Factory method for our application. It returns an instance of our Flask application. Default configuration has the TFX serving predict API on http://localhost:8503/v1/models/alpha:predict
Using waitress, starts up a production server on the input host and port
# The following are test cases for deployment