Rewrite and refactor all of pycoast to meet minimum modern standards

[djhoese]

This PR makes every change necessary to make pycoast pass basic formatting, linting, and complexity checks. To be sure of that I have enabled pre-commit.ci and added a pre-commit config that can be used by developers to ensure the quality of the code moving forward. I am almost done with the flake8 complexity issues and then there are a few bandit security issues I need to look into.

• Closes #xxxx
• Tests added
• Tests passed
• Passes git diff origin/main **/*py | flake8 --diff
• Fully documented

[coveralls]

Coverage increased (+1.0%) to 95.386% when pulling 9ffe62c on djhoese:refactor-black into b2ad512 on pytroll:main.

[djhoese]

This could still definitely use a lot more work, but I've reduce the complexity a lot for the entire cw_base.py module. Some of it still isn't pretty, but I think it requires considering changing the algorithm. The hard part for handling the grid lines was that the minor lines depend on what the major lines are and a couple other odd combinations of things. This made it hard to separate the code further.

@lobsiger CodeFactor is finally happy 😉

[mraspaud]

Ok, I got tired of writing the same comment over and over, I think black should be fixed to allow 120 characters and not split stuff randomly.

[djhoese]

Ok, your comments have been addressed. I mentioned pytest in the test document. I updated black from using its default 88 character line length to 120.

[mraspaud]

Ok, your comments have been addressed. I mentioned pytest in the test document. I updated black from using its default 88 character line length to 120.

Great, I can continue reviewing then :) Thanks!!!

[mraspaud]

LGTM. Mixing the refactoring with "blackification" didn't help either, maybe the next time we do this, we should do two separate PRs? But from what I can see, the tests didn't change in functionality, so I think we're good.

[djhoese]

Yes, agreed. This started as a "add pre-commit" PR but to include all of the hooks I wanted it required rewriting most of the package. And yes, no behavior was intentionally changed. Merging now and then I could start working toward a config interface and automatically download the coastline files.

[lobsiger]

@djhoese @mraspaud now that this refactoring PR is merged, do we have a chance to see a Pycoast V. 1.6.0 anytime soon?

[djhoese]

@lobsiger I'm currently finishing my recovery from COVID and will need to submit a final report for one of my projects next week then catch up on all my other work. After all that, yes I'd like to look into a release. However, I noticed in the newest pre-commit PR #64 that one of the tests is failing with the new versions of dependencies. This test wasn't failing when this PR last ran it's CI. Debugging this test failure is the first major hurdle to getting a new release out. If you have any free time to investigate it would be much appreciated. Just narrowing the problem down to what versions of what dependencies causes the failure would be huge.