Add GitHub workflows to check black, pylint, git-fleximod

[samsrabin]

Description of changes

• Adds workflows to test git-fleximod and pylint.
• black and pylint workflows only run when relevant files are changed
• black and pylint workflows use ctsm_pylib conda environment

Specific notes

Contributors other than yourself, if any: None

CTSM Issues Fixed:
Because I'm assuming we only want to support people running/linting our tools in ctsm_pylib and not other Python versions:

• Closes Add an action to test python code through different python versions #1580

Are answers expected to change (and if so in what way)? No

Any User Interface Changes (namelist or namelist defaults changes)? No

Does this create a need to change or add documentation? Did you do so? No

Testing performed, if any: None needed except what GitHub does.

[samsrabin]

@ekluzek Reminding you to review, as discussed. Thanks!

[samsrabin]

Blocked by #2982. Once that's done: Now unblocked, so:

• Update Python versions tested
• Use our conda env instead of requirements.txt.

[samsrabin]

#2982 is merged, so this is unblocked.

[samsrabin]

@ekluzek This is complete and ready for review. Thanks!

[samsrabin]

Blocking #2809.

[ekluzek]

I have one thing I suggest you do. But, other than that this is great. You found a solution for a problem that I didn't, so the resulting workflow is better than before.

The one suggestion is to possibly keep the two step create and then initialize in py_env_create. So we should talk if you disagree.

So I've marked as approve so you don't have to wait on anything from me. We could even talk about the two step thing later.

[ekluzek]

This also gets me thinking about adding some more tests in a future PR that could be done at some frequency. The python unit tests for example. Maybe we add some longer tests like them to happen less often then with each push, but maybe even something like nightly.

[samsrabin]

Running tests on a schedule is definitely a good idea in general. I actually have one in the docs revamp PR that checks once per week that ctsm_pylib is able to be built, because it's possible at some point that some Python module version gets taken off the conda/mamba channels, resulting in the solver failing. I added that because it's something that would happen outside of this repo that would cause the failure. We might want to run the Python unit tests weekly for the same reason, although we'd need to resolve #2984 (at least the unit test bit) first.