Skip to content

~95% on SimpleQA (e.g. Qwen3.6-27B on a 3090). Supports all local and cloud LLMs (llama.cpp, Ollama, Google, ...). 10+ search engines - arXiv, PubMed, your private documents. Everything Local & Encrypted.

7.0 KB
Raw Blame

✨ New Features

  • Add an inline, name-based Zotero collection picker with an explicit entire-library option. (#5254)
  • Add streaming NDJSON research log export endpoint (/api/research/<id>/logs/export) and log download action. (#5259)

🐛 Bug Fixes

  • Fixed parallel RAG indexing durability and vector index reconciliation. Serialized durable writes to prevent parallel-worker SQLite/FAISS corruption, hardened per-collection FAISS write locks with TOCTOU safety guards, and added refuse-to-shrink guards to index reconciliation to protect existing vector stores during transient failures. (#5235)
  • The "Load older" logs re-fetch no longer stamps a false (2×) duplicate badge on rows that were already rendered; only a genuinely new duplicate occurrence bumps the counter now. (#5253)
  • Prevent unused OpenAI-compatible endpoint settings from blocking research with other model providers. (#5255)
  • Surface benchmark result persistence failures instead of presenting them as an empty successful run. (#5256)
  • langgraph-agent fetch_content tool: pre-resolve /library/document/<uuid>[/pdf] URLs and bare [N] citation markers before the egress gate so library-only research runs can actually read their sources. The library RAG and collection search engines emit /library/document/<uuid> as a result's citation URL, and the LLM sometimes pastes the [N] citation marker back into the tool; the egress policy correctly rejected both as unsupported_scheme, so every "Reading the page" milestone on a library-only run produced no content. The pre-resolution passes library URLs through a local DB read of Document.text_content and rewrites citation markers via SearchResultsCollector.find_by_index. (#5266)
  • Fixed database lock contention in SQLite queue processor v2 by batching updates into single transactions and deduplicating per-research updates. (#5268)
  • Reduce false-positive WARNING/ERROR log noise during normal research runs: filter werkzeug Socket.IO disconnect write() before start_response errors, demote DownloadService init path logs to debug, skip the deprecated display-label fallback warning for plain engine keys like collection_<uuid>, and include exception context in document-URL warnings. (#5272)
  • langgraph-agent fetch_content no longer mis-renders egress denials as 📄 From the page: Cannot fetch … in the chat panel — denial and error strings from the tool are now logged as warnings only, and the MILESTONE channel is reserved for successful reads. The policy denied URL fetch WARNING emitted by security/egress/policy.py now inlines the offending URL (redacted to scheme://host:port) into its message body — alongside reason, scope, count, and counted — so a triager reading the persisted research_logs_<uuid>.jsonl can correlate each denial with the URL the agent attempted. With the chat-panel MILESTONE suppression, this WARNING is now the only persistent record of a denial — if the URL were missing here, there would be nothing else to read it from. The message body is what database_sink writes to app_logs.message, so this is the only layer at which the URL actually reaches the JSONL. (#5278)
  • Detailed-mode runs no longer log a misleading format_document_split appears to have over-stripped warning when the answer body is legitimately smaller than the appended sources list (e.g. runs with hundreds of sources); IntegratedReportGenerator now wraps the appended ## Sources block in a sentinel that the splitter keys off, and report_content is correctly persisted as the answer body only. The sentinel is matched at its last occurrence so a spoofed earlier occurrence inside LLM-generated prose cannot silently truncate the answer or disable the over-strip safety check, and the quick-mode save path no longer trusts the sentinel (only the detailed-mode report generator emits it). (#5283)
  • Replaced manual database is locked retry loop with tenacity in _update_task_status. The five-attempt for attempt in range(5) with time.sleep(0.05 * (2**attempt)) was rewritten using tenacity.retry(stop=stop_after_attempt(5), wait=wait_exponential(multiplier=0.05), retry=retry_if_exception(_is_database_locked_error)), matching the rest of the codebase's retry idiom. The predicate limits retries to SQLite database is locked errors; any other exception still escapes on the first attempt so the outer caller can log it without further retry. Total wall-clock backoff is unchanged (~0.05+0.10+0.20+0.40 s before the final attempt). Pins: three regression tests in tests/research_library/routes/test_rag_routes_background_coverage.py verify retry-then-success, fail-fast on non-lock errors, and exhaustion logging.
    • Stop safe_rollback from masking the original failure with a secondary InvalidRequestError or InterfaceError when the thread-local SQLAlchemy session is in a broken state — either mid-provisioning on the per-user QueuePool (concurrent-operation race) or holding a cursor invalidated by a commit/rollback that fired between execute() and the lazy-attribute fetch. Both states now reset the thread-local cache (so the next caller on the same thread gets a fresh session from the QueuePool) and log at DEBUG; other SQLAlchemyError failures still hit the loud logger.exception path (unless their message matches a known broken-session signature).
    • Persist a short [Type: value] exception prefix to app_logs rows from database_sink when the loguru record carries an active exception, so an investigator querying ResearchLog can distinguish ERROR lines that share a bare message. The full traceback remains off the encrypted-DB sink per the #4182 password-redaction policy.
    • Standardize fallback and exception handlers in TokenCountingCallback._save_to_db and SearchTracker.record_search to use logger.warning with scrub_error sanitization, ensuring DB/metrics failure logs are captured safely without exposing tracebacks, credentials, or sensitive encryption passwords on stderr.
  • LangGraph's primary search tool now identifies the selected source classification to agents.

📝 Other Changes

  • Zotero connection, picker, and sync actions now report progress and results through the app-wide notification banner. (#4961)