abhijitpal1247 / RAG-exp

This is an experimental repo for performing RAG experiments with LangChain

This is a repository where I track my experimental journey with the Retrieval-Augmented Generation (RAG) implemented using LangChain and integrated with a chat interface using Streamlit.

Overview

This project is an ongoing experiment exploring the capabilities and applications of combining LangChain's RAG for generating responses with a user-friendly chat interface built with Streamlit. The primary goal is to understand, experiment, and possibly identify some improvements over the existing conventions for implementing RAG.

Installation

To get started with these experiments on your local setup, please follow the steps below:

# Clone the repository
git clone [https://github.com/abhijitpal1247/RAG-exp.git](https://github.com/abhijitpal1247/RAG-exp.git)

# Navigate into the repository
cd RAG-exp

# Install the necessary dependencies
pip install -r requirements.txt

Create a .env files with the following environment variables stored in it
WCS_DEMO_RO_KEY
WCS_DEMO_URL
HUGGINGFACEHUB_API_TOKEN

I have used weaviate vector database for my experiments. It provides a free demo instance which came in very very handy while conducting these experiments.

Running the Streamlit UI

After installing the dependencies, you can run the Streamlit UI to interact with the RAG-powered chat system using the following command:

streamlit run main.py