chore: replace black, isort, absolufy-imports with ruff

[afuetterer]

PyPSA already uses ruff via pre-commit:
https://github.com/PyPSA/PyPSA/blob/master/.pre-commit-config.yaml

I replaced the tooling.

There are a few issues.

$ ruff check --statistics
202     UP007   [*] non-pep604-annotation
 93     UP006   [*] non-pep585-annotation
 82     F821    [ ] undefined-name
 22     F401    [ ] unused-import
 21     NPY002  [ ] numpy-legacy-random
 21     UP035   [ ] deprecated-import
 12     I001    [*] unsorted-imports
  8     W291    [*] trailing-whitespace
  6     E703    [*] useless-semicolon
  4     E712    [*] true-false-comparison
  3     F841    [*] unused-variable
  1     F601    [ ] multi-value-repeated-key-literal

[afuetterer]

I think this step can go as well:

linopy/.github/workflows/test.yml

Line 86 in b1a52dc

- name: Lint with flake8

[codecov]

Codecov Report

Attention: Patch coverage is 96.06299% with 5 lines in your changes missing coverage. Please review.

Project coverage is 86.00%. Comparing base (c6564a8) to head (dd0e54e).
Report is 5 commits behind head on master.

Files	Patch %	Lines
linopy/solvers.py	66.66%	2 Missing ⚠️
linopy/constraints.py	87.50%	0 Missing and 1 partial ⚠️
linopy/remote.py	0.00%	1 Missing ⚠️
linopy/variables.py	97.22%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #328      +/-   ##
==========================================
+ Coverage   85.95%   86.00%   +0.04%     
==========================================
  Files          17       17              
  Lines        4202     4209       +7     
  Branches      988      989       +1     
==========================================
+ Hits         3612     3620       +8     
+ Misses        431      430       -1     
  Partials      159      159              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[FabianHofmann]

Great! Still some issues in the benchmark folder with pre-commit https://results.pre-commit.ci/run/github/350276805/1721223837.ZE-YcKwTRuaSNDuB4ykaXA, we can either ignore that folder or quickly fix these

[FabianHofmann]

I think this step can go as well:

linopy/.github/workflows/test.yml

Line 86 in b1a52dc

- name: Lint with flake8

sure

[afuetterer]

Great! Still some issues in the benchmark folder with pre-commit https://results.pre-commit.ci/run/github/350276805/1721223837.ZE-YcKwTRuaSNDuB4ykaXA, we can either ignore that folder or quickly fix these

Maybe ignore F821 (snakemake is undefined)?

[FabianHofmann]

Great! Still some issues in the benchmark folder with pre-commit https://results.pre-commit.ci/run/github/350276805/1721223837.ZE-YcKwTRuaSNDuB4ykaXA, we can either ignore that folder or quickly fix these

Maybe ignore F821 (snakemake is undefined)?

sure!

[afuetterer]

$ ruff check --statistics
21      NPY002  numpy-legacy-random
 5      F821    undefined-name
 1      F401    unused-import
 1      F601    multi-value-repeated-key-literal

[afuetterer]

Which scip is relevant in this test? They can't both be used. Is this a typo?

linopy/test/test_optimization.py

Lines 410 to 413 in 717acba

	"scip": {"time_limit": 1},
	"xpress": {"maxtime": 1},
	"highs": {"time_limit": 1},
	"scip": {"limits/time": 1},

[FabianHofmann]

Which scip is relevant in this test? They can't both be used. Is this a typo?

linopy/test/test_optimization.py

Lines 410 to 413 in 717acba

	"scip": {"time_limit": 1},
	"xpress": {"maxtime": 1},
	"highs": {"time_limit": 1},
	"scip": {"limits/time": 1},

oh honestly, only one (ideally the correct one ;) ) should be used

[afuetterer]

Fabian Hofmann committed on Sep 19, 2023. :)
I guess the second (came later), will be the correct one.

[afuetterer]

$ ruff check --statistics
5       F821    undefined-name
1       F401    unused-import
1       F601    multi-value-repeated-key-literal

[lkstrp]

Could you also add a .git-blame-ignore-revs, similar as in PyPSA?
You can get a list of all pre-commit hashes via:

git log --pretty=format:'%H' --regexp-ignore-case --grep="auto fixes from pre-commit.com hooks"

And maybe don't squash this, so we keep those ruff changes separated.

The F841 leftovers I would just skip with a #noqa.

F401 I would just fix by adding it to __all__. Or is linopy.monkey_patch_xarray needed in the external API @FabianHofmann ?

[FabianHofmann]

F401 I would just fix by adding it to __all__. Or is linopy.monkey_patch_xarray needed in the external API @FabianHofmann ?

You mean whether the user should be able to import it? No, that not, but the module should internally be run when importing linopy. (hope I got your point)

[lkstrp]

Makes sense. Then it's a #noqa as well

[afuetterer]

Alright.

I will squash all ruff commits and add the one commit to .git-blame-ignore-revs?

So, you get 3 commits to merge, 1 commit adds the rule, 1 commit applies the rules and changes a lot of files, 1 commit adds the .git-blame-ignore-revs. Is that correct?

[FabianHofmann]

@afuetterer great work as always! just some minor things remaining where pre-commit complains, should be a quick thing to do, perhaps @lkstrp can also do it :)

[afuetterer]

Let me fix it quickly.

[afuetterer]

$ ruff check --statistics
2       F841    [*] unused-variable

ruff wants to remove the m in
m = n.lopf(solver_options=SOLVER_PARAMS, pyomo=False, solver_name=SOLVER).

[lkstrp]

$ ruff check --statistics
2       F841    [*] unused-variable

ruff wants to remove the m in m = n.lopf(solver_options=SOLVER_PARAMS, pyomo=False, solver_name=SOLVER).

Just add #noqa as well. They are needed for the benchmarks and it's better than inflating pyproject.toml.

As always, thanks a lot @afuetterer !

[afuetterer]

This should be it.

[lkstrp]

Beautiful!