLogMiddleware` writes to after every tool call. Each row +optionally carries a ``reverse_descriptor`` payload used by +``POST /api/threads/{thread_id}/revert/{action_id}`` to undo the action. """ from __future__ import annotations diff --git a/surfsense_backend/alembic/versions/131_add_document_revisions.py b/surfsense_backend/alembic/versions/131_add_document_revisions.py index 46c6991b6..95ce0e032 100644 --- a/surfsense_backend/alembic/versions/131_add_document_revisions.py +++ b/surfsense_backend/alembic/versions/131_add_document_revisions.py @@ -4,7 +4,7 @@ Revision ID: 131 Revises: 130 Create Date: 2026-04-28 -Tier 5.1 in the OpenCode-port plan. Adds two snapshot tables: +Adds two snapshot tables that back the per-action revert flow: * ``document_revisions``: pre-mutation snapshot of NOTE/FILE/EXTENSION docs. * ``folder_revisions``: pre-mutation snapshot of folder mkdir/move/delete. diff --git a/surfsense_backend/alembic/versions/132_add_agent_permission_rules.py b/surfsense_backend/alembic/versions/132_add_agent_permission_rules.py index 0e81eacb5..ff5b52e18 100644 --- a/surfsense_backend/alembic/versions/132_add_agent_permission_rules.py +++ b/surfsense_backend/alembic/versions/132_add_agent_permission_rules.py @@ -4,11 +4,10 @@ Revision ID: 132 Revises: 131 Create Date: 2026-04-28 -Tier 2.1 in the OpenCode-port plan. Adds the persistent ``agent_permission_rules`` -table consumed by :class:`PermissionMiddleware` at agent build time. Rules -can be scoped at search-space (``user_id`` / ``thread_id`` NULL), -user-wide (``user_id`` set, ``thread_id`` NULL), or per-thread -(``thread_id`` set). +Adds the persistent ``agent_permission_rules`` table consumed by +:class:`PermissionMiddleware` at agent build time. Rules can be scoped +at search-space (``user_id`` / ``thread_id`` NULL), user-wide +(``user_id`` set, ``thread_id`` NULL), or per-thread (``thread_id`` set). """ from __future__ import annotations diff --git a/surfsense_backend/app/agents/new_chat/chat_deepagent.py b/surfsense_backend/app/agents/new_chat/chat_deepagent.py index 3ca44dd4f..bfb94ba2d 100644 --- a/surfsense_backend/app/agents/new_chat/chat_deepagent.py +++ b/surfsense_backend/app/agents/new_chat/chat_deepagent.py @@ -353,11 +353,12 @@ async def create_surfsense_deep_agent( additional_tools=list(additional_tools) if additional_tools else None, ) - # Tier 1.6: register `invalid` tool. It is dispatched only when - # ToolCallNameRepairMiddleware rewrites a malformed call. We - # intentionally append it AFTER ``build_tools_async`` so it never - # appears in the system-prompt tool list (which is built from the - # registry, not the bound tool list). + # Register the ``invalid`` tool only when tool-call repair is on. It + # is dispatched only when :class:`ToolCallNameRepairMiddleware` + # rewrites a malformed call. We intentionally append it AFTER + # ``build_tools_async`` so it never appears in the system-prompt + # tool list (which is built from the registry, not the bound tool + # list). _flags: AgentFeatureFlags = get_flags() if _flags.enable_tool_call_repair and INVALID_TOOL_NAME not in { t.name for t in tools @@ -455,10 +456,10 @@ async def create_surfsense_deep_agent( return agent -# Tier 1.1: tools whose output is too costly / lossy to discard. Keep -# this conservative — anything listed here is *never* pruned by -# ContextEditingMiddleware. The list is filtered against actually-bound -# tool names so disabled connectors don't show up here. +# Tools whose output is too costly / lossy to discard. Keep this +# conservative — anything listed here is *never* pruned by +# :class:`ContextEditingMiddleware`. The list is filtered against +# actually-bound tool names so disabled connectors don't show up here. _PRUNE_PROTECTED_TOOL_NAMES: frozenset[str] = frozenset( { "generate_report", @@ -485,11 +486,12 @@ def _safe_exclude_tools(tools: Sequence[BaseTool]) -> tuple[str, ...]: return tuple(name for name in _PRUNE_PROTECTED_TOOL_NAMES if name in enabled) -# Tier 2.1 / cleanup: opencode `Permission.disabled` parity. Replaces the -# legacy binary ``_CONNECTOR_TYPE_TO_SEARCHABLE``-based gating with a -# declarative pass over :data:`BUILTIN_TOOLS`. Each tool that declares a -# ``required_connector`` not present in ``available_connectors`` gets a -# deny rule so any execution attempt short-circuits with permission_denied. +# Connector gating: any tool whose ``ToolDefinition.required_connector`` +# isn't actually wired up gets a synthesized permission deny rule so +# execution attempts short-circuit with ``permission_denied`` instead of +# bubbling up provider-specific 401/404 errors. Mirrors OpenCode's +# ``Permission.disabled`` (declarative, per-tool gating) — replaces the +# legacy binary ``_CONNECTOR_TYPE_TO_SEARCHABLE`` substring-heuristic. def _synthesize_connector_deny_rules( *, available_connectors: list[str] | None, @@ -503,11 +505,6 @@ def _synthesize_connector_deny_rules( 1. It is currently bound (``enabled_tool_names``). 2. It declares a ``required_connector``. 3. That connector is *not* in ``available_connectors``. - - This expresses the OpenCode ``Permission.disabled`` semantics - declaratively, replacing the substring-heuristic binary gating - that used to consult the hardcoded ``_CONNECTOR_TYPE_TO_SEARCHABLE`` - map. """ available = set(available_connectors or []) deny: list[Rule] = [] @@ -581,7 +578,7 @@ def _build_compiled_agent_blocking( "middleware": gp_middleware, } - # Tier 4.3: specialized user-facing subagents (explore, report_writer, + # Specialized user-facing subagents (explore, report_writer, # connector_negotiator). Registered through SubAgentMiddleware alongside # the general-purpose spec so the parent's `task` tool can address them # by name. Off by default until the flag flips so existing deployments @@ -629,14 +626,13 @@ def _build_compiled_agent_blocking( # ``wrap_model_call`` ordering: the FIRST middleware in the list is the # OUTERMOST wrapper. To ensure prune executes before summarization, # place ``SpillingContextEditingMiddleware`` before - # ``SurfSenseCompactionMiddleware`` (Tier 1.1 + 1.3). - # Compaction is the canonical token-budget defense after the - # cleanup tier removed ``SafeSummarizationMiddleware``. The Bedrock - # buffer-empty defense is folded into ``SurfSenseCompactionMiddleware``. + # ``SurfSenseCompactionMiddleware``. Compaction is the canonical + # token-budget defense; the Bedrock buffer-empty defense is folded + # into ``SurfSenseCompactionMiddleware``. summarization_mw = create_surfsense_compaction_middleware(llm, StateBackend) _ = flags.enable_compaction_v2 # historical flag; retained for telemetry parity - # Tier 1.1: ContextEditing prune. Trigger at 55% of model_max_input, + # ContextEditing prune. Trigger at 55% of ``max_input_tokens``, # earlier than summarization (~85%). When disabled, no edit runs. context_edit_mw = None if ( @@ -664,7 +660,10 @@ def _build_compiled_agent_blocking( backend_resolver=backend_resolver, ) - # Tier 1.4 / 1.8 / 1.9 / 1.10: built-in retry/fallback/limits. + # Resilience knobs: header-aware retry, model fallback, and + # per-thread / per-run call-count limits. The fallback / limit + # middlewares are vanilla LangChain primitives; ``RetryAfter`` is + # SurfSense's header-aware variant (see its module docstring). retry_mw = ( RetryAfterMiddleware(max_retries=3) if flags.enable_retry_after and not flags.disable_new_agent_stack @@ -700,14 +699,16 @@ def _build_compiled_agent_blocking( else None ) - # Tier 1.5: provider-compat _noop injection. + # Provider-compat ``_noop`` injection (mirrors OpenCode's + # ``llm.ts`` workaround for providers that reject empty assistant + # turns or alternating-role constraints). noop_mw = ( NoopInjectionMiddleware() if flags.enable_compaction_v2 and not flags.disable_new_agent_stack else None ) - # Tier 1.7: tool-call name repair (lowercase + invalid fallback). + # Tool-call name repair (lowercase + ``invalid`` fallback). # # ``registered_tool_names`` MUST cover every tool the model can legitimately # call. That includes the bound ``tools`` list AND every tool provided by @@ -737,18 +738,22 @@ def _build_compiled_agent_blocking( } repair_mw = ToolCallNameRepairMiddleware( registered_tool_names=registered_names, - fuzzy_match_threshold=None, # opencode parity: no fuzzy step + # Disable fuzzy matching to avoid silent rewrites; the + # lowercase + ``invalid`` fallback alone covers >95% of + # observed model errors. + fuzzy_match_threshold=None, ) - # Tier 1.11: doom-loop detector. Off by default until UI handles. + # Doom-loop detector. Off by default until the frontend handles + # ``permission == "doom_loop"`` interrupts. doom_loop_mw = ( DoomLoopMiddleware(threshold=3) if flags.enable_doom_loop and not flags.disable_new_agent_stack else None ) - # Tier 2.1: PermissionMiddleware. Layers, earliest -> latest (last - # match wins per opencode): + # PermissionMiddleware. Layers, earliest -> latest (last match wins, + # same evaluation order as OpenCode's ``permission/index.ts``): # # 1. ``surfsense_defaults`` — single ``allow */*`` rule. SurfSense # already runs per-tool HITL (see ``tools/hitl.py``) for mutating @@ -778,11 +783,11 @@ def _build_compiled_agent_blocking( ], ) - # Tier 5.2: ActionLogMiddleware. Off by default until the - # ``agent_action_log`` table is migrated. When enabled, persists one - # row per tool call with optional reverse_descriptor for - # /api/threads/{thread_id}/revert/{action_id}. Sits inside permission - # so denied calls aren't logged as completions. + # ActionLogMiddleware. Off by default until the ``agent_action_log`` + # table is migrated. When enabled, persists one row per tool call + # with optional reverse_descriptor for + # ``POST /api/threads/{thread_id}/revert/{action_id}``. Sits inside + # ``permission`` so denied calls aren't logged as completions. action_log_mw: ActionLogMiddleware | None = None if ( flags.enable_action_log @@ -804,23 +809,24 @@ def _build_compiled_agent_blocking( ) action_log_mw = None - # Tier 2.2: per-thread busy mutex. + # Per-thread busy mutex (refuse a second concurrent turn on the same + # thread; see :class:`BusyMutexMiddleware` docstring). busy_mutex_mw: BusyMutexMiddleware | None = ( BusyMutexMiddleware() if flags.enable_busy_mutex and not flags.disable_new_agent_stack else None ) - # Tier 3b: OpenTelemetry spans (model.call + tool.call). Lives just - # inside BusyMutex so it spans every retry/fallback attempt of the - # current turn but never wraps a queued/blocked turn. + # OpenTelemetry spans (model.call + tool.call). Lives just inside + # BusyMutex so it spans every retry/fallback attempt of the current + # turn but never wraps a queued/blocked turn. otel_mw: OtelSpanMiddleware | None = ( OtelSpanMiddleware() if flags.enable_otel and not flags.disable_new_agent_stack else None ) - # Tier 6: plugin entry-point loader. Off by default; opt-in via the + # Plugin entry-point loader. Off by default; opt-in via the # ``SURFSENSE_ENABLE_PLUGIN_LOADER`` flag. The allowlist is read from # the ``SURFSENSE_ALLOWED_PLUGINS`` env var (comma-separated). A future # PR can wire it through ``global_llm_config.yaml``. @@ -845,10 +851,10 @@ def _build_compiled_agent_blocking( ) plugin_middlewares = [] - # Tier 4.1: SkillsMiddleware. Loads built-in + space-authored skills - # via a CompositeBackend. Sources are layered: built-in first, space - # last, so a search-space-authored skill of the same name overrides - # the bundled one. + # SkillsMiddleware (deepagents) loads built-in + space-authored + # skills via a CompositeBackend. Sources are layered: built-in first, + # space last, so a search-space-authored skill of the same name + # overrides the bundled one. skills_mw: SkillsMiddleware | None = None if flags.enable_skills and not flags.disable_new_agent_stack: try: @@ -865,7 +871,8 @@ def _build_compiled_agent_blocking( logging.warning("SkillsMiddleware init failed; skipping: %s", exc) skills_mw = None - # Tier 2.5: LLM-driven tool selection for >30 tools. + # LangChain's LLM-driven tool selection — only enabled for stacks + # large enough to need narrowing (>30 tools). selector_mw: LLMToolSelectorMiddleware | None = None if ( flags.enable_llm_tool_selector @@ -934,12 +941,12 @@ def _build_compiled_agent_blocking( ) if filesystem_mode == FilesystemMode.CLOUD else None, - # Tier 4.1: skill loader. Placed before SubAgentMiddleware so - # subagents inherit the same skill metadata (subagent specs reference - # the same source paths via `default_skills_sources()`). + # Skill loader. Placed before SubAgentMiddleware so subagents + # inherit the same skill metadata (subagent specs reference the + # same source paths via ``default_skills_sources()``). skills_mw, SubAgentMiddleware(backend=StateBackend, subagents=subagent_specs), - # Tier 2.5: tool selection (only when >30 tools and flag on). + # Tool selection (only when >30 tools and flag on). selector_mw, # Defensive caps, then prune, then summarize. model_call_limit_mw, @@ -954,19 +961,19 @@ def _build_compiled_agent_blocking( # Tool-call repair must run after model emits but before # permission / dedup / doom-loop interpret the calls. repair_mw, - # Tier 2.1: deny/ask BEFORE the calls are forwarded to tool nodes. + # Permission deny/ask BEFORE the calls are forwarded to tool nodes. permission_mw, doom_loop_mw, - # Tier 5.2: action log sits inside permission so denied calls - # don't appear as completions, and outside dedup so each unique - # tool invocation gets its own row. + # Action log sits inside permission so denied calls don't appear + # as completions, and outside dedup so each unique tool invocation + # gets its own row. action_log_mw, PatchToolCallsMiddleware(), DedupHITLToolCallsMiddleware(agent_tools=list(tools)), - # Tier 6: plugin slot — sits just before AnthropicCache so plugin-side - # transforms see the final tool result and run before any caching - # heuristics. Multiple plugins in declared order; loader filtered by - # the admin allowlist already. + # Plugin slot — sits just before AnthropicCache so plugin-side + # transforms see the final tool result and run before any + # caching heuristics. Multiple plugins in declared order; loader + # filtered by the admin allowlist already. *plugin_middlewares, AnthropicPromptCachingMiddleware(unsupported_model_behavior="ignore"), ] diff --git a/surfsense_backend/app/agents/new_chat/errors.py b/surfsense_backend/app/agents/new_chat/errors.py index b7bac4536..a17333acc 100644 --- a/surfsense_backend/app/agents/new_chat/errors.py +++ b/surfsense_backend/app/agents/new_chat/errors.py @@ -2,10 +2,10 @@ Typed error taxonomy for the SurfSense agent stack. Used by: -- :class:`RetryAfterMiddleware` (Tier 1.4) — its ``retry_on`` callable - consults the error code to decide whether a retry is appropriate. -- :class:`PermissionMiddleware` (Tier 2.1) — emits - ``code="permission_denied"`` errors when a deny rule trips. +- :class:`RetryAfterMiddleware` — its ``retry_on`` callable consults + the error code to decide whether a retry is appropriate. +- :class:`PermissionMiddleware` — emits ``code="permission_denied"`` + errors when a deny rule trips. - All tools — return :class:`StreamingError` payloads in ``ToolMessage.additional_kwargs["error"]`` so the model and the retry/permission layers share a contract. diff --git a/surfsense_backend/app/agents/new_chat/feature_flags.py b/surfsense_backend/app/agents/new_chat/feature_flags.py index 89c4fb14f..55525abc5 100644 --- a/surfsense_backend/app/agents/new_chat/feature_flags.py +++ b/surfsense_backend/app/agents/new_chat/feature_flags.py @@ -1,9 +1,10 @@ """ Feature flags for the SurfSense new_chat agent stack. -These flags control rollout of OpenCode-pattern middleware ported into -SurfSense. They follow a "default-OFF for risky things, default-ON for -safe upgrades, master kill-switch for everything new" model. +These flags gate the newer agent middleware (some ported from OpenCode, +some sourced from ``langchain.agents.middleware`` / ``deepagents``, some +SurfSense-native). They follow a "default-OFF for risky things, +default-ON for safe upgrades, master kill-switch for everything new" model. All new middleware checks its flag at agent build time. If the master kill-switch ``SURFSENSE_DISABLE_NEW_AGENT_STACK`` is set, every new @@ -57,7 +58,7 @@ class AgentFeatureFlags: # regardless of its env value. Used for rapid rollback. disable_new_agent_stack: bool = False - # Tier 1 — Agent quality + # Agent quality — context budget, retry/limits, name-repair, doom-loop enable_context_editing: bool = False enable_compaction_v2: bool = False enable_retry_after: bool = False @@ -69,26 +70,26 @@ class AgentFeatureFlags: False # Default OFF until UI handles permission='doom_loop' ) - # Tier 2 — Safety + # Safety — permissions, concurrency, tool-set narrowing enable_permission: bool = False # Default OFF for first deploy enable_busy_mutex: bool = False enable_llm_tool_selector: bool = False # Default OFF — adds per-turn LLM cost - # Tier 4 — Skills + subagents + # Skills + subagents enable_skills: bool = False enable_specialized_subagents: bool = False enable_kb_planner_runnable: bool = False - # Tier 5 — Snapshot / revert + # Snapshot / revert enable_action_log: bool = False enable_revert_route: bool = ( False # Backend ships before UI; route returns 503 until this flips ) - # Tier 6 — Plugins + # Plugins enable_plugin_loader: bool = False - # Tier 3b — OTel (orthogonal: also requires OTEL_EXPORTER_OTLP_ENDPOINT) + # Observability — OTel (orthogonal; also requires OTEL_EXPORTER_OTLP_ENDPOINT) enable_otel: bool = False @classmethod @@ -108,7 +109,7 @@ class AgentFeatureFlags: return cls( disable_new_agent_stack=False, - # Tier 1 + # Agent quality enable_context_editing=_env_bool("SURFSENSE_ENABLE_CONTEXT_EDITING", False), enable_compaction_v2=_env_bool("SURFSENSE_ENABLE_COMPACTION_V2", False), enable_retry_after=_env_bool("SURFSENSE_ENABLE_RETRY_AFTER", False), @@ -121,13 +122,13 @@ class AgentFeatureFlags: "SURFSENSE_ENABLE_TOOL_CALL_REPAIR", False ), enable_doom_loop=_env_bool("SURFSENSE_ENABLE_DOOM_LOOP", False), - # Tier 2 + # Safety enable_permission=_env_bool("SURFSENSE_ENABLE_PERMISSION", False), enable_busy_mutex=_env_bool("SURFSENSE_ENABLE_BUSY_MUTEX", False), enable_llm_tool_selector=_env_bool( "SURFSENSE_ENABLE_LLM_TOOL_SELECTOR", False ), - # Tier 4 + # Skills + subagents enable_skills=_env_bool("SURFSENSE_ENABLE_SKILLS", False), enable_specialized_subagents=_env_bool( "SURFSENSE_ENABLE_SPECIALIZED_SUBAGENTS", False @@ -135,12 +136,12 @@ class AgentFeatureFlags: enable_kb_planner_runnable=_env_bool( "SURFSENSE_ENABLE_KB_PLANNER_RUNNABLE", False ), - # Tier 5 + # Snapshot / revert enable_action_log=_env_bool("SURFSENSE_ENABLE_ACTION_LOG", False), enable_revert_route=_env_bool("SURFSENSE_ENABLE_REVERT_ROUTE", False), - # Tier 6 + # Plugins enable_plugin_loader=_env_bool("SURFSENSE_ENABLE_PLUGIN_LOADER", False), - # Tier 3b + # Observability enable_otel=_env_bool("SURFSENSE_ENABLE_OTEL", False), ) diff --git a/surfsense_backend/app/agents/new_chat/middleware/busy_mutex.py b/surfsense_backend/app/agents/new_chat/middleware/busy_mutex.py index 1d95638d0..c57d85004 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/busy_mutex.py +++ b/surfsense_backend/app/agents/new_chat/middleware/busy_mutex.py @@ -1,11 +1,16 @@ """ BusyMutexMiddleware — per-thread asyncio lock + cancel token. -Tier 2.2 in the OpenCode-port plan. Mirrors opencode's -``Stream.scoped(AbortController)`` pattern (single-process, in-memory -lock + cooperative cancellation). For multi-worker deployments a -distributed lock backend (Redis or PostgreSQL advisory locks) is a -phase-2 follow-up. +LangChain has no built-in concept of "this thread is already running a +turn — refuse the second concurrent request". Without it, a user +double-clicking "send" or refreshing the page mid-stream can spawn two +turns racing on the same checkpoint, producing duplicated tool calls +and mangled state. + +Ported from OpenCode's ``Stream.scoped(AbortController)`` pattern: a +single-process, in-memory lock + cooperative cancellation token keyed by +``thread_id``. For multi-worker deployments a distributed lock backend +(Redis or PostgreSQL advisory locks) is a phase-2 follow-up. What this provides: - A ``WeakValueDictionary[str, asyncio.Lock]`` keyed by ``thread_id``; diff --git a/surfsense_backend/app/agents/new_chat/middleware/compaction.py b/surfsense_backend/app/agents/new_chat/middleware/compaction.py index b0a1a7ec5..16361e16b 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/compaction.py +++ b/surfsense_backend/app/agents/new_chat/middleware/compaction.py @@ -5,21 +5,22 @@ Subclasses :class:`deepagents.middleware.summarization.SummarizationMiddleware` to add SurfSense-specific behavior: 1. **Structured summary template** (OpenCode-style ``## Goal / Constraints / - Progress / Key Decisions / Next Steps / Critical Context / Relevant Files``). + Progress / Key Decisions / Next Steps / Critical Context / Relevant Files``) + — see :data:`SURFSENSE_SUMMARY_PROMPT` below. The base + ``SummarizationMiddleware`` only ships a freeform "summarize this" + prompt; the structured template is ported from OpenCode's + ``compaction.ts``. 