deepraaga-api

Flask-based API serving endpoints for DeepRaaga music generation. Extracted from the original DeepRaaga project.

Installation

You can install the API layer via pip. To ensure that the underlying neural network models are available for generation, you should install it with the models optional dependencies:

pip install deepraaga-api[models]

Overview

The deepraaga-api package provides a RESTful interface over the DeepRaaga machine learning backend. It allows external applications (like React frontends) to easily query and generate sequential notes based on learned Carnatic representations.

Usage

You can start the API directly using the bundled CLI tool:

deepraaga-api --port 8000

This will spin up a local Flask server running on http://localhost:8000.

REST Endpoints

GET /

Returns the status of the DeepRaaga API.

POST /api/generate

Generates a sequence of notes.

• Request Body:

  {
      "raga": "mayamalavagowla",
      "duration": 30,
      "temperature": 1.0
  }
  

• Response:

  {
      "notes": ["C4", "D4", "E4", "F4", "G4"],
      "raga": "mayamalavagowla"
  }
  

Programmatic Usage

You can also import the Flask app directly to mount it in another WSGI server (like Gunicorn):

from deepraaga_api.serve import app

if __name__ == '__main__':
    app.run(port=8080)

License

This project is licensed under the MIT License.