Improve AI summary UX and tighten prompt to prevent truncation

Switch to Groq-only LLM, remove GGUF dependency, speed up build and inference

Add complete RAG modules and new UI with synthesis and AI summary

Improve LIST reliability with score threshold and lexical guard; align UI with extractive synthesis and faster QA

Refactor UI and RAG modes: extractive synthesis for EXPLAIN, faster QA with warnings

Add guided tabs UI (LIST/FULLTEXT/EXPLAIN/ADVANCED)

Use models/model.gguf and configure GGUF model via env vars

Fix: download FAISS from dataset repo (repo_type=dataset)

Fix FAISS index download from HF dataset at startup

Download FAISS index at runtime (robust HF setup)

Initial commit - Code education RAG Space