MNT: Enforce ruff/flake8-bugbear rules (B) #28765

DimitriPapadopoulos · 2025-04-18T10:44:58Z

No description provided.

doc/neps/nep-0016-benchmark.py

numpy/_core/printoptions.py

numpy/_core/tests/test_deprecations.py

numpy/_core/tests/test_multiarray.py

numpy/lib/tests/test_function_base.py

numpy/ma/core.py

numpy/testing/_private/utils.py

tools/refguide_check.py

jorenham

I have the feeling that there are some threading subtleties that we might be missing here. Perhaps @ngoldbaum could also take a look at this?

numpy/_core/printoptions.py

DimitriPapadopoulos · 2025-04-19T17:54:51Z

Indeed, the underlying issue remains. I'm not familiar with ContextVar and cannot find a simple way to fix it. See Factory for contextvar default value.

ngoldbaum · 2025-04-19T21:13:54Z

I’ll try to take a look at this next week.

charris · 2025-04-25T19:18:47Z

Needs a rebase.

jorenham · 2025-04-28T16:48:09Z

this needs a rebase

numpy/testing/_private/utils.py

jorenham · 2025-04-28T16:56:18Z

tools/refguide_check.py

+    module_names = args.module_names + [
+        OTHER_MODULE_DOCS[name]
+        for name in args.module_names
+        if name in OTHER_MODULE_DOCS
+    ]
+    # remove duplicates while maintaining order
+    module_names = list(dict.fromkeys(module_names))


Technically speaking, this is a bit different now, because this also de-duplicates args.module_names, which didn't happen before. I can't imagine why that would be a problem, but I figured it could be worth noting 🤷🏻

I may be missing something, but I don't think it de-duplicates args.module_names. The code is equivalent to:

>>> args_module_names = [ 1, 1, 2, 2, 3, 3 ] >>> module_names = args_module_names + [ x for x in args_module_names if x != 2 ] >>> >>> args_module_names [1, 1, 2, 2, 3, 3] >>> module_names [1, 1, 2, 2, 3, 3, 1, 1, 3, 3] >>> >>> module_names = list(dict.fromkeys(module_names)) >>> >>> args_module_names [1, 1, 2, 2, 3, 3] >>> module_names [1, 2, 3] >>>

args.module_names is part of module_names, and module_names is what's being de-duped as a whole

How is args.module_names part of module_names? I don't get it, module_names is a whole new list containing values copied from args.module_names. See my example.

jorenham

There's nothing left for me to complain about 👌🏻. So unless @ngoldbaum has any objections, this should be good to merge.

Using `hasattr(x, "__call__")` to test if x is callable is unreliable. Use `callable(x)` for consistent results.

Do not perform function call in argument defaults; instead, perform the call within the function, or read the default from a module-level singleton variable In all these cases, the argument is designed to be persistent.

Do not assert blind exception: `Exception`

Found useless expression. Either assign it to a variable or remove it.

Loop control variable overrides iterable it iterates

abstract base class has no abstract methods or properties

Star-arg unpacking after a keyword argument is strongly discouraged

`except` handlers should only be exception classes or tuples of exception classes

Do not use mutable data structures for `ContextVar` defaults

B909 Mutation to loop iterable during iteration

Co-authored-by: Joren Hammudoglu <jhammudoglu@gmail.com>

ngoldbaum · 2025-04-28T18:07:23Z

I think you need to do a little more to remove memusage, including in the type stubs, typing tests, and the __all__ for the numpy.testing._private.utils namespace.

The memusage function is only exercised in typing tests.

DimitriPapadopoulos · 2025-04-29T07:19:19Z

I'm not sure why the compilation fails on Windows.

numpy/random/_pcg64.pyx compilation error

  [511/521] Compiling Cython source D:/a/1/s/numpy/random/_pcg64.pyx
  FAILED: numpy/random/_pcg64.cp312-win_amd64.pyd.p/numpy/random/_pcg64.pyx.c
  "cython" "-M" "--fast-fail" "-3" D:/a/1/s/numpy/random/_pcg64.pyx -o numpy/random/_pcg64.cp312-win_amd64.pyd.p/numpy/random/_pcg64.pyx.c
  [512/521] Compiling Cython source numpy/random/_generator.pyx
  [513/521] Compiling Cython source numpy/random/mtrand.pyx
  ninja: build stopped: subcommand failed.
  Activating VS 16.11.46
  INFO: automatically activated MSVC compiler environment
  INFO: autodetecting backend as ninja
  INFO: calculating backend command to run: C:\hostedtoolcache\windows\Python\3.12.10\x64\Scripts\ninja.EXE
  error: subprocess-exited-with-error
  
