Querying a database using natural language

This sample demonstrates querying a MySQL database using natural language thanks to LangChain, Chainlit, and Ollama.

Runs locally. No internet.

This project includes the classicmodels sample MySQL database.

Requirements

Docker Desktop
Visual Studio Code + Dev Containers extension
A dedicated NVIDIA GPU

Running locally

Open the repository in Visual Studio Code.
Use the Dev Containers command Reopen in a Container.
- Tip: Use Ctrl + Shift + P to open the command pallete.
Install a LLM in the llm container using Ollama.
- Command: ollama pull mistral:7b-instruct-v0.2-q5_K_M
Set the LLM_MODEL in the .env file (If you choose a different model).
When the Dev Container has loaded, launch the app using F5.