feat: add Screen 4 - Multi-Hop Reasoning page

feat: prioritize Groq model, configure LLM fallbacks, and fix graph entity highlighting by name

Fix: Resolve 0-entity extraction error by adding NVIDIA LLM integration and robust error handling

feat: complete frontend and backend integration and security configurations

feat: implement GraphRAG Phase 1 backend service layer and test suite

initialise the Graphrag knowledge ai