Agentic RAG with Knowledge Graph 🤖

A powerful AI agent system that combines traditional RAG (Retrieval-Augmented Generation) with knowledge graph capabilities to analyze and provide insights about big tech companies and their AI initiatives.

🚀 Features

• Hybrid Search: Combines vector similarity search with knowledge graph traversal
• Temporal Knowledge: Tracks how information changes over time
• Multi-LLM Support: Works with OpenAI, Ollama, Anthropic, Google Gemini, and more
• Streaming API: Real-time responses with Server-Sent Events
• Production Ready: Comprehensive testing, error handling, and monitoring

🛠️ Tech Stack

• AI Framework: Pydantic AI
• Knowledge Graph: Neo4j + Graphiti
• Vector Database: PostgreSQL + pgvector
• API: FastAPI
• LLM: Configurable (OpenAI, Ollama, etc.)

📋 Prerequisites

• Docker and Docker Compose
• 8GB+ RAM
• 20GB free disk space

🚀 Quick Start

1. Clone and Navigate

git clone https://github.com/DavidMelamed/agentic-rag-knowledge-graph.git
cd agentic-rag-knowledge-graph

2. Configure Environment

cp .env.example .env
# Edit .env with your configuration

3. Deploy with Docker

# Linux/Mac
./deploy.sh

# Windows PowerShell
.\deploy-windows.ps1

# Windows Command Prompt
deploy-windows.bat

4. Add Documents

# Create documents directory
mkdir -p documents

# Copy your markdown files or use samples
cp -r big_tech_docs/* documents/

5. Run Ingestion

docker exec agentic-rag-app python -m ingestion.ingest

6. Start Using

# CLI Interface
docker exec -it agentic-rag-app python cli.py

# API (http://localhost:8058)
curl http://localhost:8058/health

🔧 Configuration

The system supports multiple LLM providers:

OpenAI

LLM_PROVIDER=openai
LLM_API_KEY=sk-your-key
LLM_CHOICE=gpt-4-mini

Ollama (Local)

LLM_PROVIDER=ollama
LLM_BASE_URL=http://ollama:11434/v1
LLM_CHOICE=qwen2.5:14b-instruct

Anthropic (via OpenRouter)

LLM_PROVIDER=openrouter
LLM_API_KEY=your-openrouter-key
LLM_CHOICE=anthropic/claude-3-5-sonnet

📚 Documentation

• Deployment Guide
• API Documentation
• Original Project README

🏗️ Architecture

┌─────────────┐     ┌─────────────┐     ┌─────────────┐
│   FastAPI   │────▶│  Pydantic   │────▶│     LLM     │
│   Server    │     │   AI Agent  │     │  Provider   │
└─────────────┘     └─────────────┘     └─────────────┘
                            │
                ┌───────────┴───────────┐
                ▼                       ▼
        ┌─────────────┐         ┌─────────────┐
        │ PostgreSQL  │         │    Neo4j    │
        │  pgvector   │         │   Graphiti  │
        └─────────────┘         └─────────────┘

📊 Example Queries

• Semantic: "What AI research is Google working on?"
• Relational: "How are Microsoft and OpenAI connected?"
• Temporal: "Show me the timeline of Meta's AI announcements"
• Complex: "Compare AI strategies of FAANG companies"

🛠️ Development

Local Development

# Install dependencies
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt

# Run locally
python run-local.py

Testing

pytest
pytest --cov=agent --cov=ingestion

🐛 Troubleshooting

Docker Issues

• Ensure Docker Desktop is running
• Linux/WSL: Add user to docker group: sudo usermod -aG docker $USER

Port Conflicts

• Modify ports in docker-compose.yml

Memory Issues

• Increase Docker memory allocation (8GB+ recommended)

📄 License

This project is based on the ottomator-agents repository.

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

🌟 Acknowledgments

• Original project by coleam00
• Built with Pydantic AI, FastAPI, PostgreSQL, and Neo4j