This Python image is a very thin wrapper around the default Python image (https://hub.docker.com/_/python/). The wrapping is done to make it work with the Nerd CLI. For a step-by-step guide on how to use this image with the Nerd CLI, see https://www.nerdalize.com/applications/python.
How does it work?
Both images take the respective official Python image as a base. To support both Python2 and Python3 we created two images. The image names are:
Upon execution the container will do a
pip install if it finds the file
/input/requirements.txt (note that
/input is the folder that contains the contents of the input dataset).
It will then run the Python script in
/input that was given as an argument via
nerd task create (e.g.
nerd task create <workload-id> -- main.py will run
Any files written to the folder
/output can as usual be downloaded after execution using
nerd dataset download <output-id>.