I’m doing data analysis, and for reproducibility I want to make sure my results are marked with the version of the code used to produce them. I do this with

import git
repo = git.Repo(search_parent_directories=True)
last_commit_hex = repo.head.object.hexsha

But that’s not necessarily the actual state of the code if there’s uncommitted changes. So I want an automatic check in the program just before it starts running the analysis. I can use repo.is_dirty() to check for uncommitted changes in the repo. But that’s checking too much. Besides the core package, there’s also a bunch of miscellaneous files. Testing, other analysis scripts that use that core package, etc. And those don’t need to be fully committed, but will trigger is_dirty(). How do I only check for the relevant files?

So, if my file system looks like this:

Repo folder
|- package_folder
   |- __init__.py
   |- module1.py
   |- module2.py
   |- ...
|- some_analysis_script1.py
|- some_analysis_script2.py
|- some_testing.py
|- ...

I would want to check only the contents of package_folder. Getting the list of files can be done with os.listdir, but how do I check which ones have been changed?