2. **Protect SurfSense-specific SystemMessages** so injected hints (````, ````, ````, ````, ````, ````, ````) are *not* summarized away and are kept verbatim in the post-summary - message list. + message list. Mirrors OpenCode's ``PRUNE_PROTECTED_TOOLS`` philosophy + (some message types are part of the agent's contract and must survive + compaction unchanged). 3. **Sanitize ``content=None``** when feeding messages into ``get_buffer_string`` (Azure OpenAI / LiteLLM defense — when a provider streams an AIMessage containing only tool_calls and no text, ``content`` can be ``None`` and - ``get_buffer_string`` crashes iterating over ``None``). This used to live in - ``safe_summarization.py``; folded in here. - -This replaces ``app.agents.new_chat.middleware.safe_summarization``. - -Tier 1.3 in the OpenCode-port plan. + ``get_buffer_string`` crashes iterating over ``None``). SurfSense-specific. """ from __future__ import annotations @@ -42,7 +43,7 @@ if TYPE_CHECKING: logger = logging.getLogger(__name__) -# OpenCode-faithful structured summary template. Mirrors +# Structured summary template ported from OpenCode's # ``opencode/packages/opencode/src/session/compaction.ts:40-75``. Kept as a # module-level constant so unit tests can assert on its sections. SURFSENSE_SUMMARY_PROMPT = """ diff --git a/surfsense_backend/app/agents/new_chat/middleware/context_editing.py b/surfsense_backend/app/agents/new_chat/middleware/context_editing.py index 360e3e28f..39bc57c8b 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/context_editing.py +++ b/surfsense_backend/app/agents/new_chat/middleware/context_editing.py @@ -1,15 +1,15 @@ """ SpillToBackendEdit + SpillingContextEditingMiddleware. -Mirrors OpenCode's spill-to-disk behavior in -``opencode/packages/opencode/src/tool/truncate.ts``. Before -``ClearToolUsesEdit`` rewrites old ``ToolMessage.content`` to a placeholder, -we capture the full original content and write it to the runtime backend -under ``/tool_outputs/{thread_id}/{message_id}.txt``. The placeholder is -upgraded to ``"[cleared — full output at /tool_outputs/.../{id}.txt; ask the -explore subagent to read it]"`` so the agent can recover it on demand. - -Tier 1.2 in the OpenCode-port plan. +LangChain's :class:`ClearToolUsesEdit` discards old ``ToolMessage.content`` +when the context-editing budget triggers, replacing the body with a fixed +placeholder. That's lossy: anything the agent might want to revisit is +gone. The spill-to-disk pattern (originally from OpenCode's +``opencode/packages/opencode/src/tool/truncate.ts``) keeps the prune +behavior but writes the full original payload to the runtime backend +under ``/tool_outputs/{thread_id}/{message_id}.txt`` first. The +placeholder is then upgraded to point at the spill path so the agent +(or a subagent) can read it back on demand. Why this is a middleware subclass instead of a plain ``ContextEdit``: ``ContextEdit.apply`` is sync, but writing to the backend is async. We diff --git a/surfsense_backend/app/agents/new_chat/middleware/dedup_tool_calls.py b/surfsense_backend/app/agents/new_chat/middleware/dedup_tool_calls.py index 3aff524fe..c55347284 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/dedup_tool_calls.py +++ b/surfsense_backend/app/agents/new_chat/middleware/dedup_tool_calls.py @@ -9,11 +9,10 @@ the duplicate call is stripped from the AIMessage that gets checkpointed. That means it is also safe across LangGraph ``interrupt()`` boundaries: the removed call will never appear on graph resume. -Dedup-key resolution order (Tier 2.3 / cleanup in the OpenCode-port plan): +Dedup-key resolution order: 1. :class:`ToolDefinition.dedup_key` — callable provided by the registry - entry. This is the canonical mechanism after the cleanup-tier removal - of the legacy ``PRIMARY_ARG`` map. + entry. This is the canonical mechanism. 2. ``tool.metadata["hitl_dedup_key"]`` — string with a primary arg name; used by MCP / Composio tools whose schemas the registry doesn't see. @@ -72,9 +71,8 @@ class DedupHITLToolCallsMiddleware(AgentMiddleware): # type: ignore[type-arg] The dedup-resolver map is built from two sources, in priority order: 1. ``tool.metadata["dedup_key"]`` — callable provided by the registry's - ``ToolDefinition.dedup_key`` (Tier 2.3). Receives the args dict - and returns a string signature. This is the canonical mechanism - after the cleanup-tier removal of the legacy ``PRIMARY_ARG`` map. + ``ToolDefinition.dedup_key``. Receives the args dict and returns + a string signature. This is the canonical mechanism. 2. ``tool.metadata["hitl_dedup_key"]`` — string with a primary arg name; primarily used by MCP / Composio tools. """ diff --git a/surfsense_backend/app/agents/new_chat/middleware/doom_loop.py b/surfsense_backend/app/agents/new_chat/middleware/doom_loop.py index 1dde87752..850ecd1d2 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/doom_loop.py +++ b/surfsense_backend/app/agents/new_chat/middleware/doom_loop.py @@ -1,17 +1,19 @@ """ DoomLoopMiddleware — pattern-based detector for repeated identical tool calls. -Mirrors ``opencode/packages/opencode/src/session/processor.ts`` doom-loop -behavior. When the same tool with the same arguments is called N times -in a row, the agent has likely entered an infinite loop. We surface this -to the user as an interrupt with ``permission="doom_loop"`` so the UI -can render an "Are you stuck? Continue / cancel?" affordance. +LangChain has :class:`ToolCallLimitMiddleware` which caps the *total* number +of tool calls per turn — but it can't tell apart "10 distinct, useful +calls" from "the same call 10 times in a row". This middleware fills that +gap with a sliding-window check on tool-call signatures, ported from +OpenCode's ``packages/opencode/src/session/processor.ts``. -Tier 1.11 in the OpenCode-port plan. +When the same tool with the same arguments is called N times in a row, +the agent has likely entered an infinite loop. We surface this to the +user as an interrupt with ``permission="doom_loop"`` so the UI can +render an "Are you stuck? Continue / cancel?" affordance. This ships **OFF by default** until the frontend explicitly handles -``context.permission == "doom_loop"`` interrupts (the plan flips -``SURFSENSE_ENABLE_DOOM_LOOP=true`` once the UI is ready). +``context.permission == "doom_loop"`` interrupts. Wire format: uses SurfSense's existing ``interrupt()`` payload shape (see ``app/agents/new_chat/tools/hitl.py``): @@ -69,7 +71,7 @@ class DoomLoopMiddleware(AgentMiddleware[AgentState[ResponseT], ContextT, Respon Args: threshold: How many consecutive identical signatures count as a - doom loop. Default 3 (opencode parity). + doom loop. Default 3 (matches OpenCode's processor.ts). """ def __init__(self, *, threshold: int = 3) -> None: @@ -182,7 +184,7 @@ class DoomLoopMiddleware(AgentMiddleware[AgentState[ResponseT], ContextT, Respon signatures[-1] if signatures else "", ) - # Tier 3b: interrupt.raised span with permission=doom_loop attribute + # Open an interrupt.raised span with permission=doom_loop attribute # so dashboards can break out doom-loop interrupts from regular # permission asks via the ``interrupt.permission`` attribute. with ot.interrupt_span( diff --git a/surfsense_backend/app/agents/new_chat/middleware/knowledge_search.py b/surfsense_backend/app/agents/new_chat/middleware/knowledge_search.py index 08ca8e18b..0820e8c3e 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/knowledge_search.py +++ b/surfsense_backend/app/agents/new_chat/middleware/knowledge_search.py @@ -592,10 +592,11 @@ class KnowledgePriorityMiddleware(AgentMiddleware): # type: ignore[type-arg] self.available_document_types = available_document_types self.top_k = top_k self.mentioned_document_ids = mentioned_document_ids or [] - # Tier 4.2: build the kb-planner private Runnable ONCE here so we - # don't pay the create_agent compile cost (50-200ms) on every turn. - # Disabled by default behind ``enable_kb_planner_runnable``; when off - # the planner falls back to the legacy ``self.llm.ainvoke`` path. + # Build the kb-planner private Runnable ONCE here so we don't pay + # the ``create_agent`` compile cost (50-200ms) on every turn. + # Disabled by default behind ``enable_kb_planner_runnable``; when + # off the planner falls back to the legacy ``self.llm.ainvoke`` + # path. self._planner: Runnable | None = None self._planner_compile_failed = False @@ -608,9 +609,9 @@ class KnowledgePriorityMiddleware(AgentMiddleware): # type: ignore[type-arg] lazily on first call, then memoized via ``self._planner``. The compiled agent is constructed without tools — the planner's - contract is "answer with structured JSON" — but with ``RetryAfter`` - + the OpenCode-port retry/limit middleware so it shares the parent - agent's resilience guarantees. + contract is "answer with structured JSON" — but it inherits the + :class:`RetryAfterMiddleware` so transient rate-limit errors + from the planner LLM call don't fail the whole turn. """ if self._planner is not None or self._planner_compile_failed: return self._planner @@ -658,9 +659,9 @@ class KnowledgePriorityMiddleware(AgentMiddleware): # type: ignore[type-arg] loop = asyncio.get_running_loop() t0 = loop.time() - # Tier 4.2: prefer the compiled-once planner Runnable when enabled; - # otherwise fall back to ``self.llm.ainvoke``. The ``surfsense:internal`` - # tag is preserved on both paths so ``_stream_agent_events`` still + # Prefer the compiled-once planner Runnable when enabled; otherwise + # fall back to ``self.llm.ainvoke``. The ``surfsense:internal`` tag + # is preserved on both paths so ``_stream_agent_events`` still # suppresses the planner's intermediate events from the UI. planner = self._build_kb_planner_runnable() try: diff --git a/surfsense_backend/app/agents/new_chat/middleware/noop_injection.py b/surfsense_backend/app/agents/new_chat/middleware/noop_injection.py index 8628479c7..503c73ccc 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/noop_injection.py +++ b/surfsense_backend/app/agents/new_chat/middleware/noop_injection.py @@ -1,18 +1,23 @@ """ ``_noop`` provider-compatibility tool + injection middleware. -OpenCode injects a ``_noop`` tool for LiteLLM/Bedrock/Copilot when the -model call has empty tools but message history includes prior -``tool_calls`` — some providers 400 in that shape (see -``opencode/packages/opencode/src/session/llm.ts:209-228``). SurfSense uses -LiteLLM, and the compaction summarize call (no tools, history full of -tool calls) hits this. Tier 1.5 in the OpenCode-port plan. +Some providers (LiteLLM, Bedrock, Copilot) 400 when a model call has +empty ``tools`` but the message history includes prior ``tool_calls`` — +they treat that shape as malformed even though it's perfectly valid +LangChain. SurfSense hits this on the compaction summarize call (no +tools, history full of tool calls). + +Ported from OpenCode's ``packages/opencode/src/session/llm.ts:209-228``, +which discovered and codified the workaround: inject a no-op tool *only* +on those provider shapes so the request validates without ever being +called. Operation: a :class:`NoopInjectionMiddleware` ``wrap_model_call`` checks if the request has zero tools but the last AI message in history includes -``tool_calls``. If yes, it injects the ``_noop`` tool only — never globally, -mirroring opencode's gating exactly. The :func:`noop_tool` returns empty -content when called (which it should never be in practice). +``tool_calls``. If yes, it injects the ``_noop`` tool only — never +globally — mirroring OpenCode's gating exactly. The :func:`noop_tool` +returns empty content when called (which it should never be in +practice). """ from __future__ import annotations @@ -45,8 +50,9 @@ def noop_tool() -> str: # Provider markers that benefit from ``_noop`` injection. These match -# opencode's gating list. We also accept any string containing one of -# these substrings (so e.g. ``litellm`` matches ``ChatLiteLLM``). +# OpenCode's gating list (``llm.ts:209-228``). We also accept any string +# containing one of these substrings so e.g. ``litellm`` matches +# ``ChatLiteLLM``. _NOOP_NEEDED_PROVIDERS: tuple[str, ...] = ( "litellm", "bedrock", diff --git a/surfsense_backend/app/agents/new_chat/middleware/otel_span.py b/surfsense_backend/app/agents/new_chat/middleware/otel_span.py index f51d2f7bb..cfe1edae4 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/otel_span.py +++ b/surfsense_backend/app/agents/new_chat/middleware/otel_span.py @@ -3,14 +3,14 @@ OpenTelemetry span middleware for the SurfSense ``new_chat`` agent. Wraps both ``model.call`` (LLM invocations) and ``tool.call`` (tool executions) with OTel spans, attaching low-cardinality span names and -high-cardinality identifiers as attributes (per the Tier 3b plan). +high-cardinality identifiers as attributes. This middleware is intentionally a thin adapter over :mod:`app.observability.otel`; when OTel is not configured all spans collapse to no-ops and the wrapper adds <1µs overhead per call. When OTel **is** configured (``OTEL_EXPORTER_OTLP_ENDPOINT`` set), every -model and tool call gets a span with the standard attributes the -plan's dashboards expect. +model and tool call gets a span with the standard attributes our +dashboards expect. """ from __future__ import annotations diff --git a/surfsense_backend/app/agents/new_chat/middleware/permission.py b/surfsense_backend/app/agents/new_chat/middleware/permission.py index 6e1f42baf..37719e96a 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/permission.py +++ b/surfsense_backend/app/agents/new_chat/middleware/permission.py @@ -1,10 +1,15 @@ """ PermissionMiddleware — pattern-based allow/deny/ask with HITL fallback. -Mirrors ``opencode/packages/opencode/src/permission/index.ts`` but uses -SurfSense's existing ``interrupt({type, action, context})`` payload shape -(see ``app/agents/new_chat/tools/hitl.py``) so the frontend keeps -working unchanged. Tier 2.1 in the OpenCode-port plan. +LangChain's :class:`HumanInTheLoopMiddleware` only supports a static +"this tool always asks" decision per tool. There's no rule-based +allow/deny/ask layered ruleset, no glob patterns, no per-search-space or +per-thread overrides, and no auto-deny synthesis. + +This middleware ports OpenCode's ``packages/opencode/src/permission/index.ts`` +ruleset model on top of SurfSense's existing ``interrupt({type, action, +context})`` payload shape (see ``app/agents/new_chat/tools/hitl.py``) so +the frontend keeps working unchanged. Operation: 1. ``aafter_model`` inspects the latest ``AIMessage.tool_calls``. @@ -24,9 +29,9 @@ Operation: The middleware also performs a *pre-model* tool-filter step (the ``before_model`` hook) so globally denied tools are stripped from the -exposed tool list before the model gets to see them. This is -opencode's ``Permission.disabled`` equivalent and dramatically reduces -the chance the model emits a deny-only call. +exposed tool list before the model gets to see them. This mirrors +OpenCode's ``Permission.disabled`` and dramatically reduces the chance +the model emits a deny-only call. """ from __future__ import annotations @@ -117,7 +122,7 @@ class PermissionMiddleware(AgentMiddleware): # type: ignore[type-arg] self._emit_interrupt = always_emit_interrupt_payload # ------------------------------------------------------------------ - # Tool-filter step (opencode `Permission.disabled` equivalent) + # Tool-filter step (mirrors OpenCode's ``Permission.disabled``) # ------------------------------------------------------------------ def _globally_denied(self, tool_name: str) -> bool: @@ -197,8 +202,8 @@ class PermissionMiddleware(AgentMiddleware): # type: ignore[type-arg] "always": patterns, }, } - # Tier 3b: permission.asked + interrupt.raised spans (no-op when - # OTel is disabled). Both fire here so dashboards can correlate + # Open ``permission.asked`` + ``interrupt.raised`` OTel spans + # (no-op when OTel is disabled) so dashboards can correlate # "we asked X" with "interrupt was actually delivered". with ( ot.permission_asked_span( diff --git a/surfsense_backend/app/agents/new_chat/middleware/retry_after.py b/surfsense_backend/app/agents/new_chat/middleware/retry_after.py index 394bb0371..0c3d3d017 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/retry_after.py +++ b/surfsense_backend/app/agents/new_chat/middleware/retry_after.py @@ -1,10 +1,16 @@ """ RetryAfterMiddleware — Header-aware retry with custom backoff and SSE eventing. -Why standalone instead of subclassing ``ModelRetryMiddleware``: the upstream -class calls module-level ``calculate_delay`` inline (no overridable -``_calculate_delay`` hook), so a subclass cannot inject Retry-After header -delays without rewriting the loop. Tier 1.4 in the OpenCode-port plan. +LangChain's :class:`ModelRetryMiddleware` retries on exceptions but ignores +the ``Retry-After`` HTTP header — it just runs its own exponential backoff. +That wastes time when a provider has explicitly told us how long to wait. +This middleware honors the header (mirroring OpenCode's +``packages/opencode/src/session/llm.ts`` retry pathway) and emits an SSE +event so the UI can show "rate-limited, retrying in Ns". + +We can't subclass ``ModelRetryMiddleware`` cleanly because its loop calls a +module-level ``calculate_delay`` inline (no overridable +``_calculate_delay`` hook), so this is a standalone implementation. Behaviour: - Extracts ``Retry-After`` / ``retry-after-ms`` from diff --git a/surfsense_backend/app/agents/new_chat/middleware/tool_call_repair.py b/surfsense_backend/app/agents/new_chat/middleware/tool_call_repair.py index 54df0cc60..9f81a168b 100644 --- a/surfsense_backend/app/agents/new_chat/middleware/tool_call_repair.py +++ b/surfsense_backend/app/agents/new_chat/middleware/tool_call_repair.py @@ -1,10 +1,6 @@ """ ToolCallNameRepairMiddleware — two-stage tool-name repair. -Mirrors ``opencode/packages/opencode/src/session/llm.ts:339-358`` plus -``opencode/packages/opencode/src/tool/invalid.ts``. Tier 1.7 in the -OpenCode-port plan. - Operation: 1. **Stage 1 — lowercase repair:** if a tool call's ``name`` is not in the registry but ``name.lower()`` is, rewrite in place. Catches @@ -14,9 +10,13 @@ Operation: so the registered :func:`invalid_tool` returns the error to the model for self-correction. -Distinct from :class:`deepagents.middleware.PatchToolCallsMiddleware`, -which patches *dangling* tool calls (no matching ToolMessage) — that -class does not handle the wrong-name case at all. +Ported from OpenCode's ``packages/opencode/src/session/llm.ts:339-358`` ++ ``packages/opencode/src/tool/invalid.ts``. LangChain has no equivalent: +:class:`deepagents.middleware.PatchToolCallsMiddleware` patches +*dangling* tool calls (no matching ToolMessage) but does nothing about +wrong names, and the model framework's default behavior on an unknown +name is to crash the turn rather than route to a self-correction +fallback. """ from __future__ import annotations @@ -61,7 +61,8 @@ class ToolCallNameRepairMiddleware( ``invalid`` should be in this set so the fallback dispatches. fuzzy_match_threshold: Optional ``difflib`` ratio (0-1) for the fuzzy-match step that runs *between* lowercase and invalid. - Set to ``None`` to disable fuzzy matching (opencode parity). + Set to ``None`` to disable fuzzy matching (default in + OpenCode; we mirror that to avoid silent rewrites). """ def __init__( @@ -106,7 +107,7 @@ class ToolCallNameRepairMiddleware( call["response_metadata"] = metadata return call - # Optional fuzzy step (off by default for opencode parity) + # Optional fuzzy step (off by default — see class docstring) if self._fuzzy_threshold is not None: close = difflib.get_close_matches( name, registered, n=1, cutoff=self._fuzzy_threshold diff --git a/surfsense_backend/app/agents/new_chat/permissions.py b/surfsense_backend/app/agents/new_chat/permissions.py index 50a0cfbdc..523deb11f 100644 --- a/surfsense_backend/app/agents/new_chat/permissions.py +++ b/surfsense_backend/app/agents/new_chat/permissions.py @@ -1,21 +1,20 @@ """ Wildcard pattern matching + rule evaluation for the SurfSense permission system. -Mirrors ``opencode/packages/opencode/src/permission/evaluate.ts`` and -``opencode/packages/opencode/src/util/wildcard.ts`` precisely: +Ported from OpenCode's ``packages/opencode/src/permission/evaluate.ts`` and +``packages/opencode/src/util/wildcard.ts``. LangChain has no rule-based +permission evaluator, so we keep OpenCode's semantics intact: - ``Wildcard.match`` matches both the ``permission`` and the ``pattern`` fields of a rule against the requested ``(permission, pattern)`` pair. ``*`` matches any segment, ``**`` matches across separators. - The evaluator runs ``findLast`` over the **flattened** list of rules from all rulesets — last matching rule wins. -- The default fallback is ``ask`` (NOT deny), matching opencode. +- The default fallback is ``ask`` (NOT deny), matching OpenCode. - Multi-pattern requests AND together: if ANY pattern resolves to ``deny``, the whole request is denied; if ANY needs ``ask``, an interrupt is raised; only when all patterns ``allow`` does the request proceed. - -Tier 2.1 in the OpenCode-port plan. """ from __future__ import annotations diff --git a/surfsense_backend/app/agents/new_chat/plugin_loader.py b/surfsense_backend/app/agents/new_chat/plugin_loader.py index 426e28041..c52620d40 100644 --- a/surfsense_backend/app/agents/new_chat/plugin_loader.py +++ b/surfsense_backend/app/agents/new_chat/plugin_loader.py @@ -1,9 +1,10 @@ """Entry-point based plugin loader for SurfSense agent middleware. -The realization in the Tier 6 plan: LangChain's :class:`AgentMiddleware` ABC -already covers the practical surface most plugins need (``before_agent`` / -``before_model`` / ``wrap_tool_call`` / their async counterparts), so a -SurfSense-specific plugin protocol is unnecessary. +LangChain's :class:`AgentMiddleware` ABC already covers the practical +surface most plugins need (``before_agent`` / ``before_model`` / +``wrap_tool_call`` / their async counterparts), so a SurfSense-specific +plugin protocol would be redundant. We just need a way to discover and +admit third-party middleware safely. A plugin is therefore just an installable Python package that registers a factory callable under the ``surfsense.plugins`` entry-point group: diff --git a/surfsense_backend/app/agents/new_chat/plugins/year_substituter.py b/surfsense_backend/app/agents/new_chat/plugins/year_substituter.py index 3e2e631d2..2b7781b90 100644 --- a/surfsense_backend/app/agents/new_chat/plugins/year_substituter.py +++ b/surfsense_backend/app/agents/new_chat/plugins/year_substituter.py @@ -1,10 +1,10 @@ """Reference plugin: substitute ``{{year}}`` in tool descriptions. -Mirrors the OpenCode ``chat.system.transform`` example. Demonstrates the -:meth:`AgentMiddleware.awrap_tool_call` hook -- the plugin sees every tool -invocation and can rewrite the request *or* the result. This particular -plugin is read-only and only transforms the *description* the user might -see in error messages (no request mutation). +Demonstrates the :meth:`AgentMiddleware.awrap_tool_call` hook -- the +plugin sees every tool invocation and can rewrite the request *or* the +result. This particular plugin is read-only and only transforms the +*description* the user might see in error messages (no request +mutation). The plugin is built as a factory function so the entry-point loader can inject :class:`PluginContext` (containing the agent's LLM, search-space diff --git a/surfsense_backend/app/agents/new_chat/prompts/composer.py b/surfsense_backend/app/agents/new_chat/prompts/composer.py index 77b86aeef..42f8303e6 100644 --- a/surfsense_backend/app/agents/new_chat/prompts/composer.py +++ b/surfsense_backend/app/agents/new_chat/prompts/composer.py @@ -14,7 +14,13 @@ under :mod:`app.agents.new_chat.prompts`. It replaces the monolithic examples/ # one ``.md`` per tool with call examples routing/ # connector-specific routing notes (linear, slack, …) -Tier 3a in the OpenCode-port plan. +The model-family dispatch step (see :func:`detect_provider_variant`) +mirrors OpenCode's ``packages/opencode/src/session/system.ts`` — different +model families respond best to differently-styled prompts (Claude likes +XML/narrative, GPT-5 wants channel-aware pragmatic, Codex needs +terse/file:line, Gemini wants formal numbered steps, etc.). LangChain's +``dynamic_prompt`` helper supports per-call prompt swaps but ships no +out-of-the-box family classifier, so we keep our own. Backwards compatibility ======================= @@ -42,10 +48,11 @@ from app.db import ChatVisibility # When adding a new variant, also drop a matching ``providers/.md`` # file in this package and (if appropriate) extend the regex matchers below. # -# Stylistic clusters mirror OpenCode's prompt-per-family layout but adapted -# to SurfSense's "supplemental hints" architecture (each fragment is a -# focused style nudge, NOT a full system prompt — the main prompt is -# already assembled from base/ + tools/ + routing/). +# Stylistic clusters: each variant is a focused style nudge, NOT a full +# system prompt — the main prompt is already assembled from base/ + +# tools/ + routing/. The clustering itself (which models map to which +# style) follows OpenCode's ``system.ts`` family table; see the module +# docstring for credits. ProviderVariant = str # Known values: # "anthropic" — Claude family (XML-friendly, narrative todos) @@ -82,8 +89,8 @@ def detect_provider_variant(model_name: str | None) -> ProviderVariant: Order is significant: more-specific patterns are tried first so ``gpt-5-codex`` routes to ``"openai_codex"`` rather than - ``"openai_reasoning"`` (mirrors OpenCode's - ``packages/opencode/src/session/system.ts`` dispatch). + ``"openai_reasoning"`` — same dispatch order as OpenCode's + ``packages/opencode/src/session/system.ts``. """ if not model_name: return "default" diff --git a/surfsense_backend/app/agents/new_chat/subagents/__init__.py b/surfsense_backend/app/agents/new_chat/subagents/__init__.py index b9f21a0d2..7d678ec79 100644 --- a/surfsense_backend/app/agents/new_chat/subagents/__init__.py +++ b/surfsense_backend/app/agents/new_chat/subagents/__init__.py @@ -1,14 +1,17 @@ """Specialized user-facing subagents for the SurfSense agent. -Each subagent is a :class:`deepagents.SubAgent` typed-dict spec passed to -:class:`deepagents.SubAgentMiddleware`, which materializes them as ephemeral -runnables invoked via the ``task`` tool. +The :class:`deepagents.SubAgentMiddleware` already provides the +materialization machinery (each :class:`deepagents.SubAgent` typed-dict +spec is compiled into an ephemeral runnable invoked via the ``task`` +tool); what's specific to SurfSense is the *seeding* of those subagents +with declarative deny rules. Per-subagent permission rules are injected as a :class:`PermissionMiddleware` entry inside the subagent's ``middleware`` -field, mirroring opencode ``tool/task.ts`` which seeds child sessions with -deny rules for tools the parent does not want them touching (e.g. -``task``/``todowrite`` recursion, write tools for read-only research roles). +field. The auto-deny pattern (e.g. forbid ``task``/``todowrite`` +recursion, block write tools for read-only research roles) is borrowed +from OpenCode's ``packages/opencode/src/tool/task.ts``, which has +analogous logic for restricting child sessions. """ from .config import ( diff --git a/surfsense_backend/app/agents/new_chat/system_prompt.py b/surfsense_backend/app/agents/new_chat/system_prompt.py index 3919527d9..56f838d7e 100644 --- a/surfsense_backend/app/agents/new_chat/system_prompt.py +++ b/surfsense_backend/app/agents/new_chat/system_prompt.py @@ -1,13 +1,14 @@ """ Thin compatibility wrapper around :mod:`app.agents.new_chat.prompts.composer`. -Tier 3a of the OpenCode-port plan replaced the monolithic prompt strings -in this module with a fragment tree under ``prompts/`` and a composer -function. This module preserves the public function surface -(``build_surfsense_system_prompt`` / ``build_configurable_system_prompt`` / -``get_default_system_instructions`` / ``SURFSENSE_SYSTEM_PROMPT``) so that -existing call sites — `chat_deepagent.py`, anonymous chat routes, and the -configurable-prompt admin path — keep working without churn. +The composer split the previous monolithic prompt string into a fragment +tree under ``prompts/`` plus a model-family dispatch step (see the +composer module docstring for