diff --git a/docker-compose.yml b/docker-compose.yml index 0794998f0..e5989210f 100644 --- a/docker-compose.yml +++ b/docker-compose.yml @@ -45,6 +45,8 @@ services: volumes: - ./surfsense_backend/app:/app/app - shared_temp:/tmp + # Uncomment and edit the line below to enable Obsidian vault indexing + # - /path/to/your/obsidian/vault:/obsidian-vault:ro env_file: - ./surfsense_backend/.env environment: diff --git a/surfsense_backend/alembic/versions/74_no_op.py b/surfsense_backend/alembic/versions/74_no_op.py new file mode 100644 index 000000000..a5ee99b29 --- /dev/null +++ b/surfsense_backend/alembic/versions/74_no_op.py @@ -0,0 +1,29 @@ +"""No-op migration for Composio support + +Revision ID: 74 +Revises: 73 +Create Date: 2026-01-21 + +NOTE: This migration is a no-op since Composio is not supported yet. +""" + +from collections.abc import Sequence + +# revision identifiers, used by Alembic. +revision: str = "74" +down_revision: str | None = "73" +branch_labels: str | Sequence[str] | None = None +depends_on: str | Sequence[str] | None = None + + +def upgrade() -> None: + """No-op upgrade for Composio support.""" + pass + + +def downgrade() -> None: + """No-op downgrade for Composio support. + + Note: PostgreSQL does not support removing enum values directly. + """ + pass diff --git a/surfsense_backend/alembic/versions/75_add_chat_session_state_table.py b/surfsense_backend/alembic/versions/75_add_chat_session_state_table.py new file mode 100644 index 000000000..46bf7b9b9 --- /dev/null +++ b/surfsense_backend/alembic/versions/75_add_chat_session_state_table.py @@ -0,0 +1,75 @@ +"""Add chat_session_state table for live collaboration + +Revision ID: 75 +Revises: 74 + +Creates chat_session_state table to track AI responding state per thread. +Enables real-time sync via Electric SQL for shared chat collaboration. +""" + +from collections.abc import Sequence + +from alembic import op + +revision: str = "75" +down_revision: str | None = "74" +branch_labels: str | Sequence[str] | None = None +depends_on: str | Sequence[str] | None = None + + +def upgrade() -> None: + """Create chat_session_state table with Electric SQL replication.""" + op.execute( + """ + CREATE TABLE IF NOT EXISTS chat_session_state ( + id SERIAL PRIMARY KEY, + thread_id INTEGER NOT NULL REFERENCES new_chat_threads(id) ON DELETE CASCADE, + ai_responding_to_user_id UUID REFERENCES "user"(id) ON DELETE SET NULL, + updated_at TIMESTAMPTZ NOT NULL DEFAULT NOW(), + UNIQUE (thread_id) + ) + """ + ) + + op.execute( + "CREATE INDEX IF NOT EXISTS idx_chat_session_state_thread_id ON chat_session_state(thread_id)" + ) + + op.execute("ALTER TABLE chat_session_state REPLICA IDENTITY FULL;") + + op.execute( + """ + DO $$ + BEGIN + IF NOT EXISTS ( + SELECT 1 FROM pg_publication_tables + WHERE pubname = 'electric_publication_default' + AND tablename = 'chat_session_state' + ) THEN + ALTER PUBLICATION electric_publication_default ADD TABLE chat_session_state; + END IF; + END + $$; + """ + ) + + +def downgrade() -> None: + """Drop chat_session_state table and remove from Electric SQL replication.""" + op.execute( + """ + DO $$ + BEGIN + IF EXISTS ( + SELECT 1 FROM pg_publication_tables + WHERE pubname = 'electric_publication_default' + AND tablename = 'chat_session_state' + ) THEN + ALTER PUBLICATION electric_publication_default DROP TABLE chat_session_state; + END IF; + END + $$; + """ + ) + + op.execute("DROP TABLE IF EXISTS chat_session_state;") diff --git a/surfsense_backend/alembic/versions/76_add_live_collaboration_tables_electric_replication.py b/surfsense_backend/alembic/versions/76_add_live_collaboration_tables_electric_replication.py new file mode 100644 index 000000000..68bca4fc1 --- /dev/null +++ b/surfsense_backend/alembic/versions/76_add_live_collaboration_tables_electric_replication.py @@ -0,0 +1,99 @@ +"""Add live collaboration tables to Electric SQL publication + +Revision ID: 76 +Revises: 75 + +Enables real-time sync for live collaboration features: +- new_chat_messages: Live message sync between users +- chat_comments: Live comment updates + +Note: User/member info is fetched via API (membersAtom) for client-side joins, +not via Electric SQL, to keep where clauses optimized and reduce complexity. +""" + +from collections.abc import Sequence + +from alembic import op + +revision: str = "76" +down_revision: str | None = "75" +branch_labels: str | Sequence[str] | None = None +depends_on: str | Sequence[str] | None = None + + +def upgrade() -> None: + """Add live collaboration tables to Electric SQL replication.""" + # Set REPLICA IDENTITY FULL for Electric SQL sync + op.execute("ALTER TABLE new_chat_messages REPLICA IDENTITY FULL;") + op.execute("ALTER TABLE chat_comments REPLICA IDENTITY FULL;") + + # Add new_chat_messages to Electric publication + op.execute( + """ + DO $$ + BEGIN + IF NOT EXISTS ( + SELECT 1 FROM pg_publication_tables + WHERE pubname = 'electric_publication_default' + AND tablename = 'new_chat_messages' + ) THEN + ALTER PUBLICATION electric_publication_default ADD TABLE new_chat_messages; + END IF; + END + $$; + """ + ) + + # Add chat_comments to Electric publication + op.execute( + """ + DO $$ + BEGIN + IF NOT EXISTS ( + SELECT 1 FROM pg_publication_tables + WHERE pubname = 'electric_publication_default' + AND tablename = 'chat_comments' + ) THEN + ALTER PUBLICATION electric_publication_default ADD TABLE chat_comments; + END IF; + END + $$; + """ + ) + + +def downgrade() -> None: + """Remove live collaboration tables from Electric SQL replication.""" + op.execute( + """ + DO $$ + BEGIN + IF EXISTS ( + SELECT 1 FROM pg_publication_tables + WHERE pubname = 'electric_publication_default' + AND tablename = 'new_chat_messages' + ) THEN + ALTER PUBLICATION electric_publication_default DROP TABLE new_chat_messages; + END IF; + END + $$; + """ + ) + + op.execute( + """ + DO $$ + BEGIN + IF EXISTS ( + SELECT 1 FROM pg_publication_tables + WHERE pubname = 'electric_publication_default' + AND tablename = 'chat_comments' + ) THEN + ALTER PUBLICATION electric_publication_default DROP TABLE chat_comments; + END IF; + END + $$; + """ + ) + + # Note: Not reverting REPLICA IDENTITY as it doesn't harm normal operations diff --git a/surfsense_backend/alembic/versions/77_add_thread_id_to_chat_comments.py b/surfsense_backend/alembic/versions/77_add_thread_id_to_chat_comments.py new file mode 100644 index 000000000..86886eacf --- /dev/null +++ b/surfsense_backend/alembic/versions/77_add_thread_id_to_chat_comments.py @@ -0,0 +1,70 @@ +"""Add thread_id to chat_comments for denormalized Electric subscriptions + +This denormalization allows a single Electric SQL subscription per thread +instead of one per message, significantly reducing connection overhead. + +Revision ID: 77 +Revises: 76 +""" + +from collections.abc import Sequence + +from alembic import op + +revision: str = "77" +down_revision: str | None = "76" +branch_labels: str | Sequence[str] | None = None +depends_on: str | Sequence[str] | None = None + + +def upgrade() -> None: + """Add thread_id column to chat_comments and backfill from messages.""" + # Add the column (nullable initially for backfill) + op.execute( + """ + ALTER TABLE chat_comments + ADD COLUMN IF NOT EXISTS thread_id INTEGER; + """ + ) + + # Backfill thread_id from the related message + op.execute( + """ + UPDATE chat_comments c + SET thread_id = m.thread_id + FROM new_chat_messages m + WHERE c.message_id = m.id + AND c.thread_id IS NULL; + """ + ) + + # Make it NOT NULL after backfill + op.execute( + """ + ALTER TABLE chat_comments + ALTER COLUMN thread_id SET NOT NULL; + """ + ) + + # Add FK constraint + op.execute( + """ + ALTER TABLE chat_comments + ADD CONSTRAINT fk_chat_comments_thread_id + FOREIGN KEY (thread_id) REFERENCES new_chat_threads(id) ON DELETE CASCADE; + """ + ) + + # Add index for efficient Electric subscriptions by thread + op.execute( + "CREATE INDEX IF NOT EXISTS idx_chat_comments_thread_id ON chat_comments(thread_id)" + ) + + +def downgrade() -> None: + """Remove thread_id column from chat_comments.""" + op.execute("DROP INDEX IF EXISTS idx_chat_comments_thread_id") + op.execute( + "ALTER TABLE chat_comments DROP CONSTRAINT IF EXISTS fk_chat_comments_thread_id" + ) + op.execute("ALTER TABLE chat_comments DROP COLUMN IF EXISTS thread_id") diff --git a/surfsense_backend/alembic/versions/78_add_obsidian_connector.py b/surfsense_backend/alembic/versions/78_add_obsidian_connector.py new file mode 100644 index 000000000..a2a5037a8 --- /dev/null +++ b/surfsense_backend/alembic/versions/78_add_obsidian_connector.py @@ -0,0 +1,33 @@ +"""Add Obsidian connector enums + +Revision ID: 78 +Revises: 77 +Create Date: 2026-01-21 + +""" + +from collections.abc import Sequence + +from alembic import op + +# revision identifiers, used by Alembic. +revision: str = "78" +down_revision: str | None = "77" +branch_labels: str | Sequence[str] | None = None +depends_on: str | Sequence[str] | None = None + + +def upgrade() -> None: + # Add OBSIDIAN_CONNECTOR to documenttype enum + op.execute("ALTER TYPE documenttype ADD VALUE IF NOT EXISTS 'OBSIDIAN_CONNECTOR'") + + # Add OBSIDIAN_CONNECTOR to searchsourceconnectortype enum + op.execute( + "ALTER TYPE searchsourceconnectortype ADD VALUE IF NOT EXISTS 'OBSIDIAN_CONNECTOR'" + ) + + +def downgrade() -> None: + # Note: PostgreSQL doesn't support removing enum values directly. + # The values will remain in the enum type but won't be used. + pass diff --git a/surfsense_backend/alembic/versions/74_add_composio_connector_enums.py b/surfsense_backend/alembic/versions/79_add_composio_connector_enums.py similarity index 96% rename from surfsense_backend/alembic/versions/74_add_composio_connector_enums.py rename to surfsense_backend/alembic/versions/79_add_composio_connector_enums.py index 2996d9d07..0869fdcc3 100644 --- a/surfsense_backend/alembic/versions/74_add_composio_connector_enums.py +++ b/surfsense_backend/alembic/versions/79_add_composio_connector_enums.py @@ -1,8 +1,7 @@ """Add Composio connector types to SearchSourceConnectorType and DocumentType enums -Revision ID: 74 -Revises: 73 -Create Date: 2026-01-21 +Revision ID: 79 +Revises: 78 This migration adds the Composio connector enum values to both: - searchsourceconnectortype (for connector type tracking) @@ -23,8 +22,8 @@ from collections.abc import Sequence from alembic import op # revision identifiers, used by Alembic. -revision: str = "74" -down_revision: str | None = "73" +revision: str = "79" +down_revision: str | None = "78" branch_labels: str | Sequence[str] | None = None depends_on: str | Sequence[str] | None = None diff --git a/surfsense_backend/app/agents/new_chat/system_prompt.py b/surfsense_backend/app/agents/new_chat/system_prompt.py index d8202a8b0..fdb80acb9 100644 --- a/surfsense_backend/app/agents/new_chat/system_prompt.py +++ b/surfsense_backend/app/agents/new_chat/system_prompt.py @@ -34,6 +34,12 @@ You have access to the following tools: - Returns: Documentation content with chunk IDs for citations (prefixed with 'doc-', e.g., [citation:doc-123]) 1. search_knowledge_base: Search the user's personal knowledge base for relevant information. + - IMPORTANT: When searching for information (meetings, schedules, notes, tasks, etc.), ALWAYS search broadly + across ALL sources first by omitting connectors_to_search. The user may store information in various places + including calendar apps, note-taking apps (Obsidian, Notion), chat apps (Slack, Discord), and more. + - Only narrow to specific connectors if the user explicitly asks (e.g., "check my Slack" or "in my calendar"). + - Personal notes in Obsidian, Notion, or NOTE often contain schedules, meeting times, reminders, and other + important information that may not be in calendars. - Args: - query: The search query - be specific and include key terms - top_k: Number of results to retrieve (default: 10) @@ -157,6 +163,13 @@ You have access to the following tools: stating "Based on your memory..." - integrate the context seamlessly. +- User: "What time is the team meeting today?" + - Call: `search_knowledge_base(query="team meeting time today")` (searches ALL sources - calendar, notes, Obsidian, etc.) + - DO NOT limit to just calendar - the info might be in notes! + +- User: "When is my gym session?" + - Call: `search_knowledge_base(query="gym session time schedule")` (searches ALL sources) + - User: "How do I install SurfSense?" - Call: `search_surfsense_docs(query="installation setup")` @@ -175,6 +188,12 @@ You have access to the following tools: - User: "What did I discuss on Slack last week about the React migration?" - Call: `search_knowledge_base(query="React migration", connectors_to_search=["SLACK_CONNECTOR"], start_date="YYYY-MM-DD", end_date="YYYY-MM-DD")` +- User: "Check my Obsidian notes for meeting notes" + - Call: `search_knowledge_base(query="meeting notes", connectors_to_search=["OBSIDIAN_CONNECTOR"])` + +- User: "What's in my Obsidian vault about project ideas?" + - Call: `search_knowledge_base(query="project ideas", connectors_to_search=["OBSIDIAN_CONNECTOR"])` + - User: "Remember that I prefer TypeScript over JavaScript" - Call: `save_memory(content="User prefers TypeScript over JavaScript for development", category="preference")` diff --git a/surfsense_backend/app/agents/new_chat/tools/knowledge_base.py b/surfsense_backend/app/agents/new_chat/tools/knowledge_base.py index e91d865fa..552019dda 100644 --- a/surfsense_backend/app/agents/new_chat/tools/knowledge_base.py +++ b/surfsense_backend/app/agents/new_chat/tools/knowledge_base.py @@ -49,6 +49,7 @@ _ALL_CONNECTORS: list[str] = [ "BOOKSTACK_CONNECTOR", "CRAWLED_URL", "CIRCLEBACK", + "OBSIDIAN_CONNECTOR", ] @@ -508,6 +509,16 @@ async def search_knowledge_base_async( ) all_documents.extend(chunks) + elif connector == "OBSIDIAN_CONNECTOR": + _, chunks = await connector_service.search_obsidian( + user_query=query, + search_space_id=search_space_id, + top_k=top_k, + start_date=resolved_start_date, + end_date=resolved_end_date, + ) + all_documents.extend(chunks) + except Exception as e: print(f"Error searching connector {connector}: {e}") continue @@ -596,6 +607,7 @@ def create_search_knowledge_base_tool( - WEBCRAWLER_CONNECTOR: "Webpages indexed by SurfSense" (personally selected websites) - BOOKSTACK_CONNECTOR: "BookStack pages" (personal documentation) - CIRCLEBACK: "Circleback meeting notes, transcripts, and action items" (personal meeting records) + - OBSIDIAN_CONNECTOR: "Obsidian vault notes and markdown files" (personal notes and knowledge management) NOTE: `WEBCRAWLER_CONNECTOR` is mapped internally to the canonical document type `CRAWLED_URL`. diff --git a/surfsense_backend/app/config/__init__.py b/surfsense_backend/app/config/__init__.py index 71ef9c965..b0c1a2c09 100644 --- a/surfsense_backend/app/config/__init__.py +++ b/surfsense_backend/app/config/__init__.py @@ -61,6 +61,21 @@ class Config: "FFmpeg is not installed on the system. Please install it to use the Surfsense Podcaster." ) + # Deployment Mode (self-hosted or cloud) + # self-hosted: Full access to local file system connectors (Obsidian, etc.) + # cloud: Only cloud-based connectors available + DEPLOYMENT_MODE = os.getenv("SURFSENSE_DEPLOYMENT_MODE", "self-hosted") + + @classmethod + def is_self_hosted(cls) -> bool: + """Check if running in self-hosted mode.""" + return cls.DEPLOYMENT_MODE == "self-hosted" + + @classmethod + def is_cloud(cls) -> bool: + """Check if running in cloud mode.""" + return cls.DEPLOYMENT_MODE == "cloud" + # Database DATABASE_URL = os.getenv("DATABASE_URL") diff --git a/surfsense_backend/app/db.py b/surfsense_backend/app/db.py index 705e89ea7..8b6f3c718 100644 --- a/surfsense_backend/app/db.py +++ b/surfsense_backend/app/db.py @@ -53,6 +53,7 @@ class DocumentType(str, Enum): ELASTICSEARCH_CONNECTOR = "ELASTICSEARCH_CONNECTOR" BOOKSTACK_CONNECTOR = "BOOKSTACK_CONNECTOR" CIRCLEBACK = "CIRCLEBACK" + OBSIDIAN_CONNECTOR = "OBSIDIAN_CONNECTOR" NOTE = "NOTE" COMPOSIO_GOOGLE_DRIVE_CONNECTOR = "COMPOSIO_GOOGLE_DRIVE_CONNECTOR" COMPOSIO_GMAIL_CONNECTOR = "COMPOSIO_GMAIL_CONNECTOR" @@ -83,6 +84,9 @@ class SearchSourceConnectorType(str, Enum): WEBCRAWLER_CONNECTOR = "WEBCRAWLER_CONNECTOR" BOOKSTACK_CONNECTOR = "BOOKSTACK_CONNECTOR" CIRCLEBACK_CONNECTOR = "CIRCLEBACK_CONNECTOR" + OBSIDIAN_CONNECTOR = ( + "OBSIDIAN_CONNECTOR" # Self-hosted only - Local Obsidian vault indexing + ) MCP_CONNECTOR = "MCP_CONNECTOR" # Model Context Protocol - User-defined API tools COMPOSIO_GOOGLE_DRIVE_CONNECTOR = "COMPOSIO_GOOGLE_DRIVE_CONNECTOR" COMPOSIO_GMAIL_CONNECTOR = "COMPOSIO_GMAIL_CONNECTOR" @@ -419,6 +423,13 @@ class ChatComment(BaseModel, TimestampMixin): nullable=False, index=True, ) + # Denormalized thread_id for efficient Electric SQL subscriptions (one per thread) + thread_id = Column( + Integer, + ForeignKey("new_chat_threads.id", ondelete="CASCADE"), + nullable=False, + index=True, + ) parent_id = Column( Integer, ForeignKey("chat_comments.id", ondelete="CASCADE"), @@ -442,6 +453,7 @@ class ChatComment(BaseModel, TimestampMixin): # Relationships message = relationship("NewChatMessage", back_populates="comments") + thread = relationship("NewChatThread") author = relationship("User") parent = relationship( "ChatComment", remote_side="ChatComment.id", backref="replies" @@ -478,6 +490,38 @@ class ChatCommentMention(BaseModel, TimestampMixin): mentioned_user = relationship("User") +class ChatSessionState(BaseModel): + """ + Tracks real-time session state for shared chat collaboration. + One record per thread, synced via Electric SQL. + """ + + __tablename__ = "chat_session_state" + + thread_id = Column( + Integer, + ForeignKey("new_chat_threads.id", ondelete="CASCADE"), + nullable=False, + unique=True, + index=True, + ) + ai_responding_to_user_id = Column( + UUID(as_uuid=True), + ForeignKey("user.id", ondelete="SET NULL"), + nullable=True, + index=True, + ) + updated_at = Column( + TIMESTAMP(timezone=True), + nullable=False, + default=lambda: datetime.now(UTC), + onupdate=lambda: datetime.now(UTC), + ) + + thread = relationship("NewChatThread") + ai_responding_to_user = relationship("User") + + class MemoryCategory(str, Enum): """Categories for user memories.""" diff --git a/surfsense_backend/app/routes/new_chat_routes.py b/surfsense_backend/app/routes/new_chat_routes.py index 4b8600fab..25c53b69e 100644 --- a/surfsense_backend/app/routes/new_chat_routes.py +++ b/surfsense_backend/app/routes/new_chat_routes.py @@ -990,7 +990,7 @@ async def handle_new_chat( search_space_id=request.search_space_id, chat_id=request.chat_id, session=session, - user_id=str(user.id), # Pass user ID for memory tools + user_id=str(user.id), # Pass user ID for memory tools and session state llm_config_id=llm_config_id, attachments=request.attachments, mentioned_document_ids=request.mentioned_document_ids, diff --git a/surfsense_backend/app/routes/search_source_connectors_routes.py b/surfsense_backend/app/routes/search_source_connectors_routes.py index 433acac1c..db1b884e0 100644 --- a/surfsense_backend/app/routes/search_source_connectors_routes.py +++ b/surfsense_backend/app/routes/search_source_connectors_routes.py @@ -901,6 +901,25 @@ async def index_connector_content( ) response_message = "Web page indexing started in the background." + elif connector.connector_type == SearchSourceConnectorType.OBSIDIAN_CONNECTOR: + from app.config import config as app_config + from app.tasks.celery_tasks.connector_tasks import index_obsidian_vault_task + + # Obsidian connector only available in self-hosted mode + if not app_config.is_self_hosted(): + raise HTTPException( + status_code=400, + detail="Obsidian connector is only available in self-hosted mode", + ) + + logger.info( + f"Triggering Obsidian vault indexing for connector {connector_id} into search space {search_space_id} from {indexing_from} to {indexing_to}" + ) + index_obsidian_vault_task.delay( + connector_id, search_space_id, str(user.id), indexing_from, indexing_to + ) + response_message = "Obsidian vault indexing started in the background." + elif ( connector.connector_type == SearchSourceConnectorType.COMPOSIO_GOOGLE_DRIVE_CONNECTOR @@ -2195,6 +2214,58 @@ async def run_bookstack_indexing( ) +# Add new helper functions for Obsidian indexing +async def run_obsidian_indexing_with_new_session( + connector_id: int, + search_space_id: int, + user_id: str, + start_date: str, + end_date: str, +): + """Wrapper to run Obsidian indexing with its own database session.""" + logger.info( + f"Background task started: Indexing Obsidian connector {connector_id} into space {search_space_id} from {start_date} to {end_date}" + ) + async with async_session_maker() as session: + await run_obsidian_indexing( + session, connector_id, search_space_id, user_id, start_date, end_date + ) + logger.info(f"Background task finished: Indexing Obsidian connector {connector_id}") + + +async def run_obsidian_indexing( + session: AsyncSession, + connector_id: int, + search_space_id: int, + user_id: str, + start_date: str, + end_date: str, +): + """ + Background task to run Obsidian vault indexing. + + Args: + session: Database session + connector_id: ID of the Obsidian connector + search_space_id: ID of the search space + user_id: ID of the user + start_date: Start date for indexing + end_date: End date for indexing + """ + from app.tasks.connector_indexers import index_obsidian_vault + + await _run_indexing_with_notifications( + session=session, + connector_id=connector_id, + search_space_id=search_space_id, + user_id=user_id, + start_date=start_date, + end_date=end_date, + indexing_function=index_obsidian_vault, + update_timestamp_func=_update_connector_timestamp_by_id, + ) + + async def run_composio_indexing_with_new_session( connector_id: int, search_space_id: int, diff --git a/surfsense_backend/app/schemas/chat_session_state.py b/surfsense_backend/app/schemas/chat_session_state.py new file mode 100644 index 000000000..6eca0e26f --- /dev/null +++ b/surfsense_backend/app/schemas/chat_session_state.py @@ -0,0 +1,29 @@ +""" +Pydantic schemas for chat session state (live collaboration). +""" + +from datetime import datetime +from uuid import UUID + +from pydantic import BaseModel, ConfigDict + + +class RespondingUser(BaseModel): + """The user that the AI is currently responding to.""" + + id: UUID + display_name: str | None = None + email: str + + model_config = ConfigDict(from_attributes=True) + + +class ChatSessionStateResponse(BaseModel): + """Current session state for a chat thread.""" + + id: int + thread_id: int + responding_to: RespondingUser | None = None + updated_at: datetime + + model_config = ConfigDict(from_attributes=True) diff --git a/surfsense_backend/app/schemas/obsidian_auth_credentials.py b/surfsense_backend/app/schemas/obsidian_auth_credentials.py new file mode 100644 index 000000000..ab178eac8 --- /dev/null +++ b/surfsense_backend/app/schemas/obsidian_auth_credentials.py @@ -0,0 +1,59 @@ +""" +Obsidian Connector Credentials Schema. + +Obsidian is a local-first note-taking app that stores notes as markdown files. +This connector supports indexing from local file system (self-hosted only). +""" + +from pydantic import BaseModel, field_validator + + +class ObsidianAuthCredentialsBase(BaseModel): + """ + Credentials/configuration for the Obsidian connector. + + Since Obsidian vaults are local directories, this schema primarily + holds the vault path and configuration options rather than API tokens. + """ + + vault_path: str + vault_name: str | None = None + exclude_folders: list[str] | None = None + include_attachments: bool = False + + @field_validator("vault_path") + @classmethod + def validate_vault_path(cls, v: str) -> str: + """Ensure vault path is provided and stripped of whitespace.""" + if not v or not v.strip(): + raise ValueError("Vault path is required") + return v.strip() + + @field_validator("exclude_folders", mode="before") + @classmethod + def parse_exclude_folders(cls, v): + """Parse exclude_folders from string if needed.""" + if v is None: + return [".trash", ".obsidian", "templates"] + if isinstance(v, str): + return [f.strip() for f in v.split(",") if f.strip()] + return v + + def to_dict(self) -> dict: + """Convert credentials to dictionary for storage.""" + return { + "vault_path": self.vault_path, + "vault_name": self.vault_name, + "exclude_folders": self.exclude_folders, + "include_attachments": self.include_attachments, + } + + @classmethod + def from_dict(cls, data: dict) -> "ObsidianAuthCredentialsBase": + """Create credentials from dictionary.""" + return cls( + vault_path=data.get("vault_path", ""), + vault_name=data.get("vault_name"), + exclude_folders=data.get("exclude_folders"), + include_attachments=data.get("include_attachments", False), + ) diff --git a/surfsense_backend/app/services/chat_comments_service.py b/surfsense_backend/app/services/chat_comments_service.py index 6f81c0158..dc3b51238 100644 --- a/surfsense_backend/app/services/chat_comments_service.py +++ b/surfsense_backend/app/services/chat_comments_service.py @@ -281,8 +281,10 @@ async def create_comment( detail="You don't have permission to create comments in this search space", ) + thread = message.thread comment = ChatComment( message_id=message_id, + thread_id=thread.id, # Denormalized