README

This repository contains code related to RAG (Retrieval-Augmented Generation). Currently, only a simple RAG implementation is available in the simple_rag directory.

Instructions

1. Run download.py: This script downloads the necessary data for the RAG pipeline.
2. Run ingest.py: This script processes the downloaded data and ingests it into the database.
3. Run retrieve.py: This script retrieves data from the database and provides answers to user queries. Note that retrieve.py must be run in interactive mode because it contains asynchronous code that is not encapsulated within an async def function. Instead, await is called directly in the script.

Example

To run retrieve.py in interactive mode, use the following command:

python -i simple_rag/retrieve.py

Additional Information

The documentation files are downloaded from their respective documentation locations (e.g., Pydantic AI) using the download.py script. After downloading, the ingest.py script performs vector embedding updates to prepare the data for retrieval operations.