graphrag-biomedical

Implement GraphRAG on Bio-Medical Text Corpus

Pre-Requisities :-

1. OpenAI API Key ( Or Groq API Key )
2. LM Studio ( For Embedding Model to run locally )

Execute the following Commands to setup GraphRAG :

python -m pip install graphrag mkdir input

To Import Medical Text in PDF Format to Text :

python -m pip install pymupdf python parse_pdf.py

Initialise GraphRAG in the directory :

python -m graphrag.index --init --root . [ Setup settings.yaml file according to API Specifications & Input Data ]

GraphRAG Indexing :

python -m graphrag.index --root . [ Indexer is cached, as a result Rate Limit API Errors can be handled ]

Run a GraphRAG Query :

python -m graphrag.query --root . --method global "What is the main focus of the paper?" python -m graphrag.query --root . --method local "Explain the methods explored in this paper about breast cancer ?"

Results :

Global Search

Example Query : python -m graphrag.query --root . --method global "Explain the role of this paper in the context of breast cancer diagnosis"