for efficient Electric subscriptions author_id=user.id, content=content, ) @@ -299,7 +301,6 @@ async def create_comment( user_names = await get_user_names_for_mentions(session, set(mentions_map.keys())) # Create notifications for mentioned users (excluding author) - thread = message.thread author_name = user.display_name or user.email content_preview = render_mentions(content, user_names) for mentioned_user_id, mention_id in mentions_map.items(): @@ -393,8 +394,10 @@ async def create_reply( detail="You don't have permission to create comments in this search space", ) + thread = parent_comment.message.thread reply = ChatComment( message_id=parent_comment.message_id, + thread_id=thread.id, # Denormalized for efficient Electric subscriptions parent_id=comment_id, author_id=user.id, content=content, @@ -412,7 +415,6 @@ async def create_reply( user_names = await get_user_names_for_mentions(session, set(mentions_map.keys())) # Create notifications for mentioned users (excluding author) - thread = parent_comment.message.thread author_name = user.display_name or user.email content_preview = render_mentions(content, user_names) for mentioned_user_id, mention_id in mentions_map.items(): diff --git a/surfsense_backend/app/services/chat_session_state_service.py b/surfsense_backend/app/services/chat_session_state_service.py new file mode 100644 index 000000000..d82fff3a7 --- /dev/null +++ b/surfsense_backend/app/services/chat_session_state_service.py @@ -0,0 +1,65 @@ +""" +Service layer for chat session state (live collaboration). +""" + +from datetime import UTC, datetime +from uuid import UUID + +from sqlalchemy import select +from sqlalchemy.dialects.postgresql import insert +from sqlalchemy.ext.asyncio import AsyncSession +from sqlalchemy.orm import selectinload + +from app.db import ChatSessionState + + +async def get_session_state( + session: AsyncSession, + thread_id: int, +) -> ChatSessionState | None: + """Get the current session state for a thread.""" + result = await session.execute( + select(ChatSessionState) + .options(selectinload(ChatSessionState.ai_responding_to_user)) + .filter(ChatSessionState.thread_id == thread_id) + ) + return result.scalar_one_or_none() + + +async def set_ai_responding( + session: AsyncSession, + thread_id: int, + user_id: UUID, +) -> ChatSessionState: + """Mark AI as responding to a specific user. Uses upsert for atomicity.""" + now = datetime.now(UTC) + upsert_query = insert(ChatSessionState).values( + thread_id=thread_id, + ai_responding_to_user_id=user_id, + updated_at=now, + ) + upsert_query = upsert_query.on_conflict_do_update( + index_elements=["thread_id"], + set_={ + "ai_responding_to_user_id": user_id, + "updated_at": now, + }, + ) + await session.execute(upsert_query) + await session.commit() + + return await get_session_state(session, thread_id) + + +async def clear_ai_responding( + session: AsyncSession, + thread_id: int, +) -> ChatSessionState | None: + """Clear AI responding state when response is complete.""" + state = await get_session_state(session, thread_id) + if state: + state.ai_responding_to_user_id = None + state.updated_at = datetime.now(UTC) + await session.commit() + await session.refresh(state) + return state diff --git a/surfsense_backend/app/services/connector_service.py b/surfsense_backend/app/services/connector_service.py index 832aee4cc..dc43697e7 100644 --- a/surfsense_backend/app/services/connector_service.py +++ b/surfsense_backend/app/services/connector_service.py @@ -2780,3 +2780,94 @@ class ConnectorService: } return result_object, circleback_docs + + async def search_obsidian( + self, + user_query: str, + search_space_id: int, + top_k: int = 20, + start_date: datetime | None = None, + end_date: datetime | None = None, + ) -> tuple: + """ + Search for Obsidian vault notes and return both the source information and langchain documents. + + Uses combined chunk-level and document-level hybrid search with RRF fusion. + + Args: + user_query: The user's query + search_space_id: The search space ID to search in + top_k: Maximum number of results to return + start_date: Optional start date for filtering documents by updated_at + end_date: Optional end date for filtering documents by updated_at + + Returns: + tuple: (sources_info, langchain_documents) + """ + obsidian_docs = await self._combined_rrf_search( + query_text=user_query, + search_space_id=search_space_id, + document_type="OBSIDIAN_CONNECTOR", + top_k=top_k, + start_date=start_date, + end_date=end_date, + ) + + # Early return if no results + if not obsidian_docs: + return { + "id": 53, + "name": "Obsidian Vault", + "type": "OBSIDIAN_CONNECTOR", + "sources": [], + }, [] + + def _title_fn(doc_info: dict[str, Any], metadata: dict[str, Any]) -> str: + return doc_info.get("title", "Untitled Note") + + def _url_fn(doc_info: dict[str, Any], metadata: dict[str, Any]) -> str: + # Obsidian URL format: obsidian://vault_name/path + return doc_info.get("url", "") + + def _description_fn( + chunk: dict[str, Any], _doc_info: dict[str, Any], metadata: dict[str, Any] + ) -> str: + description = self._chunk_preview(chunk.get("content", ""), limit=200) + info_parts = [] + vault_name = metadata.get("vault_name") + tags = metadata.get("tags", []) + if vault_name: + info_parts.append(f"Vault: {vault_name}") + if tags and isinstance(tags, list) and len(tags) > 0: + info_parts.append(f"Tags: {', '.join(tags[:3])}") + if info_parts: + description = (description + " | " + " | ".join(info_parts)).strip(" |") + return description + + def _extra_fields_fn( + _chunk: dict[str, Any], _doc_info: dict[str, Any], metadata: dict[str, Any] + ) -> dict[str, Any]: + return { + "vault_name": metadata.get("vault_name", ""), + "file_path": metadata.get("file_path", ""), + "tags": metadata.get("tags", []), + "outgoing_links": metadata.get("outgoing_links", []), + } + + sources_list = self._build_chunk_sources_from_documents( + obsidian_docs, + title_fn=_title_fn, + url_fn=_url_fn, + description_fn=_description_fn, + extra_fields_fn=_extra_fields_fn, + ) + + # Create result object + result_object = { + "id": 53, + "name": "Obsidian Vault", + "type": "OBSIDIAN_CONNECTOR", + "sources": sources_list, + } + + return result_object, obsidian_docs diff --git a/surfsense_backend/app/services/notification_service.py b/surfsense_backend/app/services/notification_service.py index 5f7f568f6..836daeb9e 100644 --- a/surfsense_backend/app/services/notification_service.py +++ b/surfsense_backend/app/services/notification_service.py @@ -623,6 +623,28 @@ class MentionNotificationHandler(BaseNotificationHandler): def __init__(self): super().__init__("new_mention") + async def find_notification_by_mention( + self, + session: AsyncSession, + mention_id: int, + ) -> Notification | None: + """ + Find an existing notification by mention ID. + + Args: + session: Database session + mention_id: The mention ID to search for + + Returns: + Notification if found, None otherwise + """ + query = select(Notification).where( + Notification.type == self.notification_type, + Notification.notification_metadata["mention_id"].astext == str(mention_id), + ) + result = await session.execute(query) + return result.scalar_one_or_none() + async def notify_new_mention( self, session: AsyncSession, @@ -641,11 +663,12 @@ class MentionNotificationHandler(BaseNotificationHandler): ) -> Notification: """ Create notification when a user is @mentioned in a comment. + Uses mention_id for idempotency to prevent duplicate notifications. Args: session: Database session mentioned_user_id: User who was mentioned - mention_id: ID of the mention record + mention_id: ID of the mention record (used for idempotency) comment_id: ID of the comment containing the mention message_id: ID of the message being commented on thread_id: ID of the chat thread @@ -658,8 +681,16 @@ class MentionNotificationHandler(BaseNotificationHandler): search_space_id: Search space ID Returns: - Notification: The created notification + Notification: The created or existing notification """ + # Check if notification already exists for this mention (idempotency) + existing = await self.find_notification_by_mention(session, mention_id) + if existing: + logger.info( + f"Notification already exists for mention {mention_id}, returning existing" + ) + return existing + title = f"{author_name} mentioned you" message = content_preview[:100] + ("..." if len(content_preview) > 100 else "") @@ -676,21 +707,37 @@ class MentionNotificationHandler(BaseNotificationHandler): "content_preview": content_preview[:200], } - notification = Notification( - user_id=mentioned_user_id, - search_space_id=search_space_id, - type=self.notification_type, - title=title, - message=message, - notification_metadata=metadata, - ) - session.add(notification) - await session.commit() - await session.refresh(notification) - logger.info( - f"Created new_mention notification {notification.id} for user {mentioned_user_id}" - ) - return notification + try: + notification = Notification( + user_id=mentioned_user_id, + search_space_id=search_space_id, + type=self.notification_type, + title=title, + message=message, + notification_metadata=metadata, + ) + session.add(notification) + await session.commit() + await session.refresh(notification) + logger.info( + f"Created new_mention notification {notification.id} for user {mentioned_user_id}" + ) + return notification + except Exception as e: + # Handle race condition - if duplicate key error, try to fetch existing + await session.rollback() + if ( + "duplicate key" in str(e).lower() + or "unique constraint" in str(e).lower() + ): + logger.warning( + f"Duplicate notification detected for mention {mention_id}, fetching existing" + ) + existing = await self.find_notification_by_mention(session, mention_id) + if existing: + return existing + # Re-raise if not a duplicate key error or couldn't find existing + raise class NotificationService: diff --git a/surfsense_backend/app/tasks/celery_tasks/connector_tasks.py b/surfsense_backend/app/tasks/celery_tasks/connector_tasks.py index 307b5a551..81cafaa2c 100644 --- a/surfsense_backend/app/tasks/celery_tasks/connector_tasks.py +++ b/surfsense_backend/app/tasks/celery_tasks/connector_tasks.py @@ -761,6 +761,49 @@ async def _index_bookstack_pages( ) +@celery_app.task(name="index_obsidian_vault", bind=True) +def index_obsidian_vault_task( + self, + connector_id: int, + search_space_id: int, + user_id: str, + start_date: str, + end_date: str, +): + """Celery task to index Obsidian vault notes.""" + import asyncio + + loop = asyncio.new_event_loop() + asyncio.set_event_loop(loop) + + try: + loop.run_until_complete( + _index_obsidian_vault( + connector_id, search_space_id, user_id, start_date, end_date + ) + ) + finally: + loop.close() + + +async def _index_obsidian_vault( + connector_id: int, + search_space_id: int, + user_id: str, + start_date: str, + end_date: str, +): + """Index Obsidian vault with new session.""" + from app.routes.search_source_connectors_routes import ( + run_obsidian_indexing, + ) + + async with get_celery_session_maker()() as session: + await run_obsidian_indexing( + session, connector_id, search_space_id, user_id, start_date, end_date + ) + + @celery_app.task(name="index_composio_connector", bind=True) def index_composio_connector_task( self, diff --git a/surfsense_backend/app/tasks/chat/stream_new_chat.py b/surfsense_backend/app/tasks/chat/stream_new_chat.py index 7d2cf4172..31229a59b 100644 --- a/surfsense_backend/app/tasks/chat/stream_new_chat.py +++ b/surfsense_backend/app/tasks/chat/stream_new_chat.py @@ -11,6 +11,7 @@ Supports loading LLM configurations from: import json from collections.abc import AsyncGenerator +from uuid import UUID from langchain_core.messages import HumanMessage from sqlalchemy.ext.asyncio import AsyncSession @@ -27,6 +28,10 @@ from app.agents.new_chat.llm_config import ( ) from app.db import Document, SurfsenseDocsDocument from app.schemas.new_chat import ChatAttachment +from app.services.chat_session_state_service import ( + clear_ai_responding, + set_ai_responding, +) from app.services.connector_service import ConnectorService from app.services.new_streaming_service import VercelStreamingService @@ -167,9 +172,8 @@ async def stream_new_chat( search_space_id: The search space ID chat_id: The chat ID (used as LangGraph thread_id for memory) session: The database session - user_id: The current user's UUID string (for memory tools) + user_id: The current user's UUID string (for memory tools and session state) llm_config_id: The LLM configuration ID (default: -1 for first global