FloatChat AI 🌊

An intelligent oceanographic data analysis platform powered by AI, combining RAG (Retrieval-Augmented Generation) with natural language querying for ARGO float data.

🚀 Features

• AI-Powered Chat Interface: Ask questions about ocean data in natural language
• Dual Query System:
  • Semantic search using ChromaDB for descriptive queries
  • NL-to-SQL translation for analytical queries
• Interactive Dashboard: Visualize ocean temperature, salinity, and depth data
• Multiple LLM Support: Works with local Ollama and cloud providers (Groq, OpenAI, OpenRouter)
• Real-time Data Processing: Process and analyze ARGO float measurements
• Export Capabilities: Export data in CSV, NetCDF, and ASCII formats

🛠️ Tech Stack

• Backend: FastAPI, Python 3.13+
• Frontend: Streamlit
• Database: PostgreSQL
• Vector Store: ChromaDB
• LLM: Ollama (local) / Groq / OpenAI / OpenRouter (free-tier models)
• Embeddings: nomic-embed-text / sentence-transformers

📋 Prerequisites

• Python 3.13+
• PostgreSQL
• Ollama (for local LLM)
• Git

🔧 Installation

1. Clone the Repository

git clone https://github.com/NematSachdeva/FloatChat-AI_107.git
cd FloatChat-AI_107/floatchat-ai

2. Create Virtual Environment

python3 -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install Dependencies

pip install -r requirements.txt

4. Setup PostgreSQL Database

# Create database
createdb argo

# Or using psql
psql -U postgres
CREATE DATABASE argo;
\q

5. Install and Setup Ollama

# Install Ollama (macOS/Linux)
curl -fsSL https://ollama.com/install.sh | sh

# Pull required models
ollama pull gemma2:2b
ollama pull nomic-embed-text:latest

6. Configure Environment Variables

Copy the example environment file and configure it:

cp .env.example .env

Edit .env with your settings:

# Database Configuration
DB_PASSWORD=
DATABASE_URL=postgresql+psycopg://your_username@localhost:5432/argo

# LLM Configuration
LLM_PROVIDER=ollama
OLLAMA_HOST=http://localhost:11434
LLM_MODEL=gemma2:2b
EMBEDDING_MODEL=nomic-embed-text:latest

# Free online model option (OpenRouter)
# LLM_PROVIDER=openrouter
# LLM_MODEL=qwen/qwen3-8b:free
# OPENROUTER_API_KEY=your_openrouter_key

# ChromaDB Configuration
CHROMA_PATH=./chroma_db
VECTOR_STORE=persistent

# Backend URL
BACKEND_URL=http://127.0.0.1:8000

🚀 Running the Application

Start Ollama Service

ollama serve

Start Backend (FastAPI)

cd floatchat-ai
source venv/bin/activate
python3 -m uvicorn main:app --host 127.0.0.1 --port 8000 --reload

Start Frontend (Streamlit)

In a new terminal:

cd floatchat-ai
source venv/bin/activate
streamlit run streamlit_app.py

## 🌍 Use Global Argo Dataset (Seanoe GDAC)

To ingest real global Argo profile data from DOI `10.17882/42182`:

```bash
python pipeline/ingest_seanoe_argo.py
python pipeline/data_chroma_floats.py

Notes:

• pipeline/ingest_seanoe_argo.py reads the GDAC profile index and ingests a sampled subset (ARGO_MAX_PROFILES) into PostgreSQL.
• Increase ARGO_MAX_PROFILES gradually as your DB/storage budget allows.

🚂 Railway Deployment

This repository now includes railway.json and Procfile for backend deployment.

Recommended setup on Railway:

1. Create a backend service from this repo (uses Procfile/railway.json).
2. Add a PostgreSQL plugin and set DATABASE_URL from Railway.
3. Set environment variables: LLM_PROVIDER, LLM_MODEL, provider API key, VECTOR_STORE=memory.
4. Deploy and verify /health.
5. Create a second Railway service for Streamlit using start command: streamlit run streamlit_app.py --server.address=0.0.0.0 --server.port=$PORT
6. Set frontend BACKEND_URL to your backend Railway URL.

### Access the Application

- **Frontend**: http://localhost:8501
- **Backend API**: http://127.0.0.1:8000
- **API Docs**: http://127.0.0.1:8000/docs

## 📊 Usage

### Chat Interface

1. Open the frontend at http://localhost:8501
2. Type your question in the chat interface
3. Examples:
   - "What is ARGO?"
   - "Show me temperature data by depth"
   - "What are the average salinity measurements?"

### API Endpoints

#### Health Check
```bash
curl http://127.0.0.1:8000/health

Query Endpoint

curl -X POST http://127.0.0.1:8000/query \
  -H "Content-Type: application/json" \
  -d '{"query_text":"What is ARGO?"}'

🧪 Testing

Run the test suite:

pytest tests/

Run specific tests:

pytest tests/test_api_client.py
pytest tests/test_chat_interface.py

📁 Project Structure

floatchat-ai/
├── main.py                 # FastAPI backend
├── streamlit_app.py        # Streamlit frontend
├── config.py               # Configuration management
├── components/             # UI components
│   ├── api_client.py
│   ├── chat_interface.py
│   ├── data_manager.py
│   └── ...
├── tests/                  # Test suite
├── requirements.txt        # Python dependencies
├── .env.example           # Environment template
└── README.md              # This file

🔐 Security

• Never commit .env files with sensitive data
• Use environment variables for all credentials
• The .gitignore file excludes sensitive files automatically

🐛 Troubleshoot

Backend Returns 500 Error

• Ensure Ollama is running: ollama serve
• Check .env file has correct LLM_PROVIDER=ollama
• Verify models are installed: ollama list

ChromaDB Collection Error

• Delete and recreate: rm -rf chroma_db/
• Restart backend to reinitialize

Database Connection Error

• Verify PostgreSQL is running
• Check database exists: psql -l
• Verify credentials in .env

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

📝 License

This project is open source and available under the MIT License.

👥 Authors

• Nemat Sachdeva - GitHub

🙏 Acknowledgments

• ARGO float data program
• Ollama for local LLM support
• ChromaDB for vector storage
• FastAPI and Streamlit communities

📧 Contact

For questions or support, please open an issue on GitHub.

---

Made with ❤️ for oceanographic research