/ml_blink_api

Server API for ML Blink.

Primary LanguagePython

ML-Blink API

Server API for ML Blink.

Local Development

  • Install Docker
  • Create an .env file using .env.example as a starting point (i.e., cp .env.example .env)
    • Set FLASK_ENV=development
    • There's no need to set the MONGO_URI for local development
    • Set ORIGINS=http://localhost:8080
  • Unzip the beta-pack dataset images (PanSTARRS_ltd and USNO1001) in /ml_blink_api/static/beta_images in their corresponding directories
  • Run docker-compose build to build the Docker services
    • Only run this once or if the Dockerfile has been updated
  • Start: docker-compose up
  • Stop : docker-compose down
  • Run mongo 127.0.0.1:27017 to start the MongoDB shell
    • This is not required to start the application, but can be useful when you want to manually check the database's collections
  • Rebuild the application (changes to the requirements.txt or Docker configuration file require a rebuild):
docker-compose up --build

API Documentation

  • Documentation of the endpoints exposed by the ML-Blink API, the methods each of them supports, and the attributes these expect.

Server Configuration

  • A guide to configure an Ubuntu 16.04 server to serve the ml_blink_api Flask application in the SNIC Cloud.

Deployment

  • This is a very rudimentary deployment script. It assumes the server is setup as explained in the server configuration guide.
  • SSH in the server by running ssh -i <file_name>.pem ubuntu@<FLOATING_IP_ADDRESS> and run:
sudo systemctl stop celeryd
sudo service apache2 stop
cd /var/www/ml_blink_api
sudo git pull
sudo pip3 install -r requirements.txt
sudo service apache2 restart
sudo systemctl restart celeryd