Replace flake8 with ruff

[michplunkett]

Description of Changes

Replacing flake8 with ruff, which has approximate parity and is more than 10x faster: https://docs.astral.sh/ruff/faq/#how-does-ruffs-linter-compare-to-flake8

Tests and Linting

• This branch is up-to-date with the develop branch.
• pytest passes on my local development environment.
• pre-commit passes on my local development environment.

[michplunkett]

valuelist is an un-pythonic name.

[michplunkett]

Ignored C901: https://www.flake8rules.com/rules/C901.html

[abandoned-prototype]

This looks like a nice improvement. Thanks!

[abandoned-prototype]

Not sure I would agree with the changes in this file, but also don't feel strongly about it. Was this something ruff suggested?

[michplunkett]

The changes were made due to this rule. It called out 10+ other functions that had more than 10 if/else statements, and I changed the ones that could be done with the least amount of re-working.

If you'd like to see the other functions, you could remove C901 from the ignore list and rerun the linter.

[michplunkett]

I think we could reimplement C901 in another PR, but it would require a lot of changes:

ruff.....................................................................Failed
- hook id: ruff
- exit code: 1

OpenOversight/app/commands.py:183:5: C901 `update_officer_from_row` is too complex (14 > 10)
OpenOversight/app/commands.py:473:5: C901 `bulk_add_officers` is too complex (14 > 10)
OpenOversight/app/csv_imports.py:156:5: C901 `_handle_assignments_csv` is too complex (18 > 10)
OpenOversight/app/main/model_view.py:113:9: C901 `edit` is too complex (11 > 10)
OpenOversight/app/main/views.py:745:5: C901 `edit_department` is too complex (15 > 10)
OpenOversight/app/main/views.py:883:5: C901 `list_officer` is too complex (18 > 10)
OpenOversight/app/main/views.py:1374:5: C901 `label_data` is too complex (11 > 10)
OpenOversight/app/models/database_imports.py:102:5: C901 `update_officer_from_dict` is too complex (11 > 10)
OpenOversight/app/utils/forms.py:273:5: C901 `filter_by_form` is too complex (21 > 10)
Found 9 errors.

black....................................................................Passed
mypy.....................................................................Passed

[michplunkett]

This folder no longer exists, so I removed the exclude.

[sea-kelp]

I haven't had time to review all the changes in this PR but I want to push back against large refactoring changes like this without good reason.

The previous djlint change introduced several minor bugs which weren't surfaced until the changes were live in production due to the scale of the refactoring. Comparing the build times, this ruff change built in 5m 26s while an earlier build took 4m 30s.

I mention this because our fork will need to pull in changes like this to keep in sync with the upstream. I definitely don't mean to block adding important guardrails but wanted to get a temp check since we expect ruff to have rough parity with our current setup and aren't seeing the performance enhancements.

[michplunkett]

@sea-kelp, I definitely get the hesitancy. I BELIEVE most every part of the code modified here is covered in test cases. Would you feel more comfortable with these changes if I were to verify that every line change was covered by tests?

I also think there's a difference between changes in HTML versus Python syntax changes. The Python changes can be verified through a handful of mechanisms that are generally unavailable to HTML changes. Most of the changes here are also syntax changes that are more in line with Python best practices, which I think is of value as well (e.g., isinstance versus type comparisons, removal of redundant map function calls, etc.).

If there's a specific change you're curious or have questions about, I'd certainly be willing to go more into it.

[michplunkett]

Thanks for taking a look, @sea-kelp! I'll let you know when the changes are made.