Skip to content

v1.0.5: Data Preview Fix, Universal Model Filtering & Settings Environment Controls

Choose a tag to compare

View all tags
@Aniket-a14 Aniket-a14 released this 28 Aug 10:48
· 162 commits to master since this release
v1.0.5
d49cd6f

Highlights in v1.0.5

🔧 Bug Fixes

  • Dataset Preview 'No rows to show' Fix: Corrected PyArrow schema inference in _arrow_chunks where empty slices (df.iloc[:0]) caused string/object columns to default to pa.null(), silently failing Arrow IPC streaming on upload.
  • CORS Missing Headers: Added X-Arrow-Total-Rows and X-Arrow-Offset to expose_headers for proper frontend pagination.
  • DataGrid Loading State: Initialized loading=true to eliminate empty-state flash before fetch completes.

🧠 Universal Model Filtering

  • Intelligent Cloud Filtering: Added EXCLUDED_MODEL_PATTERNS to exclude embeddings, TTS, audio, image/video generation, robotics, and deep-research endpoints from cloud providers (Gemini, OpenAI, Anthropic). Gemini model list reduced from 54 raw models to 18 usable chat/coding/reasoning models.
  • Local Model Preservation: Preserved user-downloaded local models in Ollama and LM Studio without aggressive filtering.

⚙️ Settings Environment Controls & Persistence

  • Environment & Provider Defaults: Added UI controls for API_PROVIDER, DATA_MODE, DATA_SCHEMA_ONLY, GATEWAY_API_URL, GATEWAY_API_KEY, SANDBOX_EXEC_TIMEOUT, COUNCIL_ENABLED, VISION_ENABLED, CONTEXT_DOCS_ENABLED, and SKILLS_ENABLED.
  • Atomic Persistence: Changes are immediately written to backend/.env and applied in-memory via PATCH /api/config.
Assets 8
Loading