[refurb] New rule to suggest min/max over sorted() (FURB192)

[ottaviohartman]

Summary

Fixes #10463

Add FURB192 which detects violations like this:

# Bad
a = sorted(l)[0]

# Good
a = min(l)

There is a caveat that @Skylion007 has pointed out, which is that violations with reverse=True technically aren't compatible with this change, in the edge case where the unstable behavior is intended. For example:

from operator import itemgetter
data = [('red', 1), ('blue', 1), ('red', 2), ('blue', 2)]

min(data, key=itemgetter(0))  # ('blue', 1)
sorted(data, key=itemgetter(0))[0]  # ('blue', 1)
sorted(data, key=itemgetter(0), reverse=True)[-1]  # ('blue, 2')

This seems like a rare edge case, but I can make the reverse=True fixes unsafe if that's best.

Test Plan

This is unit tested.

References

https://github.com/dosisod/refurb/pull/333/files

[ottaviohartman]

This auto-generated FURB19 but I can't figure out why

[github-actions]

ruff-ecosystem results

Linter (stable)

✅ ecosystem check detected no linter changes.

Linter (preview)

ℹ️ ecosystem check detected linter changes. (+4 -0 violations, +0 -0 fixes in 3 projects; 41 projects unchanged)

apache/airflow (+1 -0 violations, +0 -0 fixes)

ruff check --no-cache --exit-zero --ignore RUF9 --output-format concise --preview --select ALL

+ airflow/providers/yandex/secrets/lockbox.py:258:16: FURB192 [*] Prefer `min` over `sorted()` to compute the minimum value in a sequence

rotki/rotki (+2 -0 violations, +0 -0 fixes)

ruff check --no-cache --exit-zero --ignore RUF9 --output-format concise --preview

+ rotkehlchen/db/dbhandler.py:269:25: FURB192 [*] Prefer `max` over `sorted()` to compute the maximum value in a sequence
+ rotkehlchen/globaldb/handler.py:127:21: FURB192 [*] Prefer `max` over `sorted()` to compute the maximum value in a sequence

indico/indico (+1 -0 violations, +0 -0 fixes)

ruff check --no-cache --exit-zero --ignore RUF9 --output-format concise --preview

+ indico/modules/events/registration/util.py:272:16: FURB192 [*] Prefer `min` over `sorted()` to compute the minimum value in a sequence

Changes by rule (1 rules affected)

code	total	+ violation	- violation	+ fix	- fix
FURB192	4	4	0	0	0

[Skylion007]

I'd definitely advocate for making unstable sort fixes unsafe. Otherwise, it could unknowingly change user code. I'd also argue we may want add to add a config option for this rule that ignores unstable sorts.

[ottaviohartman]

Is there a cleaner or more canonical way to generate this?

[ottaviohartman]

@Skylion007

I'd definitely advocate for making unstable sort fixes unsafe. Otherwise, it could unknowingly change user code.

Agreed, and updated!

I'd also argue we may want add to add a config option for this rule that ignores unstable sorts.

Shall I do this too? Haven't heard anyone else chime in yet.

[codspeed-hq]

CodSpeed Performance Report

Merging #10868 will not alter performance

_{Comparing ottaviohartman:thartman/furb192 (da36bfc) with main (925c7f8)}

Summary

✅ 30 untouched benchmarks

[zanieb]

Shall I do this too? Haven't heard anyone else chime in yet.

I probably wouldn't implement it now since the rule is just going into preview. We can use preview feedback to consider adding an option. I'd just open a tracking issue when this is merged.

[charliermarsh]

Thanks!