Graph RAG

From Local to Global: A Graph RAG Approach to Query-Focused Summarization

Approach uses an LLM to build a graph-based text index in two stages: first to derive an entity knowledge graph from the source documents, then to pregenerate community summaries for all groups of closely-related entities. Given a question, each community summary is used to generate a partial response, before all partial responses are again summarized in a final response to the user. Blog: https://www.microsoft.com/en-us/research/publication/from-local-to-global-a-graph-rag-approach-to-query-focused-summarization/

Paper link: https://arxiv.org/abs/2404.16130

Git repo: https://github.com/microsoft/graphrag

For quick start:

We will create and use use graphrag environment.

conda create -n graphrag python=3.11
# To activate this environment, use
conda activate graphrag

To install the required packages, run the following command:

pip install graphrag rich

Add openai key value in variable GRAPHRAG_API_KEY in .env file

configure settings.yaml file

change tokens_per_minute in setting.yaml file as per below https://platform.openai.com/docs/guides/rate-limits/usage-tiers

To initialize the project in the current folder:

python -m graphrag.index --init --root .

You should have you input files in ./input folder in current directory # To start graphrag indexing. using below command the data in ./input folder will be indexed

python -m graphrag.index --root .

To try global search using command:

python -m graphrag.query --root . --method global "what are the top methods in this study?"

Also Check:
https://github.com/microsoft/graphrag/blob/main/examples_notebooks/global_search.ipynb

To try local search using command:

python -m graphrag.query --root . --method local "what are the top methods in this study?"

Also Check: https://github.com/microsoft/graphrag/blob/main/examples_notebooks/local_search.ipynb