GraphRAG Visualization Tutorial

This repository provides you with documents that have already undergone the Indexing pipeline, allowing you to quickly build and query knowledge graphs with Microsoft GraphRAG. Additionally, it includes documentation for visualization to facilitate quick implementation.

English | 中文

Prerequisites

Ensure you have read through the Microsoft GraphRAG documentation in English, or my article in Chinese.

Environment Setup

Enter your OpenAI API or Azure OpenAI API key into .env.sample and rename it to .env.

Visualization Usage

Method 1: Using yFiles Graphs

Install the packages from requirements.txt, ensuring your Python version is between 3.10 and 3.12.

pip install -r requirements.txt

Then open graph-visualization.ipynb and modify the INPUT_DIR to your indexing results directory. If you have downloaded the entire repo, the default settings should work.

Method 2: Using GraphML and Third-Party Software

Enable GraphML output in settings.yaml:

snapshots:
  graphml: True

After completing the Indexing pipeline, open the generated GraphML file using tools such as Gephi or yEd Graph Editor.