- Added new environment variables for controlling task execution limits, including `SURFSENSE_SUBAGENT_INVOKE_TIMEOUT_SECONDS`, `SURFSENSE_TASK_BATCH_CONCURRENCY`, and `SURFSENSE_TASK_BATCH_MAX_SIZE`.
- Updated documentation to reflect new batch processing capabilities for `task` calls, allowing for concurrent execution of multiple subagent tasks.
- Improved error handling and receipt generation for deliverables, ensuring consistent feedback on task status.
- Refactored middleware to incorporate search space ID for better task management.
Slim composition root for the resume-chat streaming flow. Mirrors the
new_chat orchestrator but specialized for resumed turns:
* no fresh user turn, no title generation, no image-capability gate
* persists a fresh assistant shell for the resumed turn
* applies build_resume_routing to dispatch user decisions to the
correct paused subagent before invoking the agent
* shares the same stream_loop + flow-local _recover closure for in-
stream provider rate-limit recovery
Also lands flows/__init__.py, which becomes the public chat-flow API:
from app.tasks.chat.streaming.flows import stream_new_chat, stream_resume_chat
Existing wiring (routes, contract test) still imports from the legacy
app.tasks.chat.stream_new_chat module. Cutover is the next phase.
Three focused modules used by the upcoming resume-chat orchestrator:
* runtime_context: build_resume_chat_runtime_context assembles the
SurfSenseContextSchema for a resume turn (handles empty mention
lists, since resume requests do not carry fresh @-mentions).
* assistant_shell: persist_resume_assistant_shell writes a fresh
assistant row for the resumed turn so the post-stream finalize
has a target.
* resume_routing: build_resume_routing collects the pending
interrupts across paused subagents and slices the flat list of
ResumeDecision[] into the correct (thread, subagent) buckets so
LangGraph routes each decision back to the right paused tool call.
Add-only; no orchestrator yet (next commit).
Slim composition root for the new-chat streaming flow. Sequences:
1. validate inputs and load the LLM bundle (negative id => YAML)
2. open the OTEL chat_request span; set agent_mode tag
3. spawn the four pre-stream DB writes (set-ai-responding, persist
user turn, persist assistant shell, first-assistant probe)
4. reserve premium quota (with free-fallback retry on denial)
5. build connector + checkpointer + agent + input_state
6. emit first frames (message-start, step-start, initial thinking step)
7. spawn the background title generator
8. run the shared stream_loop with a flow-local _recover closure that
reroutes to the next auto-pin config on provider 429s
9. finalize: emit terminal title/token frames, shielded assistant
finalize, release-or-finalize premium quota, close session, GC,
record OTEL outcome
Public entry-point flows/new_chat/__init__ re-exports stream_new_chat.
Existing wiring (routes, tests) still imports the legacy function from
app.tasks.chat.stream_new_chat. Cutover is a later commit.
Seven focused modules that the upcoming new_chat orchestrator
composes:
* auto_pin: resolve_initial_auto_pin selects the initial config (with
vision-capable filtering and error classification).
* llm_capability: check_image_input_capability blocks routing an
image-bearing turn to a known text-only model.
* runtime_context: build_new_chat_runtime_context assembles the
SurfSenseContextSchema for a new-chat turn.
* persistence_spawn: spawn_set_ai_responding_bg, spawn_persist_user_task,
spawn_persist_assistant_shell_task, and await_persist_task background
the four pre-stream DB writes so they overlap with agent build.
* initial_thinking_step: build_initial_thinking_step +
iter_initial_thinking_step_frame produce the very first thinking-1 SSE
step ("Understanding your request" / "Analyzing referenced content").
* title_gen: spawn_title_task + maybe_emit_title_update +
await_pending_title_update background the thread-title generator and
interleave its update into the stream when ready.
* input_state: build_new_chat_input_state assembles the LangGraph
input_state (history bootstrap, mentions resolution, context blocks,
human-message construction). The heavy one.
Add-only; no orchestrator yet (next commit).
Extracts finalize_assistant_message: the post-stream server-side write
of the final assistant message (with content parts + token usage)
guarded by asyncio.shield + shielded_async_session so a client
disconnect cannot abort the persist.
Add-only; legacy stream_new_chat.py keeps its inline finalize block
until cutover.
Two cooperating modules that wrap stream_agent_events with in-stream
recovery from provider 429s:
* rate_limit_recovery: can_recover_provider_rate_limit truth-table
guard, reroute_to_next_auto_pin (selects the next eligible auto-pin
config and reloads the LLM bundle), log_rate_limit_recovered.
* stream_loop: run_stream_loop drives stream_agent_events in a
while-True loop, delegating recovery to a flow-supplied RecoverFn
callback so new_chat and resume_chat can share the same loop while
keeping their own nonlocal state.
Add-only; not yet wired into any orchestrator.
Extracts handle_terminal_exception: the shared except-branch behavior for
the chat orchestrators. Classifies the raised exception, logs the
structured chat_stream error event, and emits the terminal-error SSE
frame + done sentinel via the streaming service.
Add-only; nothing imports it yet.
Centralizes the premium-credits lifecycle for chat turns:
* needs_premium_quota: gate check (premium user + non-fallback config).
* PremiumReservation: dataclass capturing reservation state + token totals.
* reserve_premium / finalize_premium / release_premium: idempotent
reservation, commit, and rollback used by the orchestrators.
Add-only; legacy stream_new_chat.py keeps its inline quota handling
until cutover.
Six small, single-purpose modules shared by the upcoming new_chat and
resume_chat orchestrators:
* llm_bundle: dispatches negative config_id to the YAML loader and
non-negative config_id to the DB loader, returning (llm, AgentConfig).
