Add detailed error logging to debug GraphRAG issues

Fix CSV parsing logic for nano-graphrag context format

fix: Remove demo mode completely and force real GraphRAG usage

fix: Simplify nano-graphrag installation without setup.py

fix: Correct nano-graphrag installation path

fix: Install nano-graphrag at runtime from source

fix: Add nano-graphrag source code to Space

fix: Force NANO_GRAPHRAG_AVAILABLE = True for testing

fix: Fix GraphRAG function arguments and add all book data

feat: Update Gradio app with enhanced GraphRAG functionality

Final: Replace nano-graphrag with direct JSON data + OpenAI API to avoid Exit code 132 crashes

Stable: Remove problematic dependencies and add graceful fallback for GraphRAG without graspologic

Debug: Add startup logging to diagnose book loading issues and force HF Space restart

Fix: Auto-detect available books and add debug logging for book loading

Restore: Full nano-graphrag dependencies (graspologic, scipy, umap-learn) now that HF Space has CPU upgrade with sufficient resources

Fix: Remove problematic dependencies and improve error handling for matrix/graspologic issues

Fix: Add missing dependencies for nano-graphrag matrix operations (scipy, graspologic, umap-learn) and improve error handling

Add diagnostic API endpoint for troubleshooting GraphRAG setup

Fix: Add proper nano-graphrag dependencies and environment setup

Add diagnostic functionality to debug GraphRAG mode

Add Borges Graph app with GraphRAG data