An intelligent and highly visualized RAG (Retrieval-Augmented Generation) system for local knowledge base management.

Brain_Net is a comprehensive knowledge base system that intelligently manages document processing, retrieval accuracy monitoring, and provides real-time visualization through a modern web interface.

• Document Processing Pipeline: Intelligent parsing, chunking, and indexing
• Query Generation Engine: Automated test query creation for accuracy validation
• Retrieval Monitoring: Real-time accuracy tracking and bad case detection
• API Layer: RESTful endpoints for frontend communication
• LangChain Integration: LLM-powered document understanding and processing

• Modern Web Interface: Responsive, intuitive user experience
• Real-time Visualization: Live processing status and analytics
• Interactive Dashboard: Document management and query testing
• Analytics Dashboard: Retrieval accuracy metrics and insights

• PostgreSQL: Primary database for metadata, user data, and structured information
• Elasticsearch: High-performance document indexing and full-text search
• Neo4j: Knowledge graph for document relationships and semantic connections
• MinIO: Scalable object storage for document files and artifacts

• LangChain: LLM orchestration and document processing
• Query Generation: Intelligent test query creation
• Accuracy Evaluation: Automated retrieval quality assessment
• Semantic Search: Advanced document similarity and retrieval

• 📁 Document Upload: Support for multiple file formats and folder paths
• 🔍 Intelligent Chunking: Context-aware document segmentation
• 🎯 Query Generation: Automated test query creation for accuracy validation
• 📊 Real-time Monitoring: Live retrieval accuracy tracking
• 🎨 Rich Visualization: Interactive dashboards and analytics
• 🔄 Continuous Learning: System improvement through feedback loops

• Backend: FastAPI, Python 3.11+, LangChain, Pydantic
• Frontend: Next.js 14, React 18, TypeScript, Tailwind CSS
• Databases: PostgreSQL, Elasticsearch, Neo4j, MinIO
• AI/ML: LangChain, OpenAI/Anthropic APIs, Sentence Transformers
• Infrastructure: Docker, Docker Compose, Nginx
• Monitoring: Prometheus, Grafana, ELK Stack

• Docker and Docker Compose
• Node.js 18+ and npm/yarn
• Python 3.11+

# Clone the repository
git clone <repository-url>
cd brain_net

# Start the entire system
docker-compose -f docker/docker-compose.yml up -d 

# Or run components individually
cd apps/backend && source .venv/bin/activate && uvicorn main:app --reload
cd apps/frontend && npm install && npm run dev

brain_net/
├── apps/
│   ├── backend/           # FastAPI backend application
│   ├── frontend/          # Next.js frontend application
│   ├── shared/            # Shared utilities and types
│   ├── llm/               # LLM service
│   │   ├── app/
│   │   │   ├── __init__.py
│   │   │   ├── main.py
│   │   │   ├── api/
│   │   │   │   ├── v1/
│   │   │   │   │   ├── chat.py
│   │   │   │   │   ├── agents.py
│   │   │   │   │   └── rag.py
│   │   │   │   ├── core/
│   │   │   │   │   ├── config.py
│   │   │   │   │   └── agents/
│   │   │   │   │       ├── crew_ai_agents.py
│   │   │   │   │       └── langraph_agents.py
│   │   │   │   ├── services/
│   │   │   │   │   ├── rag_service.py
│   │   │   │   │   ├── embedding_service.py
│   │   │   │   │   └── vector_service.py
│   │   │   │   └── models/
│   │   │   ├── requirements.txt
│   │   │   └── Dockerfile
│   │   └── README.md
│   ├── docker/                # Docker configurations
│   ├── docs/                  # Documentation
│   └── scripts/               # Utility scripts

Detailed configuration options and environment variables are documented in each component's README.

• Backend API Documentation
• Frontend User Guide
• Deployment Guide
• Architecture Details