Skelleton Script for offering a ML model behind an API

Basic setup to offer a datascience prototype (e.g. machine learning model) behind an API using python, flask, and docker.

requires:

• serialized file of trained ML-model
• predict method that transforms the required input into a prediction using the trained ML-model
• docker and python 3 installed

outputs:

• miniconda docker container including
  • running python environment according to environment.yml
  • running main.py app within newly created environment
  • flask API that serves the datascience prototype
  • swagger spec of API based on docstring of a predict function within main.py