Orchestrator

A Django-based service for orchestrating computer vision model inference jobs using Celery and Redis. This project is containerized with Docker and provides REST API endpoints for inference job management.

Features

• Django 5.2 backend with REST API (using Django REST Framework)
• Asynchronous job processing with Celery and Redis
• Containerized with Docker and Docker Compose
• Gunicorn for production-ready serving

Requirements

• Docker & Docker Compose
• Python 3.11 (if running locally)

Setup (with Docker)

1. Clone this repository:

   git clone <repo-url>
   cd orchestrator

2. Create a .env file with the following variables:

   VT_API_URL=<your_vt_api_url>
   VT_API_TOKEN=<your_vt_api_token>
   DJANGO_SECRET_KEY=<your_secret_key>
   DJANGO_DEBUG=1

3. Build and start the services:

   docker-compose up --build

   This will start:

   • Django web server (on port 8001)
   • Celery worker
   • Redis

4. The API will be available at http://localhost:69/

API Usage

• POST /enqueue/
  • Enqueue a new vision model job.
  • Example payload:

    {
      "job_id": "<uuid>",
      "vision_model_id": 1,
      "input_image_url": "http://.../file.png"
    }

Project Structure

• core/ – Django app with tasks, views, and business logic
• orchestrator/ – Django project settings and configuration
• Dockerfile, docker-compose.yml – Containerization
• requirements.txt – Python dependencies