Commit History
-
AGGRESSIVE CACHE BUSTER v3: Force Railway/HF complete rebuild with latest BasicRAG fairness fix
- Added explicit cache-busting RUN statements to invalidate Docker layer cache - Includes timestamp and all commit hashes for deployment traceability - This will force Railway to rebuild everything from scratch, not use cache - Deploys: BasicRAG fairness fix (799cd4dc), graph traversal fix (4215b931), deployment markers (8d2485c4, 9d7f1961) - Expected result on Railway: GraphRAG ~200 tokens with 3-bullet format, fair BasicRAG scoring (~700-800 tokens)05af07f -
FORCE DOCKER REBUILD: Clear cache for latest graph traversal and fuzzy entity matching fix
9d7f196
-
Force HF Space rebuild: f4bb14ae GraphRAG prompt fix
9da0580
-
Fix Docker build with multi-stage frontend compilation and clean up fallback files
97167a5 -
Use separate build script for better frontend compilation diagnostics
2b70cd6 -
Improve frontend build: add Node memory config, better diagnostics
d458e72 -
Fix frontend build: add explicit output dir config, improve error handling
d582adb -
Add Hugging Face Spaces deployment configuration
- Create Dockerfile for containerized deployment - Add .dockerignore to optimize build - Create comprehensive HUGGINGFACE_DEPLOY.md guide HuggingFace Spaces advantages: - Free tier: 2GB+ memory (better than Render 512MB) - Docker support: No complex build issues - Perfect for ML/AI demos - Auto-redeploy on git push - Better user experience for judges Setup: ~10 minutes, completely free, more reliable than Render145846b