OCR/openspec/changes/add-office-document-support/tasks.md

Implementation Tasks

Phase 1: Dependencies & Configuration

• Install Office document processing libraries
  • Install LibreOffice via Homebrew (headless mode for conversion)
  • Verify LibreOffice installation and accessibility
  • Configure LibreOffice path in OfficeConverter
• Update JWT token configuration
  • Change ACCESS_TOKEN_EXPIRE_MINUTES to 1440 in app/core/config.py
  • Verify token expiration in authentication flow

Phase 2: Document Conversion Implementation

• Create Office document converter class
  • Add office_converter.py to services directory
  • Implement Word document conversion methods
    • convert_docx_to_pdf() for DOCX files
    • convert_doc_to_pdf() for DOC files
  • Implement PowerPoint conversion methods
    • convert_pptx_to_pdf() for PPTX files
    • convert_ppt_to_pdf() for PPT files
  • Add error handling and logging
  • Add file validation methods

Phase 3: OCR Service Integration

• Update OCR service to handle Office formats
  • Modify process_image() in ocr_service.py
  • Add Office format detection logic
  • Integrate Office-to-PDF conversion pipeline
  • Update supported formats list in configuration
• Update file manager service
  • Add Office formats to allowed extensions (file_manager.py)
  • Update file validation logic
  • Update config.py allowed extensions

Phase 4: API Updates

• File validation updated (already accepts Office formats via file_manager.py)
• Core API integration complete (Office files processed via existing endpoints)
• API documentation strings (optional enhancement)
• Add Office format examples to OpenAPI schema (optional enhancement)

Phase 5: Testing

• Create test Office documents
  • Sample DOCX with mixed Chinese/English content
  • Test document creation script (create_docx.py)
• Verify document conversion capability
  • LibreOffice headless mode verified
  • OfficeConverter service tested
• Test token validity
  • Verified 24-hour token expiration (1440 minutes)
  • Confirmed in login response
• Core functionality verified
  • Office format detection working
  • Office → PDF → Images → OCR pipeline implemented
  • File validation accepts .doc, .docx, .ppt, .pptx
• Automated integration testing
  • Fixed API endpoint paths in test script
  • Fixed configuration loading (.env file update)
  • Fixed preprocessor bugs (MIME types, validation, return order)
  • End-to-end test completed successfully (batch 24)
  • OCR accuracy: 97.39% confidence on mixed Chinese/English content
• Manual end-to-end testing
  • DOCX → PDF → Images → OCR pipeline verified
  • Processing time: ~375 seconds (includes model initialization)
  • Result output format validated (Markdown generation working)

Phase 6: Documentation

• Update README with Office format support (covered in IMPLEMENTATION.md)
• Test documents available in demo_docs/office_tests/
• API documentation update (endpoints unchanged, format list extended)
• Migration guide (no breaking changes, backward compatible)