Template application for deploy data science models as Restfull API. Build with flask, docker, jenkins and cookiecutter.
$ pip install cookiecutter $ cookiecutter https://github.com/RiccardoNizzolo/cookiecutter-devops-datascience.git
You will be asked about your basic info (name, project name, app name, etc.). This info will be used in your new project.
- Restfull API for each trained model
- Modularized design to avoid information-leaks and set-up a valid experiment
- Automatic Testing for the data science pipeline
- History of all trained models with train/validation/test performance, commit id and other statistics
- Jenkins CI/CD Pipeline with models image building
- Deploy models as Docker containers
- Flask-SQLAlchemy ready
- Easy database migrations with Flask-Migrate
- Configuration in environment variables, as per The Twelve-Factor App
- Flask's Click CLI configured with simple commands
- Utilizes best practices: Blueprints and Application Factory patterns
MIT licensed.