ResearchBuddy

Gradio Interface for LLM-Powered PDF Chats

This chatbot is designed to provide intelligent responses and answers to questions based on the content of PDF documents.Leverages Gradio as a user-friendly interface to engage with chatbots powered by OpenAI models based on langchain. Additionally, it incorporates ChromaDB for efficient data storage.

Current LLM used - GPT4-1106-preview

A base interface demo is available on this HF space for testing

Getting started

Clone this repository and add your OpenAI API key in local environment

git clone https://github.com/kushal-10/chatllms
cd chatllms

Install required dependencies

pip install -r requirements.txt

Usage

Export google api keys to save the chat history logs and access documents from a shared google drive link. (Alternatively, if using local documents, see lc_base/README_BASE.md)

export GOOGLE_CLIENT_EMAIL=".....gserviceaccount.com"
export GOOGLE_PRIVATE_KEY="-----BEGIN PRIVATE KEY-***-END PRIVATE KEY-----\n"
export GOOGLE_TOKEN_URI="..../token"
export LOGS_ID="..."
export OPENAI_API_KEY="sk-..."
export DRIVE_LINK="https://drive.google.com/drive/folders/....?usp=sharing"

To run the gradio interface

python3 app_drive.py