Commit graph

6056 commits

Author SHA1 Message Date
CREDO23
21bddc73a7 refactor(chat): add streaming/flows/shared/assistant_finalize.py
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.
2026-05-25 21:49:31 +02:00
CREDO23
b54b803dc9 refactor(chat): add streaming/flows/shared/ rate-limit recovery + stream loop
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.
2026-05-25 21:49:27 +02:00
CREDO23
2c3edb7c84 refactor(chat): add streaming/flows/shared/terminal_error.py
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.
2026-05-25 21:49:18 +02:00
CREDO23
40300d300a refactor(chat): add streaming/flows/shared/premium_quota.py
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.
2026-05-25 21:49:14 +02:00
CREDO23
e9a98ecafb refactor(chat): add streaming/flows/shared/ base helpers
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.
2026-05-25 21:49:09 +02:00
CREDO23
26c569467d refactor(chat): add streaming/agent/event_loop.stream_agent_events
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.
2026-05-25 21:48:26 +02:00
CREDO23
94bc827252 refactor(chat): add streaming/agent/ package with build_main_agent_for_thread
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.
2026-05-25 21:48:20 +02:00
CREDO23
88a58f6aff refactor(chat): add streaming/contract/ for file-write contract enforcement
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.
2026-05-25 21:48:14 +02:00
CREDO23
c13beae1ce refactor(chat): add streaming/context/ for mentioned-docs and deep-agents todos
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.
2026-05-25 21:48:08 +02:00
CREDO23
4910263c93 refactor(chat): add streaming/shared/ package for StreamResult and utils
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.
2026-05-25 21:48:04 +02:00
Rohan Verma
18c66409a0
Merge pull request #1433 from suryo12/fix/roles-manager-orphaned-hasPermission
fix(web): remove orphaned hasPermission inline body in roles-manager
2026-05-24 18:50:48 -07:00
Rohan Verma
ba18d932e6
Merge pull request #1432 from suryo12/refactor/1358-jotai-slideout-tick
refactor(web): replace slideout panel window event with jotai atom (f…
2026-05-24 18:50:17 -07:00
suryo12
d571cb23fa refactor(web): use last-seen-tick comparison for slideout listener
Replace the boolean "skip first render" ref with a ref that stores the
previously-seen tick value. The effect now compares against the stored
value and only fires when it differs, which makes the dependency
naturally used (removes the `void slideoutOpenedTick;` acknowledgement)
and self-documents the intent of the guard.

Behavior is unchanged — both forms preserve the one-shot-per-event
semantics of the prior window-event implementation. The JSDoc on
`slideoutOpenedTickAtom` is updated to describe the new pattern.
2026-05-24 18:13:36 +07:00
suryo12
29d5ee5465 fix(web): remove orphaned hasPermission inline body in roles-manager
PR #1428 (issue #1366) extracted the inline `hasPermission` callback into
a shared `canPerform` helper but left the original arrow-function body,
its dependency array, and trailing `)` behind after the new
`useCallback` block. The result was a syntactically invalid statement
that broke `pnpm build` on the `dev` branch and is now blocking every
E2E job in the PR queue.

Delete the orphaned lines so the file parses again. No behavior change —
the working `useCallback(canPerform(access, permission))` already
supplies the same predicate the duplicated body did.
2026-05-24 17:23:27 +07:00
Rohan Verma
63e5943cc8
Merge pull request #1431 from dekalouis/refactor/1361-prompt-config-use-mutation-atom
refactor: migrate PromptConfigManager save to updateSearchSpaceMutationAtom
2026-05-24 02:48:46 -07:00
suryo12
ddae506631 refactor(web): replace slideout panel window event with jotai atom (fixes #1358)
Replace the `SLIDEOUT_PANEL_OPENED_EVENT` window event with a
`slideoutOpenedTickAtom` jotai atom. The dispatcher in
`SidebarSlideOutPanel` now bumps the tick via `useSetAtom`, and the
listener in `Thread` reads it via `useAtomValue` and reacts on change
behind a ref guard that skips the initial render — preserving the
one-shot-per-open semantics of the previous event.

