SQL Buddy. Yor personal SQL assistant.

The AI powered SQL assistant that helps you write SQL queries faster and more efficiently.

This is a learning project for studying Clojure and ClojureScript.

The project's idea: to learn SQL with the help of AI in a user-friendly environment that allows running SQL code on a real database.

Currently, the project implements only the basic concept, but even in this form, it can be useful for learning SQL.

The project is designed with potential for functional expansion and could evolve into a full-fledged SQL trainer.

Run application using docker-compose

To run the application using docker-compose, you need to have Docker. Add the OPENAI_API_KEY to the setup/docker-compose.yaml file run it from the setup folder.

docker-compose up

Specify the OPEN-API key

You should set the OPEN-API key in the system environment variable OPENAI_API_KEY before running the application.

Development mode (figwheel)

To start the Figwheel compiler, navigate to the project folder and run the following command in the terminal:

lein figwheel

Figwheel will automatically push cljs changes to the browser. The server will be available at http://localhost:3449 once Figwheel starts up.

Figwheel also starts nREPL using the value of the :nrepl-port in the :figwheel config found in project.clj. By default the port is set to 7002.

The figwheel server can have unexpected behaviors in some situations such as when using websockets. In this case it's recommended to run a standalone instance of a web server as follows:

lein do clean, run

The application will now be available at http://localhost:3000.

Development mode (shadow-cljs)

Install shadow-cljs:

npm install -g shadow-cljs

Start shadow-cljs: Navigate to your project directory and start the shadow-cljs server:

shadow-cljs watch app

Run the development server: In another terminal, start the development server using Leiningen:

lein run

Optional development tools

Start the browser REPL:

$ lein repl

The Jetty server can be started by running:

(start-server)

and stopped by running:

(stop-server)

Building for release

lein do clean, uberjar

Building the Docker image

docker build -f setup/Dockerfile -t sql-buddy .

Running the Docker image

 docker run -p 3000:3000 -e "OPENAI_API_KEY=your-openai-api-key" sql-buddy  

Application Usage Instructions

• To receive a new task, click on the "New Task" button.
• The "Answer" and "Question" buttons add a prefix to the message to help the system better recognize the type of message.
• If you want to ask the AI something, click "Question". For answering a task, use "Answer".
• To generate data in the database according to the current task, click "Fill Database".
• To execute a query, click "Run Query".

To Do

Functional Tasks:

• Save current messages to the database between application sessions.
• Develop a system for tracking learning progress.
• Add the ability to select a task topic.
• Add the ability to select difficulty levels.
• Implement view like a python notebook.

Technical Tasks:

• Implement syntax highlighting for SQL code.
• Implement logging.
• Improve error handling.
• Increase test coverage.
• Code refactoring for easier testing.
• Enhance the user interface.
• Optimize token usage.
• Enable compatibility with different AI models, including local ones.