anantasty / faa_chatbot

LLM Rag Demo

This repo demonstrates the creation of a simple Retrieval Assisted Generation pipeline using langchain. You can access the api here The docs used for this can be found here.

The pipeline has 2 parts:

• Vectorstore creation
  • We read the documents, split and tokenize them
  • We create a vectorstore from the documents.
• Retrieval
  • We generate an embedding for the question
  • We then use the vectorstore to retrieve the most similar documents to a query
  • The documents are passsed into the context of the LLM for generating a coherent answer

Setup

• Clone this repo
• Set up env for the vectorstore creation and install requirements.
  • cd vectorstore_creation
  • pip install -r requirements.txt
  • Get the docs for the vectorstore creation - here
  • generate the vectorstore using python rag_index.py
• Set up a new python env for the Retrieval pipeline
  • cd gcloud
  • pip install -r requirements.txt
  • copy the my_deeplake folder from the vectorstore_creation into the gcloud folder
  • python faa_chat_api.py