RAG Copilot

The Chainlit copilot for the Chainlit documentation.

How to use?

The Chainlit Application

• cp app/.env.example app/.env
• Obtain a Literal API key here
• Get an OpenAI API key here
• Find your Pinecone API key here
• pip install -r app/requirements.txt
• chainlit run app/app.py

The Chainlit Copilot

• Make sure the Chainlit application is running.
• python -m http.server 3004 --directory copilot

Embed the documentation

To upload the latest embeddings to Pinecone:

• cp embed-documentation/.env.example embed-documentation/.env
• Get an OpenAI API key here
• Find your Pinecone API key here
• pip install -r embed-documentation/requirements.txt
• ./embed-documentation/main.sh

How to contribute?

Make all the changes you want to the application, then validate them in local against Test dataset to ship RAG.

Follow the Experiments.ipynb notebook to run your experiments against that dataset.

To have a locally exposed endpoint you can test with, run the main.py Fast API server from the root directory:

uvicorn --app-dir app main:app --host 0.0.0.0 --port 80

This will expose the http://localhost:80/app/ endpoint where you can put your question at the end.