Skip to content

Latest commit

 

History

History
12 lines (7 loc) · 581 Bytes

README.md

File metadata and controls

12 lines (7 loc) · 581 Bytes

RAG vs GraphRAG

This notebook proposes the benefits of using a graph database for GraphRAG instead of the limited in-memory RAG approach.

To get started, load the contents of this repo into an ArangoGraph Jupyter notebook and be sure to update with your own OpenAI key.

Note: the gdelt dump is zipped within resources and needs unzipped for the notebook to work. The unzip.py script handles this for you.

Upload to ArangoGraph

  1. Upload the Archive.zip and unzip.py files in the same directory
  2. Open a terminal and run python unzip.py within the above directory