Often, I would work on a Python module and add changes to it live, which I would then want to test immediately. Piece of cake! Add the changes and execute the script again. Well, that’s certainly a way, but as modules get bigger, they tend to import other modules, or do some preliminary setup work.
With time, I have developed this practice of opening up a Python REPL (though I recommend IPython or the Jupyter Console for that), importing my in-progress module and singling out separate functions I would like to test. The problem is, when I do changes to the Python code (like, add a new function), they are not immediately usable, because neither the Python REPL, nor IPython / Jupyter would auto-reload them.
Give me the code
Yes, we are coming to that
Python 3 supports inline reloading of modules using a function called, well clever enough,
reload. It used to be a built-in function in Python 2, but this is no longer the case. If you are using Python 3.2+, you should import it extra:
For Python 3.2 and 3.3:
For Python 3.4+:
Jupyter comes with a set of extensions only applicable to the Jupyter/iPython sessions, called magics. One of these “magics” is the ability to load custom extensions, one of which allows auto-reloading of modules. To enable this, you should add the following two commands, before any imports:
# you can check out the the documentation for the rest of the autoreaload modes
# by apending a question mark to %autoreload, like this:
This way, as soon as you hit
Save in your code editor, you should be able to re-run a Jupyter cell or an iPython line again, and if it is calling your module, it should automatically call the latest version.