Exoplanet RAG 🔍

A Retrieval Augmented Generation (RAG) system for querying exoplanet research papers from ArXiv. This framework demonstrates how RAG can be applied to scientific literature, enabling natural language queries about exoplanet research.

View on GitHub

Features

• Fetch and process ArXiv papers about exoplanets automatically
• Create vector embeddings for efficient similarity search
• Answer questions using retrieved context and local LLM
• Web interface built with Streamlit
• Local deployment with Ollama for privacy
• Configurable models and parameters

Technology Stack

• Framework: LangChain for RAG pipeline
• LLM: Ollama (Gemma3:1b) for local inference
• Embeddings: Sentence Transformers (all-MiniLM-L6-v2)
• Vector Store: ChromaDB
• Frontend: Streamlit web interface
• Data Source: ArXiv API for research papers

Example Queries

• “What are sub Neptunes?”
• “How are exoplanets detected?”
• “What is the habitable zone?”
• “What is an atmospheric retrieval?”

Exoplanet RAG System Interface