Skip to content

Library to support common tasks in retrieval augmented generation (RAG).

License

Notifications You must be signed in to change notification settings

br-data/rag-tools-library

Repository files navigation

rag-tools-library

Library to support common tasks in retrieval augmented generation (RAG).

This library is in a very early stage and all the documentation is AI generated.

Tutorial and Documentation

You find a brief tutorial and the documentation under br-data.github.io/rag-tools-library.

Roadmap

  • Add Google Bison to available LLMs
  • Add an offline database alternative
    • FAISS and SQLite
  • Allow users to register their own LLMs
  • Allow users to register their own Embedding models
  • Support Semantic Scholar endpoint to generate embeddings for scientific papers.
  • Support chat functionality; e.g. let the user give feedback on the result to the LLM.

Deployment

Run the build_and_deploy.sh script in the root folder. Once prompted for the username, pass __token__ and the pypi API token you've received. If you don't have an API token and feel like you should, feel free to contact the maintainers.

Contact

Marco Lehner

marco.lehner@br.de