* pre_stream_setup: builds the connector service, resolves the
Firecrawl API key, and returns the chat checkpointer.
* first_frames: iter_initial_frames + iter_final_frames emit the canonical
message-start / step-start / idle / finish / done SSE envelope.
* finalize_emit: iter_token_usage_frame emits the per-turn usage frame
from a TokenAccumulator summary.
* finally_cleanup: close_session_and_clear_ai_responding and run_gc_pass
centralize the finally-block bookkeeping.
* span: open_chat_request_span / set_agent_mode / close_chat_request_span /
record_outcome_attrs wrap the OpenTelemetry chat_request span.
Add-only; these are not yet wired into stream_new_chat.py.
Extracts the inner agent-streaming driver previously inlined as
_stream_agent_events in stream_new_chat.py.
stream_agent_events drives graph_stream.event_stream.stream_output and,
after the agent finishes, performs the post-stream safety-net work:
* commit any pending content the agent never explicitly finished
* evaluate file-operation contract outcomes and emit the appropriate
contract verdict for desktop_local_folder turns
This unit is what flows/shared/stream_loop.py wraps in the rate-limit
recovery while-loop. Add-only; no existing wiring uses it yet.
Extracts the agent-construction wrapper that the chat streamers call to
materialize the LangGraph agent for a given thread. Centralizes how we
pass the agent factory plus checkpointer, runtime context, and the
in-memory content builder.
Add-only; pre-existing inline equivalent in stream_new_chat.py stays
until cutover.
Extracts the desktop_local_folder file-operation contract helpers:
* contract_enforcement_active: gates the contract on filesystem mode.
* evaluate_file_contract_outcome: scores tool outputs as success/no-op.
* log_file_contract: structured logging of contract verdicts.
This is the unit responsible for catching agents that claim to have
written/edited a file without actually invoking the filesystem tool.
Add-only; stream_new_chat.py keeps its inline duplicates until cutover.
Extracts two pure context helpers used during input-state assembly:
* mentioned_docs.format_mentioned_surfsense_docs_as_context: renders the
user's @-mentioned SurfSense docs into the LLM context block.
* deepagents_todos.extract_todos_from_deepagents: pulls the in-progress
todo list from a deep-agents state snapshot for the title generator.
Add-only; existing call sites in stream_new_chat.py remain untouched
until cutover.
Foundation layer for the parallel refactor of stream_new_chat.py.
Extracts the StreamResult dataclass (tracks per-turn streaming state)
and a small set of shared utilities (resume_step_prefix, safe_float).
Add-only; no existing code imports from this package yet. Existing
stream_new_chat.py keeps its inline equivalents until cutover.
Resolves: surfsense_backend/app/agents/new_chat/middleware/memory_injection.py
- Took both imports: upstream moved MEMORY_HARD_LIMIT/SOFT_LIMIT to
app.services.memory; kept our perf-logger import for timing.
Pulls in upstream changes:
- Memory document feature (services/memory refactor, removal of
app.agents.new_chat.memory_extraction and background extraction in
stream_new_chat — agent now drives memory via update_memory tool).
- BACKEND_URL env refactor across web tool-ui/editor/chat/dashboard/lib.
- GitHub Actions backend test workflow + pre-commit biome bump.
- Token-display polish in MessageInfoDropdown; save_memory no-update
sentinel.
Verified: 1723 unit tests pass, ruff clean. No semantic regression in
stream_new_chat (their memory-extraction deletion and our preflight
removal touch different functions).
The preflight pattern probed the LLM with a 1-token ping before each
cold turn (when requested_llm_config_id==0, llm_config_id<0, and the
45s healthy TTL had expired) to detect 429s before fanning out into
planner/classifier/title-gen. To absorb its ~1-5s RTT cost we built the
agent speculatively in parallel; on 429 we discarded the build and
repinned.
Three problems with that design:
1. False security. Provider rate limits are token-bucket. A 1-token
ping consumes ~5 tokens; the real request consumes 10-50K. The
probe can return 200 while the real call still 429s.
2. Pure overhead in the common case. On warm-agent-cache turns the
probe dominates wall time: ~2.5s of TTFT pure tax for ~99% of users
who never see a 429.
3. The in-stream recovery loop (catch of _is_provider_rate_limited
gated by not _first_event_logged) already does the right thing
reactively: mark_runtime_cooldown -> resolve_or_get_pinned_llm_config_id
with exclude_config_ids={previous} -> rebuild agent -> retry the
stream. Preflight was never the only safety net; it was a redundant
probe in front of one.
Changes:
- Delete _preflight_llm, _settle_speculative_agent_build, and the
_PREFLIGHT_TIMEOUT_SEC / _PREFLIGHT_MAX_TOKENS constants.
- Drop the parallel agent_build_task / preflight_task plumbing in
both stream_new_chat and stream_resume_chat; build the agent inline
with await _build_main_agent_for_thread(...).
- Drop the unused is_recently_healthy / mark_healthy imports here
(still exported from auto_model_pin_service since OpenRouter
catalogue refresh and a few tests reference clear_healthy).
- Remove the obsolete preflight + settle-speculative tests from
test_stream_new_chat_contract.py.
Net: -447 LOC. ~2.5s removed from TTFT on every cold preflight-eligible
turn. 429 recovery path is unchanged - same repin/rebuild/retry, just
not paid in advance on the healthy path.
Connector kb_sync_services (gmail, onedrive, google_calendar, jira),
streaming indexers (discord, luma, teams) and the file-processor save
path all called embed_text inside async coroutines, blocking the
background worker's event loop for the duration of the embed. Wrap each
call site in asyncio.to_thread so concurrent indexing tasks stop
serialising on the embed.
