E²GraphRAG: Streamlining Graph-based RAG for High Efficiency and Effectiveness

  

E²GraphRAG is a lightweight and modular framework designed to enhance both efficiency and effectiveness in Graph-based Retrieval-Augmented Generation (RAG). It streamlines the pipeline from document parsing to answer generation via structured graph reasoning.

📁 Project Structure

.
├── README.md
├── requirements.txt
├── main.py
├── build_tree.py
├── dataloader.py
├── extract_graph.py
├── GlobalConfig.py
├── process_utils.py
├── prompt_dict.py
├── query.py
└── utils.py

📦 Datasets

We use data from:

• 📚 NovelQA Partly open-source, to obtain the full dataset, please access via a request to the original authors.
• 🔁 InfiniteBench Fully open-source and publicly available.

You can find how to obtain the data in the ./data/README.md.

Note: After obtaining the datasets, specify the data path when initializing the Dataloader class.

🚀 Getting Started

1. Install Dependencies

Ensure your environment is set up by installing the required packages:

pip install -r requirements.txt

2. Run the Pipeline

The entire pipeline—tree construction, graph extraction, and answer generation—is executed via main.py.

Step-by-step:

1. Create a config file

Prepare a YAML configuration file to define key parameters.

👉 Example: ./configs/example_config.yaml

2. Run the pipeline

bash
python main.py --config <path_to_config_file>

📬 Contact & Citation

If you use this code or find it helpful in your research, please consider citing our work. For questions or dataset access (NovelQA), please contact the original authors.

Star History