Pre-serialize BM25 index for fast startup (~16 min → ~8s)

Add entity list API, author fetching, and search source decoupling

Update chroma_store, vector_retriever, embed_chunks for rebuilt ChromaDB

Fix ChromaDB schema mismatch: auto-migrate missing 'topic' column on startup

Fix search 500 error: add LLM generation error handling and graceful fallback

Pin numpy<2.0 and chromadb==0.4.24, fix ChromaDB reset safety, update docs

Improve web UI: expand venues, add paper filters, fix dashboard chart, add About page

Ensure DB schema exists on startup for fresh deployments

ResearchRadar: RAG-powered NLP research explorer