Add React-TypeScript-Vite frontend with FastAPI backend

[Copilot]

Replaces the Streamlit UI with a modern React frontend and exposes existing RAG functionality via a FastAPI backend with SQLite persistence.

Backend (/backend)

• FastAPI app with REST + WebSocket endpoints: /api/health, /api/chat (HTTP & WS streaming), /api/documents (upload + list), /api/models, /api/auth/token
• JWT auth with optional API key support (X-API-Key header), anonymous fallback for local dev
• SQLite via SQLModel + Alembic for chat session/message persistence (ChatSession, ChatMessage)
• Existing LamaCppClient, Chroma, ConversationHandler wired through API routes without modification
• Path traversal protection on document upload (filename sanitization via Path.name)

Frontend (/frontend)

• Vite + React 19 + TypeScript scaffold with TailwindCSS v4
• ChatGPT-style dark UI: sidebar (session list, document manager, model/strategy selector, settings), message area with markdown rendering (react-markdown), streaming cursor
• Zustand for state, React Query + Axios for data fetching, native WebSocket for token streaming
• Components: Sidebar, ChatArea, MessageBubble, ChatInput, ModelSelector, SettingsPanel

Infrastructure

• docker-compose.yml: backend (FastAPI/uvicorn) + frontend (multi-stage Node build → Nginx) with volume mounts for models/, vector_store/, docs/
• Nginx config with SPA routing and /api/ reverse proxy with WebSocket upgrade
• Makefile targets: frontend-{install,build,test}, backend-{install,test}, docker-{up,down}

Tests

• 11 backend pytest tests (health, auth, chat CRUD, WebSocket streaming, models/strategies) using TestClient with in-memory SQLite
• 14 frontend Vitest + React Testing Library tests (ChatInput, MessageBubble, chatStore)
• All existing tests unaffected

Example: Chat via API

# Get a token
curl -X POST http://localhost:8000/api/auth/token -H 'Content-Type: application/json' -d '{"username":"dev"}'

# Send a message
curl -X POST http://localhost:8000/api/chat/ -H 'Content-Type: application/json' \
  -d '{"message":"What is RAG?","use_rag":true,"k":2}'

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.