Chat‑SQL

Chat with your SQL databases using an LLM-powered Streamlit app. Point it at a local SQLite database (included) or connect to your own MySQL instance, then ask questions in natural language.

Features

• Chat UI built with Streamlit
• SQLite demo database (student.db) ready to use
• Optional MySQL connection (host/user/password/db)
• Uses LangChain with Groq Llama 3 for SQL agent reasoning
• Caches DB configuration for faster reloads

Demo database

This repo includes a student.db SQLite database prepopulated with a STUDENT table (NAME, CLASS, SECTION, MARKS) for quick testing. You can also create your own database.

Quickstart

Prerequisites

• Python 3.10+
• A Groq API key

Install

python -m venv .venv
.\.venv\Scripts\activate
pip install -r requirements.txt

Run

streamlit run app.py

Then in the app sidebar:

• Enter your Groq API key
• Choose one:
  • "Use SQLLite 3 Database- Student.db" to use the bundled SQLite DB (read‑only)
  • "Connect to your MySQL Database" and provide host/user/password/db

Ask questions like:

• "Show top 3 students by MARKS"
• "Average MARKS per CLASS"

Configuration notes

• The SQLite database is opened read‑only. To modify data, connect to MySQL or replace student.db and adjust the code accordingly.
• The LLM is configured via ChatGroq with model_name="Llama3-8b-8192". Set your key in the sidebar each run.

Project structure

app.py          # Streamlit app (UI, agent, DB selection)
sqlite.py       # Helper to create a sample SQLite DB (not required at runtime)
student.db      # Sample SQLite DB used by default
requirements.txt
tests/          # Test files
  test_db_connectivity.py  # Database connectivity tests
.github/        # GitHub Actions and issue templates
  workflows/test.yml       # CI workflow (example)
  ISSUE_TEMPLATE/          # Issue templates

Development

Create/refresh the sample SQLite database (optional)

Running sqlite.py will recreate and populate the STUDENT table:

python sqlite.py

Linting & formatting

This project does not enforce a specific linter yet. Contributors may use ruff/black locally if preferred.

Testing

Basic database connectivity tests are available using pytest:

pip install pytest
pytest tests/ -v

The tests verify:

• SQLite database file exists
• Database connection works
• STUDENT table exists with correct structure
• Table contains sample data

See .github/ISSUE_TEMPLATE/add-ci-pipeline.md for adding CI/CD automation.

Contributing

We welcome contributions! Please read CONTRIBUTING.md for guidelines, local setup, and the pull request process.

Security

If you discover a security issue, please follow the instructions in SECURITY.md rather than opening a public issue.

License

This project is licensed under the MIT License. See LICENSE for details.

Acknowledgements

• Streamlit
• LangChain
• Groq Llama 3 models