Feature request: Make F401 unused-imports configurable for modules with side effects

[Hnasar]

We have a handful of modules that have import side effects. I know, I know… we should not rely on that (#4056) – but in lieu of fixing everything, would it be possible to configure a list of modules that should never be auto-removed?

I see there's https://docs.astral.sh/ruff/settings/#lint_ignore-init-module-imports but I was hoping there's a chance to avoid this footgun which comes up from time to time by specifying which modules are expected to have import side effects.

[charliermarsh]

I'm supportive of something like this. I think it's been brought up before.

[hutch3232]

Not sure if this is useful, but here is a reproducible example of how I am encountering this issue when using the pyjanitor (https://github.com/pyjanitor-devs/pyjanitor) package.

pip install pandas ruff pyjanitor
pip freeze

multipledispatch==1.0.0
natsort==8.4.0
numpy==1.26.4
packaging==24.1
pandas==2.2.2
pandas-flavor==0.6.0
pyjanitor==0.27.0
python-dateutil==2.9.0.post0
pytz==2024.1
ruff==0.4.9
scipy==1.13.1
six==1.16.0
tzdata==2024.1
xarray==2024.6.0

example.py

import janitor
import pandas as pd

df = pd.DataFrame({"col": [0.25, 0.75]})
tbl = pd.DataFrame({"lo": [0., 0.5],
                    "hi": [0.5, 1.0]})

df.conditional_join(
    tbl,
    ("col", "lo", ">="),
    ("col", "hi", "<")
)

df.conditional_join
# <pandas_flavor.register.register_dataframe_method.<locals>.inner.<locals>.AccessorMethod object at 0x0000021B23F0AE70>

ruff check
# example.py:1:8: F401 [*] `janitor` imported but unused
# Found 1 error.
# [*] 1 fixable with the `--fix` option.

[zanieb]

It shouldn't be too hard to add a configuration option, if someone is interested.

[JonathanPlasse]

I'm supportive of something like this. I think it's been brought up before.

• resources imported but unused #4250

[Avasam]

This is something Pycln allows configuring, on top of having automated side-effect detections #8149

[MarcSkovMadsen]

+1 because import hvplot.pandas, import hvplot.xarray and similar imports adds a .hvplot api for plotting to a Pandas DataFrame, XArray dataset etc.

For more info see holoviz/hvplot#1399