This removes the implicit cross-module string contract, makes the
signal traceable through React DevTools / jotai inspector, and lets
TypeScript catch typos that the string-based event API silently
swallowed.
2026-05-24 16:41:47 +07:00
dekalouis
f382cb296c refactor: migrate PromptConfigManager save to updateSearchSpaceMutationAtom 2026-05-24 15:26:40 +07:00
Rohan Verma
d53866d87d
Merge pull request #1430 from suryo12/refactor/1362-oauth-typed-contract
refactor(web): centralize OAuth callback cookie contract (fixes #1362)
2026-05-23 15:53:51 -07:00
Rohan Verma
69388fc710
Merge pull request #1429 from CREDO23/fix-desktop-redirects
[Fixes] Packaged desktop: connector redirect + linux launcher icon
2026-05-23 15:51:59 -07:00
Rohan Verma
ee87747b37
Merge pull request #1428 from guangyang1206/fix/extract-shared-haspermission-helper-1366
refactor: extract shared hasPermission helper (MODSetter/SurfSense#1366)
2026-05-23 15:51:34 -07:00
Rohan Verma
06bcc85287
Merge pull request #1426 from guangyang1206/fix/1375-move-getConnectorTelemetryMeta-to-lib
fix: move getConnectorTelemetryMeta from components/ to lib
2026-05-23 15:50:59 -07:00
Rohan Verma
da4ba09d88
Merge pull request #1427 from AnishSarkar22/feat/opentelemetry
feat: OpenTelemetry integration
2026-05-23 15:50:20 -07:00
suryo12
aa86534a52 refactor(web): centralize OAuth callback cookie contract (fixes #1362)
Replace the duplicated `OAUTH_RESULT_COOKIE` constant and inline payload
type across the callback route and connector dialog hook with a shared
`contracts/types/oauth.types.ts` module that exports:

- OAUTH_RESULT_COOKIE constant
- oauthCallbackResultSchema Zod schema
- OAuthCallbackResult type (inferred from the schema)
- parseOAuthCallbackResult() helper that returns null on invalid JSON
  or shape mismatch

The route handler now uses the shared type to constrain the cookie
payload at compile time. The consumer hook validates the cookie value
through the helper instead of an unchecked JSON.parse, removing the
silent runtime risk when the cookie is tampered with or its shape
drifts.
2026-05-23 23:22:18 +07:00
Anish Sarkar
98e3950dc8 Merge remote-tracking branch 'upstream/dev' into feat/opentelemetry 2026-05-23 03:21:08 +05:30
Anish Sarkar
6302939a72 feat(docs): add observability documentation 2026-05-23 03:18:25 +05:30
Anish Sarkar
4c8d47617d feat(env): add SURFSENSE_ENV variable for deployment environment and update observability resource attributes 2026-05-23 02:13:24 +05:30
Anish Sarkar
df698e0216 feat(observability): integrate OpenTelemetry collector and configuration for enhanced telemetry 2026-05-23 00:17:23 +05:30
CREDO23
fa75d7c54c fix(desktop): ship multi-size icons so linux launchers can find them
electron-builder was given a single 2048x2048 icon.png and dumped it into
hicolor/2048x2048/apps/, a bucket no Linux desktop environment indexes —
launchers and taskbar fell back to a generic placeholder. Pre-render the
standard sizes (16, 32, 48, 64, 128, 256, 512, 1024) and point linux.icon
at the directory so each PNG lands in the matching hicolor/NxN/apps/ slot.

Tray icon was unaffected (loaded at runtime via Electron's Tray API from
bundled resources, no theme lookup).
2026-05-22 20:04:21 +02:00
CREDO23
d5284b3076 fix(desktop): bind bundled Next.js on localhost to keep window origin stable
Setting HOSTNAME=0.0.0.0 made Next.js standalone canonicalize request.url to
http://0.0.0.0:PORT. The connector OAuth callback's NextResponse.redirect built
its Location from that URL, so navigating it flipped window.location.origin from
http://localhost:PORT to http://0.0.0.0:PORT. The backend CORS allowlist matches
localhost/127.0.0.1 only, blocking every subsequent API call until app restart —
producing the "no internet" / app-down state after connecting any connector.
2026-05-22 19:37:19 +02:00
CREDO23
b9403b1720 feat(desktop): emit PostHog events for OAuth redirect intercept and miss
Adds two diagnostic events to surface OAuth-redirect failures we can't
reproduce on Linux:

- desktop_oauth_redirect_intercepted fires from inside onBeforeRequest
  with the original host, path, and target port — confirms the rewrite
  actually ran.
- desktop_oauth_redirect_missed fires from a read-only onCompleted
  listener when a /dashboard/*/connectors/callback URL lands off-localhost,
  meaning the rewrite filter didn't catch it. This is the smoking-gun
  event for "connector OAuth dies on mac/win" reports.

Read-only; no behavior change.
2026-05-22 18:40:06 +02:00
CREDO23
1b6c238c68 refactor(desktop): harden OAuth redirect rewrite for host variants and self-hosters
The interceptor previously matched a strict `${HOSTED_FRONTEND_URL}/*`
prefix and did a naive String.replace, which broke whenever the backend
NEXT_FRONTEND_URL differed at all (apex vs www, http vs https, or a
self-hosted domain). Now:

- Match by host: apex + www. sibling, both http and https.
- Rewrite via URL parsing so only protocol/host change; query strings
  containing the host as a value are left intact.
- Read HOSTED_FRONTEND_URL through getHostedFrontendUrl() which honors
  a SURFSENSE_HOSTED_FRONTEND_URL_OVERRIDE env var, letting self-hosters
  point their builds at their own frontend without rebuilding.

