🚀 ISRO Knowledge Graph RAG

ISRO Knowledge Graph RAG is an interactive system that allows users to explore ISRO missions, satellites, launch vehicles, and centers using natural language queries. It uses GraphRAG, Cypher-based retrieval, and semantic search to generate accurate, graph-grounded answers.

Features

• 🧠 Retrieval-Augmented Generation over structured ISRO knowledge
• 🔍 Semantic Search using ChromaDB instead of keyword matching
• 🕸️ Graph-based reasoning using Cypher queries (via Neo4j)
• ⚡ Fast responses by retrieving only relevant triplets / subgraphs
• 💬 Clean Gradio chatbot UI for interacting with the graph
• 🔌 Fully local - no external APIs or cloud databases required

🛠️ Setup Instructions

1. Clone the repository

git clone https://github.com/Anjali-Mittal/KnwoledgeGraph_RAG.git
cd KnwoledgeGraph_RAG

2. Install dependencies

Use a virtual environment (recommended):

python -m venv venv
source venv/bin/activate  # or venv\Scripts\activate on Windows
pip install -r requirements.txt

3. Add environment variables

Create a .env (DO NOT upload this to GitHub):

NEO4J_USER=neo4j
NEO4J_PASSWORD=your_password
OPENROUTER_API_KEY=your_key

For HuggingFace deployment, add these to HF Spaces → Settings → Secrets.

4. Start the local Neo4j database

Make sure Neo4j Desktop or Memgraph Light is running.

5. Run the app

bash python app.py

The Gradio UI will open you can now ask any ISRO-related question.

📂 Directory Structure

ISRO_KnowledgeGraph_RAG/
├── app.py                       # Main Gradio application
├── backend.py                   # Retrieval + LLM logic
├── cypher_generator.py          # Natural language → Cypher conversion
├── fallback_llm.py              # Fallback logic when no context is found
├── get_response.py              # Final answer generation
├── data/
│   └── triplets.csv             # ISRO Knowledge Graph dataset
├── embedding/
│   ├── chroma_utils.py          # ChromaDB operations
│   └── models/                  # (Optional) stored embedding models
├── ingest/
│   └── main_ingest.py           # Ingestion pipeline for embeddings
├── config/
│   └── settings.py              # Central config
├── static/
│   └── graph.js                 # Visualization logic
├── .env                         # Not stored in repo
├── requirements.txt
└── README.md

🛰️ Example Queries

Users can ask things like:

• What missions did ISRO launch in 2019?
• Show me the relationship between PSLV and Chandrayaan.
• Which centers handle satellite integration?
• Explain the structure of ISRO launch vehicles.

🤝 Contributing

Pull requests are welcome — especially improvements to:

• Dataset coverage
• Graph consistency
• Model performance
• UI/UX

🧾 License

MIT License © 2025 Anjali Mittal

---

Made with ❤️ by Anjali Mittal