[pydocstyle] Handle arguments with the same names as sections (D417)

[ntBre]

Summary

Fixes #16007. The logic from the last fix for this (#9427) was sufficient, it just wasn't being applied because Attributes sections aren't expected to have nested sections. I just deleted the outer conditional, which should hopefully fix this for all section types.

Test Plan

New regression test, plus the existing D417 tests.

[github-actions]

ruff-ecosystem results

Linter (stable)

✅ ecosystem check detected no linter changes.

Linter (preview)

✅ ecosystem check detected no linter changes.

[MichaReiser]

Would you mind adding a test where one of the function arguments is named like a section header? I'm curious what happens if that argument is undocumented.

[ntBre]

Is this what you had in mind? I added these two cases, both of which cause D417. The first one is right, Args is not documented, but the second is still a false positive that I can keep working on.

# undocumented argument with the same name as a section
def f(payload, Args):
    """
    Send a message.

    Args:
        payload:
            The message payload.
    """


# documented argument with the same name as a section
def f(payload, Args):
    """
    Send a message.

    Args:
        payload:
            The message payload.

        Args:
            The other arguments.
    """

[MichaReiser]

Yes, that's what I had in mind

[ntBre]

These cases should work now! I initially deleted the inner conditional here

ruff/crates/ruff_linter/src/docstrings/sections.rs

Lines 525 to 529 in 1f7a29d

	if previous_section.indent_size < indent_size {
	if section_kind.as_str() != verbatim {
	return false;
	}
	}

which worked well for D417 but broke other rules like D214 - overindented-section. So I went for the more involved approach of passing down a Definition and extracting known parameter names from it.

One other thing I considered is just searching for a known parameter at line 514 instead of collecting a HashSet up front.