Stock Market ML App

Prerequisites

Before you begin, make sure you have the following installed on your machine:

• Docker: Install Docker
• Docker Compose: Install Docker Compose

Getting Started

1. Clone the repository:

   git clone https://github.com/vkmguy/ml_training_hub.git
   cd ml_training_hub

2. Build and start the services:

   docker-compose up --build

   This command will build the Docker images and start the services.

3. Access your application:

   • Django Admin: localhost:8001/admin (setup your username and password)

   docker exec -it mlTrainingHub bash
   python manage.py createsuperuser

4. To stop the services, press Ctrl + C in the terminal where docker-compose is running.

Services

Django Web App

The Django web app allows you to interact with the Stock Market ML model.

• GET request to train the ML model: localhost:8001/api/trainModel
• GET request to get the last 10 ML model accuracies: localhost:8001/api/accuracies
• POST request to predict the direction of market: localhost:8001/api/predict

  {
   "open": 0.11,
   "low": 0.1,
   "high": 0.11,
   "volume": 202048089,
   "dividends": 0,
   "stock_splits": 0
  }

  {
   "prediction": 0.0
  }

• GET request to train the ML model: localhost:8001/api/trainModel

  {
  "message": "ML model training task has been triggered successfully."
  }

• GET request to fetch the ML model accuracy: localhost:8001/api/accuracies

  [
  {
     "metrics": {
         "recall": 0.8824884792626728,
         "accuracy": 0.4818880351262349,
         "precision": 0.4763681592039801
     },
     "timestamp": "2024-02-07T21:09:59.251336Z"
  },
  {
     "metrics": {
         "recall": 0.7235023041474654,
         "accuracy": 0.49286498353457736,
         "precision": 0.47865853658536583
     },
     "timestamp": "2024-02-07T20:33:16.008314Z"
  }
  ]

  • GET model visualization for its metrics: localhost:8001/visualization

Celery Worker

The Celery worker processes asynchronous tasks, such as training the ML model. Check the Celery container for logs and task execution.

• start celery beat and worker so that the jobs can be scheduled, inside the docerk container

  docker exec -it mlTrainingHub bash
  celery -A ml_training_hub beat --detach
  celery -A ml_training_hub worker --detach

Troubleshooting

• If you encounter issues, check the logs of individual containers for more details.

License

This project is licensed under the MIT License.