Ready Tensor Publications Conversational Assistant

This project provides a Streamlit chatbot that helps users explore Ready Tensor publications by asking natural language questions. The assistant uses a sample of 20-50 publications from project_1_publications.json and answers questions using Google Gemini.

How to Run

1. Install dependencies:

   pip install -r requirements.txt

2. Set your Gemini API key:

   • Get your API key from https://aistudio.google.com/app/apikey
   • Set it as an environment variable:

     export GOOGLE_API_KEY=your-key-here

   • Or add it to Streamlit secrets (.streamlit/secrets.toml):

     GOOGLE_API_KEY = "your-key-here"

3. Run the app:

   streamlit run app.py

Usage

• Ask questions like:
  • What's this publication about?
  • What models or tools were used?
  • Any limitations or assumptions?

The assistant will answer using only the information from the sample publications.