/recursive-retrieval

Primary LanguageJupyter Notebook

Legal Document RAG: Multi-Graph Multi-Agent Recursive Retrieval through Clauses

Setup

  • Instructions are in the notebook legal_recursive_retrieval.ipynb
  • Create a venv if desired
    • python3 -m venv venv
    • source venv/bin/activate
  • Run pip install -r requirements.txt to install the necessary packages.

Important

  • If you are providing a workspace, it should be empty and should only contain documents used build the lexical graph.

Integration with Reducto

  • As of the notebook's creation date, any developers interested in the Reducto API must reach out to founders to get a key for the demo.
  • However, the free demo on the website allows up to 10 pages of text to be analyzed
  • The API response JSONs can be copied and saved as <pdf filename without extension>.json under docs_structure/reducto/.

  • In our example, bnm_compliance.pdf in /docs was saved as bnm_compliance.json in /docs_structure/reducto/.
  • This already has been done for you so you can run the notebook as-is – follow the steps above if you want to try with your own documents.