GraphRAG Experiments

An experimental project for building and querying knowledge graphs from markdown files using LlamaIndex.

Features

• Build knowledge graphs from markdown files
• Interactive visualization using Streamlit and Pyvis
• Query the knowledge graph using natural language
• Highlight relevant nodes and relationships in the visualization
• Temporal weighting of nodes based on document age

For detailed information about the system architecture and implementation, see architecture.md.

Setup

1. Create a Python environment (3.9 or later)

Install uv if you haven't already

curl -LsSf https://astral.sh/uv/install.sh | sh

Create and activate a new virtual environment

uv venv
source .venv/bin/activate  # On Unix/macOS
# or
# .venv\Scripts\activate  # On Windows

2. Install dependencies using uv:

   uv pip install -e .

3. Create a .env file with your OpenAI API key:

   OPENAI_API_KEY=your_key_here
   

Usage

Web Interface

Run the web interface:

./run_webapp.sh

This will start a Streamlit app where you can:

1. Upload markdown files
2. View the knowledge graph visualization
3. Query the graph using natural language
4. See highlighted nodes and relationships based on your query

CLI Interface

Run the CLI tool:

graphrag process path/to/markdown/files

Development

The project uses:

• LlamaIndex for knowledge graph creation and querying
• Streamlit for the web interface
• Pyvis for graph visualization
• NetworkX for graph operations

License

MIT License