Default behavior is identical when override is unset and backend host
matches the baked-in value.
2026-05-22 18:39:47 +02:00
CREDO23
fe98c17b1d fix(desktop): declare surfsense:// scheme in macOS Info.plist
Linux registered the scheme via desktop-file MIME, but mac.extendInfo
never declared CFBundleURLTypes, leaving install-time LaunchServices
unaware of the protocol. The runtime app.setAsDefaultProtocolClient
call still runs as a fallback.
2026-05-22 18:39:27 +02:00
CREDO23
6ee7c04d02 fix(desktop): recover deep-link URL from argv on win/linux cold start
setupDeepLinks() only listened for second-instance and open-url events.
On Windows/Linux a fresh launch via `surfsense://` delivers the URL in
argv of the first instance, where it was silently dropped. Scan argv on
setup so the existing handlePendingDeepLink() pass picks it up.
2026-05-22 18:39:22 +02:00
CREDO23
d97b2830c5 fix: resolve desktop KB prompt self-contradiction on chunk_ids
The citations fix (cacb27e0) added a "Chunk citations in your prose"
section to system_prompt_desktop.md telling the KB subagent to always
leave `evidence.chunk_ids` null and emit no `[citation:...]` markers in
desktop mode, but left the pre-existing line declaring that
`chunk_ids` apply to `<priority_documents>` hits. The two rules
contradicted each other; the model picked one per turn.

Strike the stale conditional clause and point at the dedicated section
as the single source of truth. Matches the parallel line in
system_prompt_cloud.md and the already-consistent
system_prompt_readonly_desktop.md.
2026-05-22 17:24:57 +02:00
Anish Sarkar
51e4d8b489 feat(tasks): enhance Celery task telemetry with queue metadata and latency tracking 2026-05-22 18:19:38 +05:30
Anish Sarkar
7a3b278b75 feat(connectors): add retry and auth telemetry events 2026-05-22 17:50:02 +05:30
Anish Sarkar
c4abbd6e20 feat(pipeline): enrich ETL and indexing failure telemetry 2026-05-22 17:49:46 +05:30
Anish Sarkar
6e03ab044a feat(tasks): measure Celery queue latency 2026-05-22 17:49:02 +05:30
Anish Sarkar
dc893281ba feat(chat): add model retry and stream lifecycle events 2026-05-22 17:48:43 +05:30
Anish Sarkar
dbb652d4f8 feat(observability): add telemetry error and event helpers 2026-05-22 17:48:01 +05:30
Anish Sarkar
0fe2bba5a8 refactor(chat): replace TooltipIconButton with Button for scroll functionality 2026-05-22 17:46:35 +05:30
Anish Sarkar
87a4dcfd05 feat(tasks): record indexing heartbeat metrics 2026-05-22 13:50:32 +05:30
Anish Sarkar
7c07c220fc feat(connectors): add connector sync spans 2026-05-22 13:49:59 +05:30
Anish Sarkar
4e3a6dff46 feat(etl): instrument extraction spans and outcomes 2026-05-22 13:49:42 +05:30
Anish Sarkar
8bca29fe0d feat(agents): track subagent invocation telemetry 2026-05-22 13:48:57 +05:30
Anish Sarkar
5a6b92c2b6 feat(chat): instrument streamed chat request telemetry 2026-05-22 13:48:19 +05:30
Anish Sarkar
f7f49de109 feat(observability): add chat subagent and ETL telemetry primitives 2026-05-22 13:47:50 +05:30
Anish Sarkar
21d9b1f218 fix(observability): sanitize outbound HTTP span URLs 2026-05-22 13:47:10 +05:30
guangyang1206
a1397a51b5 fix: move getConnectorTelemetryMeta from components/ to lib/ to fix dependency inversion
lib/posthog/events.ts was importing from components/assistant-ui/...,
creating an inverted dependency layer (lib → components).

Move ConnectorTelemetryMeta type, CONNECTOR_TELEMETRY_REGISTRY,
getConnectorTelemetryMeta, and getReauthEndpoint into the new
lib/connector-telemetry.ts module so that lib/ no longer depends upward
into the UI tree.

connector-constants.ts now re-exports from the new module for
backward compatibility.

Fixes #1375
2026-05-22 12:15:20 +08:00
DESKTOP-RTLN3BA\$punk
fc3b39483b Merge commit '334729754f' into dev 2026-05-21 21:14:15 -07:00