Figure out notebook pre-commit hooks vs. post-commit stripping

[jcohenadad]

Oops! I became lazy to remove my notebook outputs before committing, thinking that GHA would take care of it, but I guess the difference with a pre-commit hook, is that the commit with the notebook outputs becomes part of the git history, and then a bot strips it out and add another commit to it:

This is pretty annoying. I guess I should be looking at pre-commit hook approach, unless there is another solution?

[jcohenadad]

In fact, the auto-strip commit is annoying because it systematically creates a conflict on the local branch:

We should probably get rid of it...

[mathieuboudreau]

Maybe this? https://zhauniarovich.com/post/2020/2020-06-clearing-jupyter-output/

[mguaypaq]

The updated post using git attributes is pretty nice:
https://zhauniarovich.com/post/2020/2020-10-clearing-jupyter-output-p3/

It's very similar to how git-annex already works.

[jcohenadad]

Thank you @mathieuboudreau and @mguaypaq , I will try that and post feedback