Spaces:
Sleeping
Sleeping
Commit History
-
Add automatic Docling parsing display in indexing flow
- Add /api/parse-docling endpoint for bulk file parsing - Create DoclingOutput component showing complete parsing results: - Breakdown numbers (elements, chars, words, pages) - Element type counts - Complete element list with full text - Download JSON button - Continue to Indexing button - Add parseWithDocling API function - Add "parse" step to ProcessingStatus - Update Sidebar to: - Parse files with Docling after Dropbox fetch - Display DoclingOutput for user review - Continue indexing on user confirmation - Remove manual "Test Parsing" button (now automatic) 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>aa663e1 -
Add parsing evaluation UI to frontend
- Add evalParsing and getSupportedFormats API functions - Create ParsingEval component to display Docling parsing results - Add "Test Parsing" button to each staged file in Sidebar - Expand supported file formats (DOCX, PPTX, XLSX, HTML, images) - Increase max file size to 10MB 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>2eda359 -
Add /api/eval/parsing endpoint for Docling evaluation
- Fetches file from Dropbox - Runs Docling parsing - Returns element breakdown and metrics 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>2413602 -
Add evaluation framework for Docling + RAG pipeline
- eval_spot_check.py: Manual parsing inspection - eval_parsing.py: Automated structure metrics - eval_retrieval.py: Precision/recall/MRR - eval_embeddings.py: Semantic similarity tests - tests/eval_data/: Test document structure 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>c6a48e0 -
Fix libgl package name for Debian Trixie
🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>3a9dfa1 -
Fix Dockerfile for HF Spaces user permissions
🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>77879fa -
Add Docling integration for multi-format document processing
- Add docling_loader.py for PDF, DOCX, PPTX, HTML, image support - Add structure-aware chunking (preserves tables, sections) - Update ingestion API with use_docling and use_structure options - Update Dockerfile with system deps for Docling - Remove render.yaml (using HuggingFace Spaces only) 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>7e07738
-
Initial commit: RAG Document Assistant with Zero-Storage Privacy
f866820