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 the
Archive.zipandunzip.pyfiles in the same directory - Open a terminal and run
python unzip.pywithin the above directory