[FEATURE] Enhance RAG System with Advanced Features

[anhmtk]

📋 Description

Build upon the existing Vector DB MVP to add advanced RAG capabilities that make StillMe's learning more intelligent and contextual.

🎯 Goals

• Multi-modal embeddings (text, images, code)
• Advanced similarity search algorithms
• Context ranking and relevance scoring
• Query expansion and refinement
• Knowledge graph integration

🛠️ Technical Requirements

• Extend current ChromaDB implementation
• Add support for different embedding models
• Implement advanced search algorithms
• Add context ranking mechanisms
• Create knowledge graph connections

📁 Files to Modify

• backend/vector_db/ - All RAG components
• backend/learning/ - Learning modules
• backend/api/main.py - API endpoints

🧠 Advanced Features

• Multi-modal embeddings (text + images)
• Query expansion using synonyms/related terms
• Context ranking by relevance and recency
• Knowledge graph for concept relationships
• Advanced filtering and search options
• Performance optimization for large datasets

✅ Acceptance Criteria

• Multi-modal search works correctly
• Query expansion improves search results
• Context ranking is accurate and useful
• Knowledge graph shows concept relationships
• Performance is optimized for large datasets