Add _FilterArgs NamedTuple to ImageFilter

[radarhere]

See if you think this makes the code clearer.

[akx]

I like the idea. Minor comments - things I had to look up to understand the flow, so maybe someone else would have to, too :)

[akx]

I'd personally like to call this divisor, because that's what it practically is (and in fact, that's the "argument name" in _filter() in _imaging.c).

[radarhere]

Kernel calls it scale.

Pillow/src/PIL/ImageFilter.py

Lines 72 to 78 in 5eee920

	def __init__(
	self,
	size: tuple[int, int],
	kernel: Sequence[float],
	scale: float | None = None,
	offset: float = 0,
	) -> None:

I would rather be consistent with the existing Python API.

[akx]

Sure thing.

[akx]

I'd kind of want to note that these are unpacked directly to ImagingCore.filter().

[radarhere]

Ok, I've added a comment.

[codspeed-hq]

Merging this PR will improve performance by 89.2%

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 4 improved benchmarks
✅ 599 untouched benchmarks
⏩ 335 skipped benchmarks¹

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
⚡	test_filter[1237x811-L-Smooth]	53.1 ms	16.5 ms	×3.2
⚡	test_filter[1237x811-L-Sharpen]	52.9 ms	16.5 ms	×3.2
⚡	test_filter[1237x811-I;16-Sharpen]	153.3 ms	137.3 ms	+11.64%
⚡	test_filter[1237x811-I;16-Smooth]	153.4 ms	137.5 ms	+11.63%

Tip

Curious why performance improved? Comment @codspeedbot explain why performance improved on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing radarhere:filter (749c66e) with main (e7f9779)²}

Footnotes

1. 335 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

2. No successful run was found on main (3c7080c) during the generation of this report, so e7f9779 was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