config) - messages: Optional chat history from frontend (list of ChatMessage) attachments: Optional attachments with extracted content mentioned_document_ids: Optional list of document IDs mentioned with @ in the chat mentioned_surfsense_doc_ids: Optional list of SurfSense doc IDs mentioned with @ in the chat @@ -183,6 +187,9 @@ async def stream_new_chat( current_text_id: str | None = None try: + # Mark AI as responding to this user for live collaboration + if user_id: + await set_ai_responding(session, chat_id, UUID(user_id)) # Load LLM config - supports both YAML (negative IDs) and database (positive IDs) agent_config: AgentConfig | None = None @@ -1147,3 +1154,7 @@ async def stream_new_chat( yield streaming_service.format_finish_step() yield streaming_service.format_finish() yield streaming_service.format_done() + + finally: + # Clear AI responding state for live collaboration + await clear_ai_responding(session, chat_id) diff --git a/surfsense_backend/app/tasks/connector_indexers/__init__.py b/surfsense_backend/app/tasks/connector_indexers/__init__.py index 8f25e6fdd..9a1d17fd5 100644 --- a/surfsense_backend/app/tasks/connector_indexers/__init__.py +++ b/surfsense_backend/app/tasks/connector_indexers/__init__.py @@ -46,6 +46,7 @@ from .luma_indexer import index_luma_events # Documentation and knowledge management from .notion_indexer import index_notion_pages +from .obsidian_indexer import index_obsidian_vault from .slack_indexer import index_slack_messages from .webcrawler_indexer import index_crawled_urls @@ -68,6 +69,7 @@ __all__ = [ # noqa: RUF022 "index_linear_issues", # Documentation and knowledge management "index_notion_pages", + "index_obsidian_vault", "index_crawled_urls", # Communication platforms "index_slack_messages", diff --git a/surfsense_backend/app/tasks/connector_indexers/obsidian_indexer.py b/surfsense_backend/app/tasks/connector_indexers/obsidian_indexer.py new file mode 100644 index 000000000..4c4dab4c2 --- /dev/null +++ b/surfsense_backend/app/tasks/connector_indexers/obsidian_indexer.py @@ -0,0 +1,516 @@ +""" +Obsidian connector indexer. + +Indexes markdown notes from a local Obsidian vault. +This connector is only available in self-hosted mode. +""" + +import os +import re +from datetime import UTC, datetime +from pathlib import Path + +import yaml +from sqlalchemy.exc import SQLAlchemyError +from sqlalchemy.ext.asyncio import AsyncSession + +from app.config import config +from app.db import Document, DocumentType, SearchSourceConnectorType +from app.services.llm_service import get_user_long_context_llm +from app.services.task_logging_service import TaskLoggingService +from app.utils.document_converters import ( + create_document_chunks, + generate_content_hash, + generate_document_summary, + generate_unique_identifier_hash, +) + +from .base import ( + build_document_metadata_string, + check_document_by_unique_identifier, + get_connector_by_id, + get_current_timestamp, + logger, + update_connector_last_indexed, +) + + +def parse_frontmatter(content: str) -> tuple[dict | None, str]: + """ + Parse YAML frontmatter from markdown content. + + Args: + content: The full markdown content + + Returns: + Tuple of (frontmatter dict or None, content without frontmatter) + """ + if not content.startswith("---"): + return None, content + + # Find the closing --- + end_match = re.search(r"\n---\n", content[3:]) + if not end_match: + return None, content + + frontmatter_str = content[3 : end_match.start() + 3] + remaining_content = content[end_match.end() + 3 :] + + try: + frontmatter = yaml.safe_load(frontmatter_str) + return frontmatter, remaining_content.strip() + except yaml.YAMLError: + return None, content + + +def extract_wiki_links(content: str) -> list[str]: + """ + Extract [[wiki-style links]] from content. + + Args: + content: Markdown content + + Returns: + List of linked note names + """ + # Match [[link]] or [[link|alias]] + pattern = r"\[\[([^\]|]+)(?:\|[^\]]+)?\]\]" + matches = re.findall(pattern, content) + return list(set(matches)) + + +def extract_tags(content: str) -> list[str]: + """ + Extract #tags from content (both inline and frontmatter). + + Args: + content: Markdown content + + Returns: + List of tags (without # prefix) + """ + # Match #tag but not ## headers + pattern = r"(? list[dict]: + """ + Scan an Obsidian vault for markdown files. + + Args: + vault_path: Path to the Obsidian vault + exclude_folders: List of folder names to exclude + + Returns: + List of file info dicts with path, name, modified time + """ + if exclude_folders is None: + exclude_folders = [".trash", ".obsidian", "templates"] + + vault = Path(vault_path) + if not vault.exists(): + raise ValueError(f"Vault path does not exist: {vault_path}") + + files = [] + for md_file in vault.rglob("*.md"): + # Check if file is in an excluded folder + relative_path = md_file.relative_to(vault) + parts = relative_path.parts + + if any(excluded in parts for excluded in exclude_folders): + continue + + try: + stat = md_file.stat() + files.append( + { + "path": str(md_file), + "relative_path": str(relative_path), + "name": md_file.stem, + "modified_at": datetime.fromtimestamp(stat.st_mtime, tz=UTC), + "created_at": datetime.fromtimestamp(stat.st_ctime, tz=UTC), + "size": stat.st_size, + } + ) + except OSError as e: + logger.warning(f"Could not stat file {md_file}: {e}") + + return files + + +async def index_obsidian_vault( + session: AsyncSession, + connector_id: int, + search_space_id: int, + user_id: str, + start_date: str | None = None, + end_date: str | None = None, + update_last_indexed: bool = True, +) -> tuple[int, str | None]: + """ + Index notes from a local Obsidian vault. + + This indexer is only available in self-hosted mode as it requires + direct file system access to the user's Obsidian vault. + + Args: + session: Database session + connector_id: ID of the Obsidian connector + search_space_id: ID of the search space to store documents in + user_id: ID of the user + start_date: Start date for filtering (YYYY-MM-DD format) - optional + end_date: End date for filtering (YYYY-MM-DD format) - optional + update_last_indexed: Whether to update the last_indexed_at timestamp + + Returns: + Tuple containing (number of documents indexed, error message or None) + """ + task_logger = TaskLoggingService(session, search_space_id) + + # Check if self-hosted mode + if not config.is_self_hosted(): + return 0, "Obsidian connector is only available in self-hosted mode" + + # Log task start + log_entry = await task_logger.log_task_start( + task_name="obsidian_vault_indexing", + source="connector_indexing_task", + message=f"Starting Obsidian vault indexing for connector {connector_id}", + metadata={ + "connector_id": connector_id, + "user_id": str(user_id), + "start_date": start_date, + "end_date": end_date, + }, + ) + + try: + # Get the connector + await task_logger.log_task_progress( + log_entry, + f"Retrieving Obsidian connector {connector_id} from database", + {"stage": "connector_retrieval"}, + ) + + connector = await get_connector_by_id( + session, connector_id, SearchSourceConnectorType.OBSIDIAN_CONNECTOR + ) + + if not connector: + await task_logger.log_task_failure( + log_entry, + f"Connector with ID {connector_id} not found or is not an Obsidian connector", + "Connector not found", + {"error_type": "ConnectorNotFound"}, + ) + return ( + 0, + f"Connector with ID {connector_id} not found or is not an Obsidian connector", + ) + + # Get vault path from connector config + vault_path = connector.config.get("vault_path") + if not vault_path: + await task_logger.log_task_failure( + log_entry, + "Vault path not configured for this connector", + "Missing vault path", + {"error_type": "MissingVaultPath"}, + ) + return 0, "Vault path not configured for this connector" + + # Validate vault path exists + if not os.path.exists(vault_path): + await task_logger.log_task_failure( + log_entry, + f"Vault path does not exist: {vault_path}", + "Vault path not found", + {"error_type": "VaultNotFound", "vault_path": vault_path}, + ) + return 0, f"Vault path does not exist: {vault_path}" + + # Get configuration options + exclude_folders = connector.config.get( + "exclude_folders", [".trash", ".obsidian", "templates"] + ) + vault_name = connector.config.get("vault_name") or os.path.basename(vault_path) + + await task_logger.log_task_progress( + log_entry, + f"Scanning Obsidian vault: {vault_name}", + {"stage": "vault_scan", "vault_path": vault_path}, + ) + + # Scan vault for markdown files + try: + files = scan_vault(vault_path, exclude_folders) + except Exception as e: + await task_logger.log_task_failure( + log_entry, + f"Failed to scan vault: {e}", + "Vault scan error", + {"error_type": "VaultScanError"}, + ) + return 0, f"Failed to scan vault: {e}" + + logger.info(f"Found {len(files)} markdown files in vault") + + await task_logger.log_task_progress( + log_entry, + f"Found {len(files)} markdown files to process", + {"stage": "files_discovered", "file_count": len(files)}, + ) + + # Filter by date if provided (handle "undefined" string from frontend) + # Also handle inverted dates (start > end) by skipping filtering + start_dt = None + end_dt = None + + if start_date and start_date != "undefined": + start_dt = datetime.strptime(start_date, "%Y-%m-%d").replace(tzinfo=UTC) + + if end_date and end_date != "undefined": + # Make end_date inclusive (end of day) + end_dt = datetime.strptime(end_date, "%Y-%m-%d").replace(tzinfo=UTC) + end_dt = end_dt.replace(hour=23, minute=59, second=59) + + # Only apply date filtering if dates are valid and in correct order + if start_dt and end_dt and start_dt > end_dt: + logger.warning( + f"start_date ({start_date}) is after end_date ({end_date}), skipping date filter" + ) + else: + if start_dt: + files = [f for f in files if f["modified_at"] >= start_dt] + logger.info( + f"After start_date filter ({start_date}): {len(files)} files" + ) + if end_dt: + files = [f for f in files if f["modified_at"] <= end_dt] + logger.info(f"After end_date filter ({end_date}): {len(files)} files") + + logger.info(f"Processing {len(files)} files after date filtering") + + # Get LLM for summarization + long_context_llm = await get_user_long_context_llm( + session, user_id, search_space_id + ) + + indexed_count = 0 + skipped_count = 0 + + for file_info in files: + try: + file_path = file_info["path"] + relative_path = file_info["relative_path"] + + # Read file content + try: + with open(file_path, encoding="utf-8") as f: + content = f.read() + except UnicodeDecodeError: + logger.warning(f"Could not decode file {file_path}, skipping") + skipped_count += 1 + continue + + if not content.strip(): + logger.debug(f"Empty file {file_path}, skipping") + skipped_count += 1 + continue + + # Parse frontmatter and extract metadata + frontmatter, body_content = parse_frontmatter(content) + wiki_links = extract_wiki_links(content) + tags = extract_tags(content) + + # Get title from frontmatter or filename + title = file_info["name"] + if frontmatter: + title = frontmatter.get("title", title) + # Also extract tags from frontmatter + fm_tags = frontmatter.get("tags", []) + if isinstance(fm_tags, list): + tags = list({*tags, *fm_tags}) + elif isinstance(fm_tags, str): + tags = list({*tags, fm_tags}) + + # Generate unique identifier using vault name and relative path + unique_identifier = f"{vault_name}:{relative_path}" + unique_identifier_hash = generate_unique_identifier_hash( + DocumentType.OBSIDIAN_CONNECTOR, + unique_identifier, + search_space_id, + ) + + # Check for existing document + existing_document = await check_document_by_unique_identifier( + session, unique_identifier_hash + ) + + # Generate content hash + content_hash = generate_content_hash(content, search_space_id) + + # Build metadata + document_metadata = { + "vault_name": vault_name, + "file_path": relative_path, + "tags": tags, + "outgoing_links": wiki_links, + "frontmatter": frontmatter, + "modified_at": file_info["modified_at"].isoformat(), + "created_at": file_info["created_at"].isoformat(), + "word_count": len(body_content.split()), + } + + # Build document content with metadata + metadata_sections = [ + ( + "METADATA", + [ + f"Title: {title}", + f"Vault: {vault_name}", + f"Path: {relative_path}", + f"Tags: {', '.join(tags) if tags else 'None'}", + f"Links to: {', '.join(wiki_links) if wiki_links else 'None'}", + ], + ), + ("CONTENT", [body_content]), + ] + document_string = build_document_metadata_string(metadata_sections) + + if existing_document: + # Check if content has changed + if existing_document.content_hash == content_hash: + logger.debug(f"Note {title} unchanged, skipping") + skipped_count += 1 + continue + + # Update existing document + logger.info(f"Updating note: {title}") + + # Generate new summary if content changed + if long_context_llm: + new_summary, _ = await generate_document_summary( + document_string, + long_context_llm, + document_metadata, + ) + # Store summary in metadata + document_metadata["summary"] = new_summary + + # Add URL and connector_id to metadata + document_metadata["url"] = ( + f"obsidian://{vault_name}/{relative_path}" + ) + document_metadata["connector_id"] = connector_id + + existing_document.content = document_string + existing_document.content_hash = content_hash + existing_document.document_metadata = document_metadata + existing_document.updated_at = get_current_timestamp() + + # Update embedding + embedding = config.embedding_model_instance.embed(document_string) + existing_document.embedding = embedding + + # Update chunks - delete old and create new + existing_document.chunks.clear() + new_chunks = await create_document_chunks(document_string) + existing_document.chunks = new_chunks + + indexed_count += 1 + + else: + # Create new document + logger.info(f"Indexing new note: {title}") + + # Generate summary + summary_content = "" + if long_context_llm: + summary_content, _ = await generate_document_summary( + document_string, + long_context_llm, + document_metadata, + ) + + # Generate embedding + embedding = config.embedding_model_instance.embed(document_string) + + # Add URL and summary to metadata + document_metadata["url"] = ( + f"obsidian://{vault_name}/{relative_path}" + ) + document_metadata["summary"] = summary_content + document_metadata["connector_id"] = connector_id + + # Create chunks + chunks = await create_document_chunks(document_string) + + # Create document + new_document = Document( + search_space_id=search_space_id, + title=title, + document_type=DocumentType.OBSIDIAN_CONNECTOR, + content=document_string, + content_hash=content_hash, + unique_identifier_hash=unique_identifier_hash, + document_metadata=document_metadata, + embedding=embedding, + chunks=chunks, + updated_at=get_current_timestamp(), + ) + + session.add(new_document) + + indexed_count += 1 + + except Exception as e: + logger.exception( + f"Error processing file {file_info.get('path', 'unknown')}: {e}" + ) + skipped_count += 1 + continue + + # Update connector's last indexed timestamp + await update_connector_last_indexed(session, connector, update_last_indexed) + + # Commit all changes + await session.commit() + + await task_logger.log_task_success( + log_entry, + f"Successfully indexed {indexed_count} Obsidian notes (skipped {skipped_count})", + { + "indexed_count": indexed_count, + "skipped_count": skipped_count, + "total_files": len(files), + }, + ) + + return indexed_count, None + + except SQLAlchemyError as e: + logger.exception(f"Database error during Obsidian indexing: {e}") + await session.rollback() + await task_logger.log_task_failure( + log_entry, + f"Database error during Obsidian indexing: {e}", + "Database error", + {"error_type": "SQLAlchemyError"}, + ) + return 0, f"Database error: {e}" + + except Exception as e: + logger.exception(f"Error during Obsidian indexing: {e}") + await task_logger.log_task_failure( + log_entry, + f"Error during Obsidian indexing: {e}", + "Unexpected error", + {"error_type": type(e).__name__}, + ) + return 0, str(e) diff --git a/surfsense_backend/pyproject.toml b/surfsense_backend/pyproject.toml index f96619607..ffe9e5232 100644 --- a/surfsense_backend/pyproject.toml +++ b/surfsense_backend/pyproject.toml @@ -46,12 +46,9 @@ dependencies = [ "boto3>=1.35.0", "langchain-community>=0.3.31", "langchain-unstructured>=1.0.0", - "langchain>=1.2.0", "litellm>=1.80.10", "langchain-litellm>=0.3.5", - "langgraph>=1.0.5", "fake-useragent>=2.2.0", - "deepagents>=0.3.0", "trafilatura>=2.0.0", "fastapi-users[oauth,sqlalchemy]>=15.0.3", "chonkie[all]>=1.5.0", @@ -62,6 +59,9 @@ dependencies = [ "sse-starlette>=3.1.1,<3.1.2", "gitingest>=0.3.1", "composio>=0.10.9", + "deepagents>=0.3.8", + "langchain>=1.2.6", + "langgraph>=1.0.5", ] [dependency-groups] diff --git a/surfsense_backend/uv.lock b/surfsense_backend/uv.lock index 96282380b..18f04288e 100644 --- a/surfsense_backend/uv.lock +++ b/surfsense_backend/uv.lock @@ -195,7 +195,7 @@ wheels = [ [[package]] name = "anthropic" -version = "0.75.0" +version = "0.76.0" source = { registry = "https://pypi.org/simple" } dependencies = [ { name = "anyio" }, @@ -207,9 +207,9 @@ dependencies = [ { name = "sniffio" }, { name = "typing-extensions" }, ] -sdist = { url = "https://files.pythonhosted.org/packages/04/1f/08e95f4b7e2d35205ae5dcbb4ae97e7d477fc521c275c02609e2931ece2d/anthropic-0.75.0.tar.gz", hash = "sha256:e8607422f4ab616db2ea5baacc215dd5f028da99ce2f022e33c7c535b29f3dfb", size = 439565 } +sdist = { url = "https://files.pythonhosted.org/packages/6e/be/d11abafaa15d6304826438170f7574d750218f49a106c54424a40cef4494/anthropic-0.76.0.tar.gz", hash = "sha256:e0cae6a368986d5cf6df743dfbb1b9519e6a9eee9c6c942ad8121c0b34416ffe", size = 495483 } wheels = [ - { url = "https://files.pythonhosted.org/packages/60/1c/1cd02b7ae64302a6e06724bf80a96401d5313708651d277b1458504a1730/anthropic-0.75.0-py3-none-any.whl", hash = "sha256:ea8317271b6c15d80225a9f3c670152746e88805a7a61e14d4a374577164965b", size = 388164 }, + { url = "https://files.pythonhosted.org/packages/e5/70/7b0fd9c1a738f59d3babe2b4212031c34ab7d0fda4ffef15b58a55c5bcea/anthropic-0.76.0-py3-none-any.whl", hash = "sha256:81efa3113901192af2f0fe977d3ec73fdadb1e691586306c4256cd6d5ccc331c", size = 390309 }, ] [[package]] @@ -1231,17 +1231,18 @@ wheels = [ [[package]] name = "deepagents" -version = "0.3.0" +version = "0.3.8" source = { registry = "https://pypi.org/simple" } dependencies = [ { name = "langchain" }, { name = "langchain-anthropic" }, { name = "langchain-core" }, + { name = "langchain-google-genai" }, { name = "wcmatch" }, ] -sdist = { url = "https://files.pythonhosted.org/packages/cc/c7/d3c2840bd0e66b6cd5948aa69625e129328ad261308e18fcb9a9420709da/deepagents-0.3.0.tar.gz", hash = "sha256:3dd4d2ed53efb1ef78aeb1020a5696c0ec7e58e627b305a6665d33fe6fbdedff", size = 51387 } +sdist = { url = "https://files.pythonhosted.org/packages/47/69/d8dd80dd5c0c81393cc32623dd51e642c8607ab798276506b3b3e89b1f20/deepagents-0.3.8.tar.gz", hash = "sha256:4b8252f8deaad449ce39426cc2233a597ee079b9b690647a26d128c16d6c6eb8", size = 73956 } wheels = [ - { url = "https://files.pythonhosted.org/packages/22/e9/60bab7f37ff38bf982ea578e457ed1878ded613a3425462bcd07b00487e9/deepagents-0.3.0-py3-none-any.whl", hash = "sha256:9e23532d8d535dc2b0b4e0834453a1223a6a8f81b77947c0faf54537d05ce89a", size = 54065 }, + { url = "https://files.pythonhosted.org/packages/a6/6a/35968909bd3184eafee97326bcfd16c99bf9b0e03aadb3327eaf7229ea11/deepagents-0.3.8-py3-none-any.whl", hash = "sha256:7c76205dc014173d795402045b51505517954c7c4a508175f8e4a529f51928cc", size = 79161 }, ] [[package]] @@ -1992,9 +1993,9 @@ dependencies = [ { name = "starlette" }, { name = "tiktoken" }, ] -sdist = { url = "https://files.pythonhosted.org/packages/d6/fe/a915f0c32a3d7920206a677f73c185b3eadf4ec151fb05aedd52e64713f7/gitingest-0.3.1.tar.gz", hash = "sha256:4587cab873d4e08bdb16d612bb153c23e0ce59771a1d57a438239c5e39f05ebf", size = 70681, upload-time = "2025-07-31T13:56:19.845Z" } +sdist = { url = "https://files.pythonhosted.org/packages/d6/fe/a915f0c32a3d7920206a677f73c185b3eadf4ec151fb05aedd52e64713f7/gitingest-0.3.1.tar.gz", hash = "sha256:4587cab873d4e08bdb16d612bb153c23e0ce59771a1d57a438239c5e39f05ebf", size = 70681 } wheels = [ - { url = "https://files.pythonhosted.org/packages/00/15/f200ab2e73287e67d1dce6fbacf421552ae9fbafdc5f0cc8dd0d2fe4fc47/gitingest-0.3.1-py3-none-any.whl", hash = "sha256:8143a5e6a7140ede9f680e13d3931ac07c82ac9bd8bab9ad1fba017c8c1e8666", size = 68343, upload-time = "2025-07-31T13:56:17.729Z" }, + { url = "https://files.pythonhosted.org/packages/00/15/f200ab2e73287e67d1dce6fbacf421552ae9fbafdc5f0cc8dd0d2fe4fc47/gitingest-0.3.1-py3-none-any.whl", hash = "sha256:8143a5e6a7140ede9f680e13d3931ac07c82ac9bd8bab9ad1fba017c8c1e8666", size = 68343 }, ] [[package]] @@ -2916,30 +2917,30 @@ wheels = [ [[package]] name = "langchain" -version = "1.2.0" +version = "1.2.6" source = { registry = "https://pypi.org/simple" } dependencies = [ { name = "langchain-core" }, { name = "langgraph" }, { name = "pydantic" }, ] -sdist = { url = "https://files.pythonhosted.org/packages/b1/12/3a74c22abdfddd877dfc2ee666d516f9132877fcd25eb4dd694835c59c79/langchain-1.2.0.tar.gz", hash = "sha256:a087d1e2b2969819e29a91a6d5f98302aafe31bd49ba377ecee3bf5a5dcfe14a", size = 536126 } +sdist = { url = "https://files.pythonhosted.org/packages/f5/bc/d8f506a525baadee99a65c6cc28c1c35c9eaf1cb2009f048e9861d81a600/langchain-1.2.6.tar.gz", hash = "sha256:7d46cbf719d860a16f6fc182d5d3de17453dda187f3d43e9c40ac352a5094fdd", size = 553127 } wheels = [ - { url = "https://files.pythonhosted.org/packages/23/00/4e3fa0d90f5a5c376ccb8ca983d0f0f7287783dfac48702e18f01d24673b/langchain-1.2.0-py3-none-any.whl", hash = "sha256:82f0d17aa4fbb11560b30e1e7d4aeb75e3ad71ce09b85c90ab208b181a24ffac", size = 102828 }, + { url = "https://files.pythonhosted.org/packages/3f/28/d5dc4cb06ccb29d62a590d446072964766555e85863f5044c6e644c07d0d/langchain-1.2.6-py3-none-any.whl", hash = "sha256:a9a6c39f03c09b6eb0f1b47e267ad2a2fd04e124dfaa9753bd6c11d2fe7d944e", size = 108458 }, ] [[package]] name = "langchain-anthropic" -version = "1.3.0" +version = "1.3.1" source = { registry = "https://pypi.org/simple" } dependencies = [ { name = "anthropic" }, { name = "langchain-core" }, { name = "pydantic" }, ] -sdist = { url = "https://files.pythonhosted.org/packages/de/50/cc3b3e0410d86de457d7a100dde763fc1c33c4ce884e883659aa4cf95538/langchain_anthropic-1.3.0.tar.gz", hash = "sha256:497a937ee0310c588196bff37f39f02d43d87bff3a12d16278bdbc3bd0e9a80b", size = 707207 } +sdist = { url = "https://files.pythonhosted.org/packages/0d/b6/ac5ee84e15bf79844c9c791f99a614c7ec7e1a63c2947e55977be01a81b4/langchain_anthropic-1.3.1.tar.gz", hash = "sha256:4f3d7a4a7729ab1aeaf62d32c87d4d227c1b5421668ca9e3734562b383470b07", size = 708940 } wheels = [ - { url = "https://files.pythonhosted.org/packages/29/ca/0725bc347a9c226da9d76f85bf7d03115caec7dbc87876af68579c4ab24e/langchain_anthropic-1.3.0-py3-none-any.whl", hash = "sha256:3823560e1df15d6082636baa04f87cb59052ba70aada0eba381c4679b1ce0eba", size = 45724 }, + { url = "https://files.pythonhosted.org/packages/9a/4f/7a5b32764addf4b757545b89899b9d76688176f19e4ee89868e3b8bbfd0f/langchain_anthropic-1.3.1-py3-none-any.whl", hash = "sha256:1fc28cf8037c30597ee6172fc2ff9e345efe8149a8c2a39897b1eebba2948322", size = 46328 }, ] [[package]] @@ -2967,7 +2968,7 @@ wheels = [ [[package]] name = "langchain-core" -version = "1.2.1" +version = "1.2.7" source = { registry = "https://pypi.org/simple" } dependencies = [ { name = "jsonpatch" }, @@ -2979,9 +2980,24 @@ dependencies = [ { name = "typing-extensions" }, { name = "uuid-utils" }, ] -sdist = { url = "https://files.pythonhosted.org/packages/f5/a0/2177f4ef4dfbea8edeba377b7b4889d177b8356ce186640e4651b240fd4d/langchain_core-1.2.1.tar.gz", hash = "sha256:131e6ad105b47ec2adc4d4d973f569276688f48cd890ba44603d48e76d9993ce", size = 802986 } +sdist = { url = "https://files.pythonhosted.org/packages/a2/0e/664d8d81b3493e09cbab72448d2f9d693d1fa5aa2bcc488602203a9b6da0/langchain_core-1.2.7.tar.gz", hash = "sha256:e1460639f96c352b4a41c375f25aeb8d16ffc1769499fb1c20503aad59305ced", size = 837039 } wheels = [ - { url = "https://files.pythonhosted.org/packages/cc/95/98c47dbb4b6098934ff70e0f52efef3a85505dbcccc9eb63587e21fde4c9/langchain_core-1.2.1-py3-none-any.whl", hash = "sha256:2f63859f85dc3d95f768e35fed605702e3ff5aa3e92c7b253103119613e79768", size = 475972 }, + { url = "https://files.pythonhosted.org/packages/6e/6f/34a9fba14d191a67f7e2ee3dbce3e9b86d2fa7310e2c7f2c713583481bd2/langchain_core-1.2.7-py3-none-any.whl", hash = "sha256:452f4fef7a3d883357b22600788d37e3d8854ef29da345b7ac7099f33c31828b", size = 490232 }, +] + +[[package]] +name = "langchain-google-genai" +version = "4.2.0" +source = { registry = "https://pypi.org/simple" } +dependencies = [ + { name = "filetype" }, + { name = "google-genai" }, + { name = "langchain-core" }, + { name = "pydantic" }, +] +sdist = { url = "https://files.pythonhosted.org/packages/d8/0b/eae2305e207574dc633983a8a82a745e0ede1bce1f3a9daff24d2341fadc/langchain_google_genai-4.2.0.tar.gz", hash = "sha256:9a8d9bfc35354983ed29079cefff53c3e7c9c2a44b6ba75cc8f13a0cf8b55c33", size = 277361 } +wheels = [ + { url = "https://files.pythonhosted.org/packages/22/51/39942c0083139652494bb354dddf0ed397703a4882302f7b48aeca531c96/langchain_google_genai-4.2.0-py3-none-any.whl", hash = "sha256:856041aaafceff65a4ef0d5acf5731f2db95229ff041132af011aec51e8279d9", size = 66452 }, ] [[package]] @@ -4516,9 +4532,9 @@ wheels = [ name = "pathspec" version = "1.0.3" source = { registry = "https://pypi.org/simple" } -sdist = { url = "https://files.pythonhosted.org/packages/4c/b2/bb8e495d5262bfec41ab5cb18f522f1012933347fb5d9e62452d446baca2/pathspec-1.0.3.tar.gz", hash = "sha256:bac5cf97ae2c2876e2d25ebb15078eb04d76e4b98921ee31c6f85ade8b59444d", size = 130841, upload-time = "2026-01-09T15:46:46.009Z" } +sdist = { url = "https://files.pythonhosted.org/packages/4c/b2/bb8e495d5262bfec41ab5cb18f522f1012933347fb5d9e62452d446baca2/pathspec-1.0.3.tar.gz", hash = "sha256:bac5cf97ae2c2876e2d25ebb15078eb04d76e4b98921ee31c6f85ade8b59444d", size = 130841 } wheels = [ - { url = "https://files.pythonhosted.org/packages/32/2b/121e912bd60eebd623f873fd090de0e84f322972ab25a7f9044c056804ed/pathspec-1.0.3-py3-none-any.whl", hash = "sha256:e80767021c1cc524aa3fb14bedda9c34406591343cc42797b386ce7b9354fb6c", size = 55021, upload-time = "2026-01-09T15:46:44.652Z" }, + { url = "https://files.pythonhosted.org/packages/32/2b/121e912bd60eebd623f873fd090de0e84f322972ab25a7f9044c056804ed/pathspec-1.0.3-py3-none-any.whl", hash = "sha256:e80767021c1cc524aa3fb14bedda9c34406591343cc42797b386ce7b9354fb6c", size = 55021 }, ] [[package]] @@ -6606,7 +6622,7 @@ requires-dist = [ { name = "chonkie", extras = ["all"], specifier = ">=1.5.0" }, { name = "composio", specifier = ">=0.10.9" }, { name = "datasets", specifier = ">=2.21.0" }, - { name = "deepagents", specifier = ">=0.3.0" }, + { name = "deepagents", specifier = ">=0.3.8" }, { name = "discord-py", specifier = ">=2.5.2" }, { name = "docling", specifier = ">=2.15.0" }, { name = "elasticsearch", specifier = ">=9.1.1" }, @@ -6622,7 +6638,7 @@ requires-dist = [ { name = "google-api-python-client", specifier = ">=2.156.0" }, { name = "google-auth-oauthlib", specifier = ">=1.2.1" }, { name = "kokoro", specifier = ">=0.9.4" }, - { name = "langchain", specifier = ">=1.2.0" }, + { name = "langchain", specifier = ">=1.2.6" }, { name = "langchain-community", specifier = ">=0.3.31" }, { name = "langchain-litellm", specifier = ">=0.3.5" }, { name = "langchain-unstructured", specifier = ">=1.0.0" }, diff --git a/surfsense_web/app/dashboard/[search_space_id]/layout.tsx b/surfsense_web/app/dashboard/[search_space_id]/layout.tsx index 1631f00b9..e434d85fb 100644 --- a/surfsense_web/app/dashboard/[search_space_id]/layout.tsx +++ b/surfsense_web/app/dashboard/[search_space_id]/layout.tsx @@ -39,12 +39,6 @@ export default function DashboardLayout({ icon: "SquareLibrary", items: [], }, - { - title: "Logs", - url: `/dashboard/${search_space_id}/logs`, - icon: "Logs", - items: [], - }, ]; return ( 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 88c532cf2..4d43b7f64 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 @@ -24,6 +24,7 @@ import { // extractWriteTodosFromContent, hydratePlanStateAtom, } from "@/atoms/chat/plan-state.atom"; +import { membersAtom } from "@/atoms/members/members-query.atoms"; import { currentUserAtom } from "@/atoms/user/user-query.atoms"; import { Thread } from "@/components/assistant-ui/thread"; import { ChatHeader } from "@/components/new-chat/chat-header"; @@ -32,7 +33,9 @@ import { DisplayImageToolUI } from "@/components/tool-ui/display-image"; import { GeneratePodcastToolUI } from "@/components/tool-ui/generate-podcast"; import { LinkPreviewToolUI } from "@/components/tool-ui/link-preview"; import { ScrapeWebpageToolUI } from "@/components/tool-ui/scrape-webpage"; -import { SaveMemoryToolUI, RecallMemoryToolUI } from "@/components/tool-ui/user-memory"; +import { RecallMemoryToolUI, SaveMemoryToolUI } from "@/components/tool-ui/user-memory"; +import { useChatSessionStateSync } from "@/hooks/use-chat-session-state"; +import { useMessagesElectric } from "@/hooks/use-messages-electric"; // import { WriteTodosToolUI } from "@/components/tool-ui/write-todos"; import { getBearerToken } from "@/lib/auth-utils"; import { createAttachmentAdapter, extractAttachmentContent } from "@/lib/chat/attachment-adapter"; @@ -258,6 +261,53 @@ export default function NewChatPage() { // Get current user for author info in shared chats const { data: currentUser } = useAtomValue(currentUserAtom); + // Live collaboration: sync session state and messages via Electric SQL + useChatSessionStateSync(threadId); + const { data: membersData } = useAtomValue(membersAtom); + + const handleElectricMessagesUpdate = useCallback( + ( + electricMessages: { + id: number; + thread_id: number; + role: string; + content: unknown; + author_id: string | null; + created_at: string; + }[] + ) => { + if (isRunning) { + return; + } + + setMessages((prev) => { + if (electricMessages.length < prev.length) { + return prev; + } + + return electricMessages.map((msg) => { + const member = msg.author_id + ? membersData?.find((m) => m.user_id === msg.author_id) + : null; + + return convertToThreadMessage({ + id: msg.id, + thread_id: msg.thread_id, + role: msg.role.toLowerCase() as "user" | "assistant" | "system", + content: msg.content, + author_id: msg.author_id, + created_at: msg.created_at, + author_display_name: member?.user_display_name ?? null, + author_avatar_url: member?.user_avatar_url ?? null, + }); + }); + }); + }, + [isRunning, membersData] + ); + + useMessagesElectric(threadId, handleElectricMessagesUpdate); + // Create the attachment adapter for file processing const attachmentAdapter = useMemo(() => createAttachmentAdapter(), []); @@ -587,8 +637,6 @@ export default function NewChatPage() { content: persistContent, }) .then(() => { - // For new threads, the backend updates the title from the first user message - // Invalidate threads query so sidebar shows the updated title in real-time if (isNewThread) { queryClient.invalidateQueries({ queryKey: ["threads", String(searchSpaceId)] }); } diff --git a/surfsense_web/atoms/chat/chat-session-state.atom.ts b/surfsense_web/atoms/chat/chat-session-state.atom.ts new file mode 100644 index 000000000..4d83a45d4 --- /dev/null +++ b/surfsense_web/atoms/chat/chat-session-state.atom.ts @@ -0,0 +1,15 @@ +"use client"; + +import { atom } from "jotai"; + +export interface ChatSessionStateData { + threadId: number; + isAiResponding: boolean; + respondingToUserId: string | null; +} + +/** + * Global chat session state atom. + * Updated by useChatSessionStateSync hook, read anywhere. + */ +export const chatSessionStateAtom = atom(null); diff --git a/surfsense_web/atoms/members/members-query.atoms.ts b/surfsense_web/atoms/members/members-query.atoms.ts index 8ed56ef0c..f486dc02b 100644 --- a/surfsense_web/atoms/members/members-query.atoms.ts +++ b/surfsense_web/atoms/members/members-query.atoms.ts @@ -9,7 +9,7 @@ export const membersAtom = atomWithQuery((get) => { return { queryKey: cacheKeys.members.all(searchSpaceId?.toString() ?? ""), enabled: !!searchSpaceId, - staleTime: 5 * 60 * 1000, // 5 minutes + staleTime: 3 * 1000, // 3 seconds - short staleness for live collaboration queryFn: async () => { if (!searchSpaceId) { return []; diff --git a/surfsense_web/components/assistant-ui/chat-session-status.tsx b/surfsense_web/components/assistant-ui/chat-session-status.tsx new file mode 100644 index 000000000..88fea6b8c --- /dev/null +++ b/surfsense_web/components/assistant-ui/chat-session-status.tsx @@ -0,0 +1,50 @@ +"use client"; + +import { Loader2 } from "lucide-react"; +import type { FC } from "react"; +import { cn } from "@/lib/utils"; + +interface ChatSessionStatusProps { + isAiResponding: boolean; + respondingToUserId: string | null; + currentUserId: string | null; + members: Array<{ + user_id: string; + user_display_name?: string | null; + user_email?: string | null; + }>; + className?: string; +} + +export const ChatSessionStatus: FC = ({ + isAiResponding, + respondingToUserId, + currentUserId, + members, + className, +}) => { + if (!isAiResponding || !respondingToUserId) { + return null; + } + + if (respondingToUserId === currentUserId) { + return null; + } + + const respondingUser = members.find((m) => m.user_id === respondingToUserId); + const displayName = + respondingUser?.user_display_name || respondingUser?.user_email || "another user"; + + return ( +
+ + Currently responding to {displayName} +
+ ); +}; diff --git a/surfsense_web/components/assistant-ui/connector-popup/connect-forms/components/obsidian-connect-form.tsx b/surfsense_web/components/assistant-ui/connector-popup/connect-forms/components/obsidian-connect-form.tsx new file mode 100644 index 000000000..064e10e2f --- /dev/null +++ b/surfsense_web/components/assistant-ui/connector-popup/connect-forms/components/obsidian-connect-form.tsx @@ -0,0 +1,464 @@ +"use client"; + +import { zodResolver } from "@hookform/resolvers/zod"; +import { FolderOpen, Info } from "lucide-react"; +import type { FC } from "react"; +import { useRef, useState } from "react"; +import { useForm } from "react-hook-form"; +import * as z from "zod"; +import { + Accordion, + AccordionContent, + AccordionItem, + AccordionTrigger, +} from "@/components/ui/accordion"; +import { Alert, AlertDescription, AlertTitle } from "@/components/ui/alert"; +import { + Form, + FormControl, + FormDescription, + FormField, + FormItem, + FormLabel, + FormMessage, +} from "@/components/ui/form"; +import { Input } from "@/components/ui/input"; +import { Label } from "@/components/ui/label"; +import { + Select, + SelectContent, + SelectItem, + SelectTrigger, + SelectValue, +} from "@/components/ui/select"; +import { Switch } from "@/components/ui/switch"; +import { EnumConnectorName } from "@/contracts/enums/connector"; +import { getConnectorBenefits } from "../connector-benefits"; +import type { ConnectFormProps } from "../index"; + +const obsidianConnectorFormSchema = z.object({ + name: z.string().min(3, { + message: "Connector name must be at least 3 characters.", + }), + vault_path: z.string().min(1, { + message: "Vault path is required.", + }), + vault_name: z.string().min(1, { + message: "Vault name is required.", + }), + exclude_folders: z.string().optional(), + include_attachments: z.boolean(), +}); + +type ObsidianConnectorFormValues = z.infer; + +export const ObsidianConnectForm: FC = ({ onSubmit, isSubmitting }) => { + const isSubmittingRef = useRef(false); + const [periodicEnabled, setPeriodicEnabled] = useState(true); + const [frequencyMinutes, setFrequencyMinutes] = useState("60"); + const form = useForm({ + resolver: zodResolver(obsidianConnectorFormSchema), + defaultValues: { + name: "Obsidian Vault", + vault_path: "", + vault_name: "", + exclude_folders: ".obsidian,.trash", + include_attachments: false, + }, + }); + + const handleSubmit = async (values: ObsidianConnectorFormValues) => { + // Prevent multiple submissions + if (isSubmittingRef.current || isSubmitting) { + return; + } + + isSubmittingRef.current = true; + try { + // Parse exclude_folders into an array + const excludeFolders = values.exclude_folders + ? values.exclude_folders + .split(",") + .map((f) => f.trim()) + .filter(Boolean) + : [".obsidian", ".trash"]; + + await onSubmit({ + name: values.name, + connector_type: EnumConnectorName.OBSIDIAN_CONNECTOR, + config: { + vault_path: values.vault_path, + vault_name: values.vault_name, + exclude_folders: excludeFolders, + include_attachments: values.include_attachments, + }, + is_indexable: true, + is_active: true, + last_indexed_at: null, + periodic_indexing_enabled: periodicEnabled, + indexing_frequency_minutes: periodicEnabled ? Number.parseInt(frequencyMinutes, 10) : null, + next_scheduled_at: null, + periodicEnabled, + frequencyMinutes, + }); + } finally { + isSubmittingRef.current = false; + } + }; + + return ( +
+ + +
+ Self-Hosted Only + + This connector requires direct file system access and only works with self-hosted + SurfSense installations. + +
+ + +
+
+ + ( + + Connector Name + + + + + A friendly name to identify this connector. + + + + )} + /> + + ( + + Vault Path + + + + + The absolute path to your Obsidian vault on the server. This must be accessible + from the SurfSense backend. + + + + )} + /> + + ( + + Vault Name + + + + + A display name for your vault. This will be used in search results. + + + + )} + /> + + ( + + Exclude Folders + + + + + Comma-separated list of folder names to exclude from indexing. + + + + )} + /> + + ( + +
+ Include Attachments + + Index attachment folders and embedded files (images, PDFs, etc.) + +
+ + + + + )} + /> + + {/* Indexing Configuration */} +
+

