Spaces:
Paused
Paused
Commit History
-
Support multi-format uploads (PDF/images/Excel/Word/PPT/ODF/text) with AI text extraction
- Backend: extend ALLOWED_DATASET_EXTENSIONS + BINARY_EXTRACTION_EXTS - Add backend/document_parser.py: extract text from PDF (pypdf), images (pytesseract OCR), XLSX (openpyxl), DOCX (python-docx), PPTX (python-pptx), legacy/ODF via LibreOffice - Inject extracted text as SYSTEM context so text-only LLM backends can understand files - Frontend: widen file accept + extension checks + error message - Dockerfile: install tesseract-ocr(+eng, vie), libreoffice, and python doc deps21c88c7
-
Remove all custom files to restore 4bac0e2
Verified signature8ad01c8 -
Fix main.py: Route ALL model IDs through OpenRouter regardless of prefix
Verified signaturecb850a0 -
ROOT FIX: Override agent.py with correct OpenRouter models + routing
Verified signatureb53ebf8 -
Override backend/main.py with OpenRouter patch at startup
Verified signatureabcc4f0