Python(fix): clean docstrings captured in pytest

[alexluck-sift]

The pytest plugin stored test step descriptions straight from __doc__. For leaf steps (test functions) the raw docstring was used as-is, so an indented multi-line docstring kept its source indentation on every line after the first, plus surrounding blank lines. For hierarchy parents (package/module/class) only .strip() was applied, which trims the ends but leaves interior indentation in place.

Both extraction points now use inspect.getdoc, which dedents interior lines, trims surrounding blank lines, and falls back to an inherited docstring when the object itself has none. The existing try/except guards and the empty-to-None behavior are preserved.

The leaf-step test is updated to assert the stored description against the dedented literal and to confirm it differs from the raw __doc__, so a regression back to raw storage fails the test.

[github-actions]

Python docs preview: https://sift-stack.github.io/sift/python/pr-658/

Deployed from cca7eab. The link may take up to a minute to become live as GitHub Pages propagates.