Indexing Configuration

+ + {/* Periodic Sync Config */} +
+
+
+

Enable Periodic Sync

+

+ Automatically re-index at regular intervals +

+
+ +
+ + {periodicEnabled && ( +
+
+ + +
+
+ )} +
+
+ + +
+ + {/* What you get section */} + {getConnectorBenefits(EnumConnectorName.OBSIDIAN_CONNECTOR) && ( +
+

+ What you get with Obsidian integration: +

+
    + {getConnectorBenefits(EnumConnectorName.OBSIDIAN_CONNECTOR)?.map((benefit) => ( +
  • {benefit}
    • + ))} +
+
+ )} + + {/* Documentation Section */} + + + + Documentation + + +
+

How it works

+

+ The Obsidian connector scans your local Obsidian vault directory and indexes all + Markdown files. It preserves your note structure and extracts metadata from YAML + frontmatter. +

+
    +
  • + The connector parses frontmatter metadata (title, tags, aliases, dates, etc.) +
    • +
  • Wiki-style links ([[note]]) are extracted and preserved
    • +
  • Inline tags (#tag) are recognized and indexed
    • +
  • Content is chunked intelligently for optimal search results
    • +
  • + Subsequent indexing runs use content hashing to skip unchanged files for faster + sync +
    • +
+
+ +
+
+

Setup

+ + + + File System Access Required + + + The SurfSense backend must have read access to your Obsidian vault directory. + For Docker deployments, mount your vault as a volume. + + + +
+
+

+ Step 1: Locate your vault +

+
    +
  1. + macOS/Linux: Right-click any note in Obsidian → "Reveal in + Finder" to see the vault folder +
    2. +
  2. + Windows: Right-click any note → "Show in system explorer" +
    3. +
  3. + Or: Click the vault switcher (bottom-left icon) → "Open + folder" next to your vault name +
    4. +
+
+ +
+

+ Step 2: Enter the path +

+

+ Running locally (no Docker): Use the direct path to your + vault: +

+ 
+											{`/Users/yourname/Documents/MyObsidianVault`}
+
+

+ Running in Docker: Mount your vault as a volume in + docker-compose.yml: +

+ 
+											{`volumes:
+  - /path/to/your/vault:/app/obsidian_vaults/my-vault:ro`}
+
+

+ Then use /app/obsidian_vaults/my-vault as your vault path. +

+
+ +
+

+ Step 3: Configure exclusions +

+

+ Common folders to exclude: +

+
    +
  • + .obsidian - Obsidian config (always recommended) +
    • +
  • + .trash - Obsidian's trash folder +
    • +
  • + templates - If you have a templates folder +
    • +
  • + daily-notes - If you want to exclude daily notes +
    • +
+
+
+
+
+ +
+
+

What Gets Indexed

+ + + Indexed Content + +

The Obsidian connector indexes:

+
    +
  • All Markdown files (.md) in your vault
    • +
  • YAML frontmatter metadata (title, tags, aliases, dates)
    • +
  • Wiki-style links between notes
    • +
  • Inline tags throughout your notes
    • +
  • Full note content with proper chunking
    • +
+ + +
+
+ + + +
+ ); +}; diff --git a/surfsense_web/components/assistant-ui/connector-popup/connect-forms/connector-benefits.ts b/surfsense_web/components/assistant-ui/connector-popup/connect-forms/connector-benefits.ts index 9a74132ed..392de4bc8 100644 --- a/surfsense_web/components/assistant-ui/connector-popup/connect-forms/connector-benefits.ts +++ b/surfsense_web/components/assistant-ui/connector-popup/connect-forms/connector-benefits.ts @@ -108,6 +108,14 @@ export function getConnectorBenefits(connectorType: string): string[] | null { "Real-time updates via webhook integration", "No manual indexing required - meetings are added automatically", ], + OBSIDIAN_CONNECTOR: [ + "Search through all your Obsidian notes and knowledge base", + "Access note content with YAML frontmatter metadata preserved", + "Wiki-style links ([[note]]) and #tags are indexed", + "Connect your personal knowledge base directly to your search space", + "Incremental sync - only changed files are re-indexed", + "Full support for your vault's folder structure", + ], }; return benefits[connectorType] || null; diff --git a/surfsense_web/components/assistant-ui/connector-popup/connect-forms/index.tsx b/surfsense_web/components/assistant-ui/connector-popup/connect-forms/index.tsx index 4710f9ad5..ffaeb1478 100644 --- a/surfsense_web/components/assistant-ui/connector-popup/connect-forms/index.tsx +++ b/surfsense_web/components/assistant-ui/connector-popup/connect-forms/index.tsx @@ -7,6 +7,7 @@ import { GithubConnectForm } from "./components/github-connect-form"; import { LinkupApiConnectForm } from "./components/linkup-api-connect-form"; import { LumaConnectForm } from "./components/luma-connect-form"; import { MCPConnectForm } from "./components/mcp-connect-form"; +import { ObsidianConnectForm } from "./components/obsidian-connect-form"; import { SearxngConnectForm } from "./components/searxng-connect-form"; import { TavilyApiConnectForm } from "./components/tavily-api-connect-form"; @@ -58,6 +59,8 @@ export function getConnectFormComponent(connectorType: string): ConnectFormCompo return CirclebackConnectForm; case "MCP_CONNECTOR": return MCPConnectForm; + case "OBSIDIAN_CONNECTOR": + return ObsidianConnectForm; // Add other connector types here as needed default: return null; diff --git a/surfsense_web/components/assistant-ui/connector-popup/connector-configs/components/obsidian-config.tsx b/surfsense_web/components/assistant-ui/connector-popup/connector-configs/components/obsidian-config.tsx new file mode 100644 index 000000000..f79f2953a --- /dev/null +++ b/surfsense_web/components/assistant-ui/connector-popup/connector-configs/components/obsidian-config.tsx @@ -0,0 +1,191 @@ +"use client"; + +import { FolderOpen } from "lucide-react"; +import type { FC } from "react"; +import { useEffect, useState } from "react"; +import { Input } from "@/components/ui/input"; +import { Label } from "@/components/ui/label"; +import { Switch } from "@/components/ui/switch"; +import type { ConnectorConfigProps } from "../index"; + +export interface ObsidianConfigProps extends ConnectorConfigProps { + onNameChange?: (name: string) => void; +} + +export const ObsidianConfig: FC = ({ + connector, + onConfigChange, + onNameChange, +}) => { + const [vaultPath, setVaultPath] = useState( + (connector.config?.vault_path as string) || "" + ); + const [vaultName, setVaultName] = useState( + (connector.config?.vault_name as string) || "" + ); + const [excludeFolders, setExcludeFolders] = useState(() => { + const folders = connector.config?.exclude_folders; + if (Array.isArray(folders)) { + return folders.join(", "); + } + return (folders as string) || ".obsidian, .trash"; + }); + const [includeAttachments, setIncludeAttachments] = useState( + (connector.config?.include_attachments as boolean) || false + ); + const [name, setName] = useState(connector.name || ""); + + // Update values when connector changes + useEffect(() => { + const path = (connector.config?.vault_path as string) || ""; + const vName = (connector.config?.vault_name as string) || ""; + const folders = connector.config?.exclude_folders; + const attachments = (connector.config?.include_attachments as boolean) || false; + + setVaultPath(path); + setVaultName(vName); + setIncludeAttachments(attachments); + setName(connector.name || ""); + + if (Array.isArray(folders)) { + setExcludeFolders(folders.join(", ")); + } else if (typeof folders === "string") { + setExcludeFolders(folders); + } + }, [connector.config, connector.name]); + + const handleVaultPathChange = (value: string) => { + setVaultPath(value); + if (onConfigChange) { + onConfigChange({ + ...connector.config, + vault_path: value, + }); + } + }; + + const handleVaultNameChange = (value: string) => { + setVaultName(value); + if (onConfigChange) { + onConfigChange({ + ...connector.config, + vault_name: value, + }); + } + }; + + const handleExcludeFoldersChange = (value: string) => { + setExcludeFolders(value); + const foldersArray = value + .split(",") + .map((f) => f.trim()) + .filter(Boolean); + if (onConfigChange) { + onConfigChange({ + ...connector.config, + exclude_folders: foldersArray, + }); + } + }; + + const handleIncludeAttachmentsChange = (value: boolean) => { + setIncludeAttachments(value); + if (onConfigChange) { + onConfigChange({ + ...connector.config, + include_attachments: value, + }); + } + }; + + const handleNameChange = (value: string) => { + setName(value); + if (onNameChange) { + onNameChange(value); + } + }; + + return ( +
+ {/* Connector Name */} +
+
+ + handleNameChange(e.target.value)} + placeholder="My Obsidian Vault" + className="border-slate-400/20 focus-visible:border-slate-400/40" + /> +

+ A friendly name to identify this connector. +

+
+
+ + {/* Configuration */} +
+
+

+ + Vault Configuration +

+
+ +
+
+ + handleVaultPathChange(e.target.value)} + placeholder="/path/to/your/obsidian/vault" + className="border-slate-400/20 focus-visible:border-slate-400/40 font-mono" + /> +

+ The absolute path to your Obsidian vault on the server. +

+
+ +
+ + handleVaultNameChange(e.target.value)} + placeholder="My Knowledge Base" + className="border-slate-400/20 focus-visible:border-slate-400/40" + /> +

+ A display name for your vault in search results. +

+
+ +
+ + handleExcludeFoldersChange(e.target.value)} + placeholder=".obsidian, .trash, templates" + className="border-slate-400/20 focus-visible:border-slate-400/40 font-mono" + /> +