  Preparing metadata (pyproject.toml) did not run successfully.
  exit code: 1
  
  See above for output.
  
  note: This error originates from a subprocess, and is likely not a problem with pip.
  full command: 'C:\hostedtoolcache\windows\Python\3.12.10\x64\python.exe' 'C:\hostedtoolcache\windows\Python\3.12.10\x64\Lib\site-packages\pip\_vendor\pyproject_hooks\_in_process\_in_process.py' prepare_metadata_for_build_wheel 'C:\Users\VSSADM~1\AppData\Local\Temp\tmpxa2humy0'
  cwd: D:\a\1\s
  Preparing metadata (pyproject.toml): finished with status 'error'
error: metadata-generation-failed

Encountered error while generating package metadata.

ngoldbaum · 2025-04-29T13:50:57Z

That's odd, let's see if it's reproducible. I re-triggered the failed build.

ngoldbaum · 2025-04-29T16:46:50Z

numpy/testing/_private/utils.py

@@ -36,7 +36,7 @@
        'assert_equal', 'assert_almost_equal', 'assert_approx_equal',
        'assert_array_equal', 'assert_array_less', 'assert_string_equal',
        'assert_array_almost_equal', 'assert_raises', 'build_err_msg',
-        'decorate_methods', 'jiffies', 'memusage', 'print_assert_equal',


Argh, I'm really sorry for not realizing this until now, but it turns out that np.testing._private.utils.memusage is re-exported in the public NumPy API as np.testing.memusage:

In [1]: import numpy as np In [2]: np.testing.memusage Out[2]: <function numpy.testing._private.utils.memusage()> In [3]: np.testing.memusage? Signature: np.testing.memusage() Docstring: Return memory usage of running python. [Not implemented] File: ~/.pyenv/versions/3.13.3/lib/python3.13/site-packages/numpy/testing/_private/utils.py Type: function

So, we definitely shouldn't delete it. Sorry for pointing you in completely the wrong direction.

Now, that said, there's all kinds of cruft in the np.testing namespace we should deprecate and clean up, but that should be in a separate PR.

Can you revert the changes that deleted memusage? Again, sorry for pointing you in the wrong direction in my prior code review. I'm happy to do this myself if you would like me to fix my own mistake :)

Sure, will do that.

This reverts commit 371b363.

github-actions bot added the 03 - Maintenance label Apr 18, 2025

jorenham reviewed Apr 18, 2025

View reviewed changes

DimitriPapadopoulos force-pushed the B branch 2 times, most recently from 0b85500 to 6731f90 Compare April 18, 2025 18:37

jorenham reviewed Apr 19, 2025

View reviewed changes

numpy/_core/printoptions.py Show resolved Hide resolved

DimitriPapadopoulos force-pushed the B branch from 6731f90 to 81d24ab Compare April 27, 2025 09:58

jorenham reviewed Apr 28, 2025

View reviewed changes

numpy/testing/_private/utils.py Show resolved Hide resolved

jorenham reviewed Apr 28, 2025

View reviewed changes

jorenham approved these changes Apr 28, 2025

View reviewed changes

DimitriPapadopoulos and others added 16 commits April 28, 2025 19:10

STY: Apply ruff/flake8-bugbear rule B004

c6727aa

Using `hasattr(x, "__call__")` to test if x is callable is unreliable. Use `callable(x)` for consistent results.

MNT: Apply ruff/flake8-bugbear rule B008

7127ab6

Do not perform function call in argument defaults; instead, perform the call within the function, or read the default from a module-level singleton variable In all these cases, the argument is designed to be persistent.

MNT: Apply ruff/flake8-bugbear rule B017

9261be7

Do not assert blind exception: `Exception`

MNT: Apply ruff/flake8-bugbear rule B018

e056d68

Found useless expression. Either assign it to a variable or remove it.

