Summary

Currently, there isn't documentation that highlights the ability to use full context of the documents uploaded for RAG. This causes new users like myself to run into subpar RAG results in certain scenarios. Adding the RAG_USE_FULL_CONTEXT to the .env.example file for better visibility.

Relevant PRs

• RAG API Documentation Update: added explanation for RAG_USE_FULL_CONTEXT variable LibreChat-AI/librechat.ai#146
• Updated documentation to highlight existence of RAG_USE_FULL_CONTEXT variable rag_api#90

Change Type

Please delete any irrelevant options.

• This change requires a documentation update

Testing

The functionality of using RAG_USE_FULL_CONTEXT has already existed and been tested. Though, I tested it again and it works. The results of the RAG output visibly includes more context when setting RAG_USE_FULL_CONTEXT=true and not having the variable defined at all in .env. It is more apparent when using smaller models (8b params or less).

Test Configuration:

Checklist

Please delete any irrelevant options.

• My code adheres to this project's style guidelines
• I have performed a self-review of my own code
• I have made pertinent documentation changes
• Local unit tests pass with my changes
• A pull request for updating the documentation has been submitted.