Honest eval narrative + query rewrite off by default

Add query rewriting + corrective RAG + 3-stage RAGAS ablation

Add reranking + RAGAS evaluation

Always enrich retrieval query with previous user message

Add conversation memory: history-aware LLM + context-aware retrieval

Fix startup: run ingest in background thread so Gradio launches immediately

Deploy: auto-ingest, hybrid RAG, streaming, UMAP viz, 16 LLMs