+ Comma-separated list of folder names to exclude from indexing. +

+
+ +
+
+ +

+ Index attachment folders and embedded files +

+
+ +
+
+
+
+ ); +}; diff --git a/surfsense_web/components/assistant-ui/connector-popup/connector-configs/index.tsx b/surfsense_web/components/assistant-ui/connector-popup/connector-configs/index.tsx index 160185b1e..1a713a5a0 100644 --- a/surfsense_web/components/assistant-ui/connector-popup/connector-configs/index.tsx +++ b/surfsense_web/components/assistant-ui/connector-popup/connector-configs/index.tsx @@ -5,8 +5,8 @@ import type { SearchSourceConnector } from "@/contracts/types/connector.types"; import { BaiduSearchApiConfig } from "./components/baidu-search-api-config"; import { BookStackConfig } from "./components/bookstack-config"; import { CirclebackConfig } from "./components/circleback-config"; -import { ComposioConfig } from "./components/composio-config"; import { ClickUpConfig } from "./components/clickup-config"; +import { ComposioConfig } from "./components/composio-config"; import { ConfluenceConfig } from "./components/confluence-config"; import { DiscordConfig } from "./components/discord-config"; import { ElasticsearchConfig } from "./components/elasticsearch-config"; @@ -16,6 +16,7 @@ import { JiraConfig } from "./components/jira-config"; import { LinkupApiConfig } from "./components/linkup-api-config"; import { LumaConfig } from "./components/luma-config"; import { MCPConfig } from "./components/mcp-config"; +import { ObsidianConfig } from "./components/obsidian-config"; import { SearxngConfig } from "./components/searxng-config"; import { SlackConfig } from "./components/slack-config"; import { TavilyApiConfig } from "./components/tavily-api-config"; @@ -74,6 +75,8 @@ export function getConnectorConfigComponent( return CirclebackConfig; case "MCP_CONNECTOR": return MCPConfig; + case "OBSIDIAN_CONNECTOR": + return ObsidianConfig; case "COMPOSIO_GOOGLE_DRIVE_CONNECTOR": case "COMPOSIO_GMAIL_CONNECTOR": case "COMPOSIO_GOOGLE_CALENDAR_CONNECTOR": diff --git a/surfsense_web/components/assistant-ui/connector-popup/connector-configs/views/connector-connect-view.tsx b/surfsense_web/components/assistant-ui/connector-popup/connector-configs/views/connector-connect-view.tsx index 7666b775f..49d1ebacc 100644 --- a/surfsense_web/components/assistant-ui/connector-popup/connector-configs/views/connector-connect-view.tsx +++ b/surfsense_web/components/assistant-ui/connector-popup/connector-configs/views/connector-connect-view.tsx @@ -57,6 +57,7 @@ export const ConnectorConnectView: FC = ({ LUMA_CONNECTOR: "luma-connect-form", CIRCLEBACK_CONNECTOR: "circleback-connect-form", MCP_CONNECTOR: "mcp-connect-form", + OBSIDIAN_CONNECTOR: "obsidian-connect-form", }; const formId = formIdMap[connectorType]; if (formId) { @@ -141,12 +142,10 @@ export const ConnectorConnectView: FC = ({ Connecting + ) : connectorType === "MCP_CONNECTOR" ? ( + "Connect" ) : ( - <> - {connectorType === "MCP_CONNECTOR" - ? "Connect" - : `Connect ${getConnectorTypeDisplay(connectorType)}`} - + `Connect ${getConnectorTypeDisplay(connectorType)}` )} diff --git a/surfsense_web/components/assistant-ui/connector-popup/constants/connector-constants.ts b/surfsense_web/components/assistant-ui/connector-popup/constants/connector-constants.ts index 11066f28a..a3e8ae272 100644 --- a/surfsense_web/components/assistant-ui/connector-popup/constants/connector-constants.ts +++ b/surfsense_web/components/assistant-ui/connector-popup/constants/connector-constants.ts @@ -166,6 +166,13 @@ export const OTHER_CONNECTORS = [ description: "Connect to MCP servers for AI tools", connectorType: EnumConnectorName.MCP_CONNECTOR, }, + { + id: "obsidian-connector", + title: "Obsidian", + description: "Index your Obsidian vault (self-hosted only)", + connectorType: EnumConnectorName.OBSIDIAN_CONNECTOR, + selfHostedOnly: true, + }, ] as const; // Composio Connectors - Individual entries for each supported toolkit diff --git a/surfsense_web/components/assistant-ui/connector-popup/tabs/all-connectors-tab.tsx b/surfsense_web/components/assistant-ui/connector-popup/tabs/all-connectors-tab.tsx index 6b38a37d2..0268ab761 100644 --- a/surfsense_web/components/assistant-ui/connector-popup/tabs/all-connectors-tab.tsx +++ b/surfsense_web/components/assistant-ui/connector-popup/tabs/all-connectors-tab.tsx @@ -3,12 +3,13 @@ import type { FC } from "react"; import { EnumConnectorName } from "@/contracts/enums/connector"; import type { SearchSourceConnector } from "@/contracts/types/connector.types"; +import { isSelfHosted } from "@/lib/env-config"; import { ConnectorCard } from "../components/connector-card"; import { + COMPOSIO_CONNECTORS, CRAWLERS, OAUTH_CONNECTORS, OTHER_CONNECTORS, - COMPOSIO_CONNECTORS, } from "../constants/connector-constants"; import { getDocumentCountForConnector } from "../utils/connector-document-mapping"; @@ -57,23 +58,31 @@ export const AllConnectorsTab: FC = ({ onManage, onViewAccountsList, }) => { - // Filter connectors based on search + // Check if self-hosted mode (for showing self-hosted only connectors) + const selfHosted = isSelfHosted(); + + // Filter connectors based on search and deployment mode const filteredOAuth = OAUTH_CONNECTORS.filter( (c) => - c.title.toLowerCase().includes(searchQuery.toLowerCase()) || - c.description.toLowerCase().includes(searchQuery.toLowerCase()) + // Filter by search query + (c.title.toLowerCase().includes(searchQuery.toLowerCase()) || + c.description.toLowerCase().includes(searchQuery.toLowerCase())) && + // Filter self-hosted only connectors in cloud mode + (!("selfHostedOnly" in c) || !c.selfHostedOnly || selfHosted) ); const filteredCrawlers = CRAWLERS.filter( (c) => - c.title.toLowerCase().includes(searchQuery.toLowerCase()) || - c.description.toLowerCase().includes(searchQuery.toLowerCase()) + (c.title.toLowerCase().includes(searchQuery.toLowerCase()) || + c.description.toLowerCase().includes(searchQuery.toLowerCase())) && + (!("selfHostedOnly" in c) || !c.selfHostedOnly || selfHosted) ); const filteredOther = OTHER_CONNECTORS.filter( (c) => - c.title.toLowerCase().includes(searchQuery.toLowerCase()) || - c.description.toLowerCase().includes(searchQuery.toLowerCase()) + (c.title.toLowerCase().includes(searchQuery.toLowerCase()) || + c.description.toLowerCase().includes(searchQuery.toLowerCase())) && + (!("selfHostedOnly" in c) || !c.selfHostedOnly || selfHosted) ); // Filter Composio connectors diff --git a/surfsense_web/components/assistant-ui/connector-popup/utils/connector-document-mapping.ts b/surfsense_web/components/assistant-ui/connector-popup/utils/connector-document-mapping.ts index ded3bdcca..090207bbb 100644 --- a/surfsense_web/components/assistant-ui/connector-popup/utils/connector-document-mapping.ts +++ b/surfsense_web/components/assistant-ui/connector-popup/utils/connector-document-mapping.ts @@ -26,6 +26,7 @@ export const CONNECTOR_TO_DOCUMENT_TYPE: Record = { ELASTICSEARCH_CONNECTOR: "ELASTICSEARCH_CONNECTOR", BOOKSTACK_CONNECTOR: "BOOKSTACK_CONNECTOR", CIRCLEBACK_CONNECTOR: "CIRCLEBACK", + OBSIDIAN_CONNECTOR: "OBSIDIAN_CONNECTOR", // Special mappings (connector type differs from document type) GOOGLE_DRIVE_CONNECTOR: "GOOGLE_DRIVE_FILE", diff --git a/surfsense_web/components/assistant-ui/thread.tsx b/surfsense_web/components/assistant-ui/thread.tsx index eaf30fc96..042b932a4 100644 --- a/surfsense_web/components/assistant-ui/thread.tsx +++ b/surfsense_web/components/assistant-ui/thread.tsx @@ -26,11 +26,13 @@ import { import { useParams } from "next/navigation"; import { type FC, useCallback, useContext, useEffect, useMemo, useRef, useState } from "react"; import { createPortal } from "react-dom"; +import { chatSessionStateAtom } from "@/atoms/chat/chat-session-state.atom"; import { showCommentsGutterAtom } from "@/atoms/chat/current-thread.atom"; import { mentionedDocumentIdsAtom, mentionedDocumentsAtom, } from "@/atoms/chat/mentioned-documents.atom"; +import { membersAtom } from "@/atoms/members/members-query.atoms"; import { globalNewLLMConfigsAtom, llmPreferencesAtom, @@ -39,6 +41,7 @@ import { import { currentUserAtom } from "@/atoms/user/user-query.atoms"; import { AssistantMessage } from "@/components/assistant-ui/assistant-message"; import { ComposerAddAttachment, ComposerAttachments } from "@/components/assistant-ui/attachment"; +import { ChatSessionStatus } from "@/components/assistant-ui/chat-session-status"; import { ConnectorIndicator } from "@/components/assistant-ui/connector-popup"; import { InlineMentionEditor, @@ -59,6 +62,7 @@ import { import type { ThinkingStep } from "@/components/tool-ui/deepagent-thinking"; import { Button } from "@/components/ui/button"; import type { Document } from "@/contracts/types/document.types"; +import { useCommentsElectric } from "@/hooks/use-comments-electric"; import { cn } from "@/lib/utils"; interface ThreadProps { @@ -86,6 +90,7 @@ const ThreadContent: FC<{ header?: React.ReactNode }> = ({ header }) => { > { const editorRef = useRef(null); const editorContainerRef = useRef(null); const documentPickerRef = useRef(null); - const { search_space_id } = useParams(); + const { search_space_id, chat_id } = useParams(); const setMentionedDocumentIds = useSetAtom(mentionedDocumentIdsAtom); const composerRuntime = useComposerRuntime(); const hasAutoFocusedRef = useRef(false); @@ -223,6 +228,23 @@ const Composer: FC = () => { const isThreadEmpty = useAssistantState(({ thread }) => thread.isEmpty); const isThreadRunning = useAssistantState(({ thread }) => thread.isRunning); + // Live collaboration state + const { data: currentUser } = useAtomValue(currentUserAtom); + const { data: members } = useAtomValue(membersAtom); + const threadId = useMemo(() => { + if (Array.isArray(chat_id) && chat_id.length > 0) { + return Number.parseInt(chat_id[0], 10) || null; + } + return typeof chat_id === "string" ? Number.parseInt(chat_id, 10) || null : null; + }, [chat_id]); + const sessionState = useAtomValue(chatSessionStateAtom); + const isAiResponding = sessionState?.isAiResponding ?? false; + const respondingToUserId = sessionState?.respondingToUserId ?? null; + const isBlockedByOtherUser = isAiResponding && respondingToUserId !== currentUser?.id; + + // Sync comments for the entire thread via Electric SQL (one subscription per thread) + useCommentsElectric(threadId); + // Auto-focus editor on new chat page after mount useEffect(() => { if (isThreadEmpty && !hasAutoFocusedRef.current && editorRef.current) { @@ -298,9 +320,9 @@ const Composer: FC = () => { [showDocumentPopover] ); - // Submit message (blocked during streaming or when document picker is open) + // Submit message (blocked during streaming, document picker open, or AI responding to another user) const handleSubmit = useCallback(() => { - if (isThreadRunning) { + if (isThreadRunning || isBlockedByOtherUser) { return; } if (!showDocumentPopover) { @@ -315,6 +337,7 @@ const Composer: FC = () => { }, [ showDocumentPopover, isThreadRunning, + isBlockedByOtherUser, composerRuntime, setMentionedDocuments, setMentionedDocumentIds, @@ -374,7 +397,13 @@ const Composer: FC = () => { ); return ( - + + {/* Inline editor with @mention support */} @@ -417,13 +446,17 @@ const Composer: FC = () => { />, document.body )} - + ); }; -const ComposerAction: FC = () => { +interface ComposerActionProps { + isBlockedByOtherUser?: boolean; +} + +const ComposerAction: FC = ({ isBlockedByOtherUser = false }) => { // Check if any attachments are still being processed (running AND progress < 100) // When progress is 100, processing is done but waiting for send() const hasProcessingAttachments = useAssistantState(({ composer }) => @@ -458,7 +491,8 @@ const ComposerAction: FC = () => { return userConfigs?.some((c) => c.id === agentLlmId) ?? false; }, [preferences, globalConfigs, userConfigs]); - const isSendDisabled = hasProcessingAttachments || isComposerEmpty || !hasModelConfigured; + const isSendDisabled = + hasProcessingAttachments || isComposerEmpty || !hasModelConfigured || isBlockedByOtherUser; return (
@@ -487,13 +521,15 @@ const ComposerAction: FC = () => { - {/* Platform navigation */} - {navItems.length > 0 && ( - - )} - {/* Scrollable content */} {isCollapsed ? ( @@ -235,7 +230,12 @@ export function Sidebar({ {/* Footer */} -
+
+ {/* Platform navigation */} + {navItems.length > 0 && ( + + )} + {pageUsage && !isCollapsed && ( )} diff --git a/surfsense_web/components/layout/ui/sidebar/SidebarHeader.tsx b/surfsense_web/components/layout/ui/sidebar/SidebarHeader.tsx index 6ca057819..5c8c89152 100644 --- a/surfsense_web/components/layout/ui/sidebar/SidebarHeader.tsx +++ b/surfsense_web/components/layout/ui/sidebar/SidebarHeader.tsx @@ -1,6 +1,7 @@ "use client"; -import { ChevronsUpDown, Settings, Users } from "lucide-react"; +import { ChevronsUpDown, ScrollText, Settings, Users } from "lucide-react"; +import { useParams, useRouter } from "next/navigation"; import { useTranslations } from "next-intl"; import { Button } from "@/components/ui/button"; import { @@ -29,6 +30,9 @@ export function SidebarHeader({ className, }: SidebarHeaderProps) { const t = useTranslations("sidebar"); + const router = useRouter(); + const params = useParams(); + const searchSpaceId = params.search_space_id as string; return (
@@ -52,6 +56,10 @@ export function SidebarHeader({ {t("manage_members")} + router.push(`/dashboard/${searchSpaceId}/logs`)}> + + {t("logs")} + diff --git a/surfsense_web/components/layout/ui/sidebar/SidebarUserProfile.tsx b/surfsense_web/components/layout/ui/sidebar/SidebarUserProfile.tsx index 907fd89f8..982d88e8b 100644 --- a/surfsense_web/components/layout/ui/sidebar/SidebarUserProfile.tsx +++ b/surfsense_web/components/layout/ui/sidebar/SidebarUserProfile.tsx @@ -1,6 +1,6 @@ "use client"; -import { ChevronUp, Laptop, Languages, LogOut, Moon, Settings, Sun } from "lucide-react"; +import { ChevronUp, Languages, Laptop, LogOut, Moon, Settings, Sun } from "lucide-react"; import { useTranslations } from "next-intl"; import { DropdownMenu, diff --git a/surfsense_web/components/new-chat/model-selector.tsx b/surfsense_web/components/new-chat/model-selector.tsx index af9378e34..bc7591c51 100644 --- a/surfsense_web/components/new-chat/model-selector.tsx +++ b/surfsense_web/components/new-chat/model-selector.tsx @@ -123,6 +123,11 @@ export function ModelSelector({ onEdit, onAddNew, className }: ModelSelectorProp ); }, [userConfigs, searchQuery]); + // Total model count for conditional search display + const totalModels = useMemo(() => { + return (globalConfigs?.length ?? 0) + (userConfigs?.length ?? 0); + }, [globalConfigs, userConfigs]); + const handleSelectConfig = useCallback( async (config: NewLLMConfigPublic | GlobalNewLLMConfig) => { // If already selected, just close @@ -212,14 +217,16 @@ export function ModelSelector({ onEdit, onAddNew, className }: ModelSelectorProp shouldFilter={false} className="rounded-lg relative [&_[data-slot=command-input-wrapper]]:border-0 [&_[data-slot=command-input-wrapper]]:px-0 [&_[data-slot=command-input-wrapper]]:gap-2" > -
- -
+ {totalModels > 3 && ( +
+ +
+ )} @@ -245,7 +252,7 @@ export function ModelSelector({ onEdit, onAddNew, className }: ModelSelectorProp value={`global-${config.id}`} onSelect={() => handleSelectConfig(config)} className={cn( - "mx-2 rounded-lg mb-1 cursor-pointer transition-all", + "mx-2 rounded-lg mb-1 cursor-pointer group transition-all", "hover:bg-accent/50", isSelected && "bg-accent/80" )} @@ -276,7 +283,7 @@ export function ModelSelector({ onEdit, onAddNew, className }: ModelSelectorProp