RAG for Proteins

A demonstration of Retrieval-Augmented Generation (RAG) applied to protein analysis, currently focusing on antibody data from SAbDab (Structural Antibody Database).

🌟 Features

Current:

• RAG-powered chat interface for antibody structure analysis
• Integration with SAbDab for antibody structural data
• Local LLM support via Ollama (currently using Qwen 7B)
• Persistent vector storage for efficient retrieval
• Clean, modular architecture with separate components for:
  • Data loading and preprocessing
  • Vector storage management
  • LLM interface
  • RAG pipeline coordination
  • Streamlit UI

🚀 Quick Start

Prerequisites

• Python 3.10+
• Ollama installed and running
• Required Python packages:

pip install -r requirements.txt