Initially all you should have to do is to put the command you run from the command line into an exec node, but specify the full path the the python command and also the full path to the python file. Add debug nodes on all outputs of the exec node, give them names so you can tell which output is which, and set them Output Complete Message and see what happens. Maybe a good idea to initially modify the python to just run once rather than looping.
However, it really would be better to use MQTT, you will not regret the few hours needed to understand and use MQTT. Particularly if you intend to keep using node-red.