Vanna + Apache Druid (Local Setup)

This repo demonstrates how to connect the open source Vanna library to a locally running Apache Druid instance using SQLAlchemy and PyDruid.

🧠 What is Vanna?

Vanna AI is a lightweight, open-source Python library that translates natural language questions into SQL queries for your database.

🚀 What’s in this demo?

• Runs Apache Druid locally (via its quickstart)
• Connects Vanna to Druid using the pydruid SQLAlchemy dialect
• Lets you ask natural language questions and get back query results

📄 Files

• vanna_druid.py: The full working example that:
  • Connects to a local Druid instance
  • Trains Vanna with a couple of example queries
  • Runs a natural language question end-to-end

⚙️ Prerequisites

• Python 3.8+

• Apache Druid running locally (use start-micro-quickstart)

• Install required packages:

  pip install vanna openai sqlalchemy pydruid pandas chromadb

🧪 Run It

1. Start Druid:

   export DRUID_SKIP_PORT_CHECK=1
   export DRUID_SKIP_JAVA_CHECK=1
   bin/start-micro-quickstart

2. Run the Vanna example:

   python vanna_druid.py

3. Ask natural language questions like:

   What are the top 5 most active Wikipedia channels?
   

---

🐳 Want to run it in the same network as Druid?

We can provide a Docker container that runs Vanna in the same network as your Druid deployment. Just reach out via GitHub Issues or vanna.ai.

---

🧠 Learn More

• Vanna GitHub: https://github.com/vanna-ai/vanna
• Apache Druid: https://druid.apache.org/