MNT: Apply ruff/flake8-bugbear rule B020

166c886

Loop control variable overrides iterable it iterates

MNT: Apply ruff/flake8-bugbear rule B024

a5d4c34

abstract base class has no abstract methods or properties

MNT: Apply ruff/flake8-bugbear rule B026

cfef76d

Star-arg unpacking after a keyword argument is strongly discouraged

MNT: Apply ruff/flake8-bugbear rule B030

97ca947

`except` handlers should only be exception classes or tuples of exception classes

MNT: Apply ruff/flake8-bugbear B039

795c40a

Do not use mutable data structures for `ContextVar` defaults

MNT: Apply ruff/flake8-bugbear rule B909

4577317

B909 Mutation to loop iterable during iteration

MNT: Enforce ruff/flake8-bugbear rules (B)

0ba82cc

MNT: Better solution for B008

178936b

Co-authored-by: Joren Hammudoglu <jhammudoglu@gmail.com>

MNT: Better solution for B020

16752e8

Co-authored-by: Joren Hammudoglu <jhammudoglu@gmail.com>

MNT: Better solution for B030

713b177

Co-authored-by: Joren Hammudoglu <jhammudoglu@gmail.com>

MNT: Better solution for B909

b51d677

Co-authored-by: Joren Hammudoglu <jhammudoglu@gmail.com>

MNT: Fix new CI issue

754e534

DimitriPapadopoulos force-pushed the B branch from 74cc316 to 7e3940a Compare April 28, 2025 17:26

DimitriPapadopoulos force-pushed the B branch from 7e3940a to 741eb67 Compare April 28, 2025 17:26

DimitriPapadopoulos force-pushed the B branch from 741eb67 to e1a7bb3 Compare April 28, 2025 21:57

MNT: get rid of dead code

371b363

The memusage function is only exercised in typing tests.

DimitriPapadopoulos force-pushed the B branch from e1a7bb3 to 371b363 Compare April 28, 2025 22:02

ngoldbaum reviewed Apr 29, 2025

View reviewed changes

MNT: Revert "MNT: get rid of dead code"

a1c4d2d

This reverts commit 371b363.

ngoldbaum approved these changes Apr 29, 2025

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

MNT: Enforce ruff/flake8-bugbear rules (B) #28765

MNT: Enforce ruff/flake8-bugbear rules (B) #28765

DimitriPapadopoulos commented Apr 18, 2025

jorenham left a comment

DimitriPapadopoulos commented Apr 19, 2025

ngoldbaum commented Apr 19, 2025

charris commented Apr 25, 2025

jorenham commented Apr 28, 2025

jorenham Apr 28, 2025 •

edited

Loading

DimitriPapadopoulos Apr 28, 2025

jorenham Apr 28, 2025

DimitriPapadopoulos Apr 28, 2025

jorenham left a comment

ngoldbaum commented Apr 28, 2025

DimitriPapadopoulos commented Apr 29, 2025 •

edited

Loading

ngoldbaum commented Apr 29, 2025

ngoldbaum Apr 29, 2025

DimitriPapadopoulos Apr 29, 2025

MNT: Enforce ruff/flake8-bugbear rules (B) #28765

Are you sure you want to change the base?

MNT: Enforce ruff/flake8-bugbear rules (B) #28765

Conversation

DimitriPapadopoulos commented Apr 18, 2025

jorenham left a comment

Choose a reason for hiding this comment

DimitriPapadopoulos commented Apr 19, 2025

ngoldbaum commented Apr 19, 2025

charris commented Apr 25, 2025

jorenham commented Apr 28, 2025

jorenham Apr 28, 2025 • edited Loading

Choose a reason for hiding this comment

DimitriPapadopoulos Apr 28, 2025

Choose a reason for hiding this comment

jorenham Apr 28, 2025

Choose a reason for hiding this comment

DimitriPapadopoulos Apr 28, 2025

Choose a reason for hiding this comment

jorenham left a comment

Choose a reason for hiding this comment

ngoldbaum commented Apr 28, 2025

DimitriPapadopoulos commented Apr 29, 2025 • edited Loading

ngoldbaum commented Apr 29, 2025

ngoldbaum Apr 29, 2025

Choose a reason for hiding this comment

DimitriPapadopoulos Apr 29, 2025

Choose a reason for hiding this comment

jorenham Apr 28, 2025 •

edited

Loading

DimitriPapadopoulos commented Apr 29, 2025 •

edited

Loading