credits). This module preserves the public +function surface (``build_surfsense_system_prompt`` / +``build_configurable_system_prompt`` / +``get_default_system_instructions`` / ``SURFSENSE_SYSTEM_PROMPT``) so +that existing call sites — `chat_deepagent.py`, anonymous chat routes, +and the configurable-prompt admin path — keep working without churn. For new call sites prefer importing ``compose_system_prompt`` directly from :mod:`app.agents.new_chat.prompts.composer`. diff --git a/surfsense_backend/app/agents/new_chat/tools/invalid_tool.py b/surfsense_backend/app/agents/new_chat/tools/invalid_tool.py index df10fcbe3..ea4bc0bc1 100644 --- a/surfsense_backend/app/agents/new_chat/tools/invalid_tool.py +++ b/surfsense_backend/app/agents/new_chat/tools/invalid_tool.py @@ -6,8 +6,9 @@ tool, :class:`ToolCallNameRepairMiddleware` rewrites the call to ``invalid`` with the original name and a parser/validation error string. This tool's execution then returns that error to the model so it can self-correct. -Mirrors ``opencode/packages/opencode/src/tool/invalid.ts``. Tier 1.6 in -the OpenCode-port plan. +Ported from OpenCode's ``packages/opencode/src/tool/invalid.ts`` — +LangChain has no equivalent fallback path; the default behavior on an +unknown tool name is a hard ``ToolNotFoundError`` which kills the turn. Critically, the :class:`ToolDefinition` for this tool is **excluded** from the system-prompt tool list and from ``LLMToolSelectorMiddleware`` selection diff --git a/surfsense_backend/app/agents/new_chat/tools/registry.py b/surfsense_backend/app/agents/new_chat/tools/registry.py index fce1bf872..e8bab36fd 100644 --- a/surfsense_backend/app/agents/new_chat/tools/registry.py +++ b/surfsense_backend/app/agents/new_chat/tools/registry.py @@ -132,12 +132,10 @@ class ToolDefinition: that must be in ``available_connectors`` for the tool to be enabled. dedup_key: Optional callable that maps a tool's ``args`` dict to a string signature used by :class:`DedupHITLToolCallsMiddleware` - to drop duplicate calls. Replaces the legacy hardcoded - ``_NATIVE_HITL_TOOL_DEDUP_KEYS`` map (Tier 2.3 in the - OpenCode-port plan). + to drop duplicate calls within a single LLM response. reverse: Optional callable that, given the tool's ``(args, result)``, returns a ``ReverseDescriptor`` describing the inverse tool - invocation. Consumed by the snapshot/revert pipeline (Tier 5). + invocation. Consumed by the snapshot/revert pipeline. """ diff --git a/surfsense_backend/app/observability/otel.py b/surfsense_backend/app/observability/otel.py index 4f2257ab7..6791ab499 100644 --- a/surfsense_backend/app/observability/otel.py +++ b/surfsense_backend/app/observability/otel.py @@ -1,12 +1,10 @@ """ OpenTelemetry instrumentation helpers for the SurfSense agent stack. -Tier 3b in the OpenCode-port plan. - Goals ===== -- Provide one tiny, ergonomic API for the spans listed in the plan +- Provide one tiny, ergonomic API for the spans we care about (``tool.call``, ``model.call``, ``kb.search``, ``kb.persist``, ``compaction.run``, ``interrupt.raised``, ``permission.asked``). - Keep span **names** low-cardinality (``tool.call`` rather than diff --git a/surfsense_backend/app/routes/agent_revert_route.py b/surfsense_backend/app/routes/agent_revert_route.py index cbe4e7417..12484ff53 100644 --- a/surfsense_backend/app/routes/agent_revert_route.py +++ b/surfsense_backend/app/routes/agent_revert_route.py @@ -1,9 +1,9 @@ """POST ``/api/threads/{thread_id}/revert/{action_id}``: undo an agent action. -Per the Tier 5 plan, the route ships **before** the UI lights up the per-message -"Undo from here" affordance. To prevent accidental usage during the gap we -return ``503 Service Unavailable`` until the -``SURFSENSE_ENABLE_REVERT_ROUTE`` flag flips. Once enabled, the route runs: +The route ships **before** the UI lights up the per-message "Undo from +here" affordance. To prevent accidental usage during the gap we return +``503 Service Unavailable`` until the ``SURFSENSE_ENABLE_REVERT_ROUTE`` +flag flips. Once enabled, the route runs: 1. Authentication via :func:`current_active_user`. 2. Action lookup; 404 if the action does not belong to the thread. diff --git a/surfsense_backend/tests/unit/agents/new_chat/prompts/test_composer.py b/surfsense_backend/tests/unit/agents/new_chat/prompts/test_composer.py index aa0c215b9..397b1c787 100644 --- a/surfsense_backend/tests/unit/agents/new_chat/prompts/test_composer.py +++ b/surfsense_backend/tests/unit/agents/new_chat/prompts/test_composer.py @@ -1,4 +1,4 @@ -"""Tests for the prompt fragment composer (Tier 3a).""" +"""Tests for the prompt fragment composer.""" from __future__ import annotations diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_otel_span.py b/surfsense_backend/tests/unit/agents/new_chat/test_otel_span.py index e5b171612..55434c04d 100644 --- a/surfsense_backend/tests/unit/agents/new_chat/test_otel_span.py +++ b/surfsense_backend/tests/unit/agents/new_chat/test_otel_span.py @@ -1,4 +1,4 @@ -"""Tests for the OtelSpanMiddleware adapter (Tier 3b).""" +"""Tests for the OtelSpanMiddleware adapter.""" from __future__ import annotations diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_permissions.py b/surfsense_backend/tests/unit/agents/new_chat/test_permissions.py index 4924f2aee..8ec16617a 100644 --- a/surfsense_backend/tests/unit/agents/new_chat/test_permissions.py +++ b/surfsense_backend/tests/unit/agents/new_chat/test_permissions.py @@ -1,4 +1,4 @@ -"""Tests for the wildcard matcher and rule evaluator (opencode evaluate.ts parity).""" +"""Tests for the wildcard matcher and rule evaluator (parity with OpenCode evaluate.ts).""" from __future__ import annotations diff --git a/surfsense_backend/tests/unit/agents/new_chat/test_plugin_loader.py b/surfsense_backend/tests/unit/agents/new_chat/test_plugin_loader.py index c2118c697..5dbf765a7 100644 --- a/surfsense_backend/tests/unit/agents/new_chat/test_plugin_loader.py +++ b/surfsense_backend/tests/unit/agents/new_chat/test_plugin_loader.py @@ -1,4 +1,4 @@ -"""Unit tests for the SurfSense plugin entry-point loader (Tier 6).""" +"""Unit tests for the SurfSense plugin entry-point loader.""" from __future__ import annotations diff --git a/surfsense_backend/tests/unit/observability/test_otel.py b/surfsense_backend/tests/unit/observability/test_otel.py index 583142098..fc5813973 100644 --- a/surfsense_backend/tests/unit/observability/test_otel.py +++ b/surfsense_backend/tests/unit/observability/test_otel.py @@ -1,4 +1,4 @@ -"""Tests for the SurfSense OpenTelemetry shim (Tier 3b).""" +"""Tests for the SurfSense OpenTelemetry shim.""" from __future__ import annotations diff --git a/surfsense_backend/tests/unit/services/test_revert_service.py b/surfsense_backend/tests/unit/services/test_revert_service.py index e2cbe383a..a81e52041 100644 --- a/surfsense_backend/tests/unit/services/test_revert_service.py +++ b/surfsense_backend/tests/unit/services/test_revert_service.py @@ -1,4 +1,4 @@ -"""Unit tests for the agent revert service (Tier 5.3).""" +"""Unit tests for the agent revert service.""" from __future__ import annotations