Enable numpydoc docstring validation for astropy.uncertainty and fix the issues - #20447

Description

This enables the numpydoc pre-commit check for astropy.uncertainty (one subpackage of #20353) and fixes the issues it reports.

Most of them come from docstrings that describe the parameters in prose rather than in a Parameters section, so numpydoc cannot match them against the signature:

• Distribution.pdf_mean / pdf_std / pdf_var said "Arguments are as for numpy.mean"; these now have proper Parameters sections.
• Distribution.pdf_percentiles / pdf_histogram and the three distributions.py factory functions said "Additional keywords are passed into ..." as a trailing sentence; these are now **kwargs entries.
• broadcast_arrays and concatenate in function_helpers.py had no Parameters section at all.

The rest are trailing blank lines at the end of a docstring (GL03), missing descriptions for ArrayDistribution.view and get_n_samples, and a few things in the private _result_as_distribution (missing space before the colon, no description for the return value).

Two descriptions were wrong rather than missing, so I corrected them while I was there:

• broadcast_arrays: subok does not control whether subclasses of Distribution are allowed (NdarrayDistribution is one either way). What it controls is whether ndarray subclasses such as Quantity are preserved.
• concatenate: arrays was documented as a sequence of array-like, but entries without a shape are passed straight through to numpy.concatenate, so a list or a scalar raises. Also pdf_percentiles referred to a fracs parameter that does not exist and declared the result dimensionless, while it actually keeps the unit of the samples.

One exclusion was needed. Distribution documents samples in the class docstring but takes it in __new__, and the pre-commit hook only reads signatures from __init__, so it reports samples as an unknown parameter. I have added it to [tool.numpydoc_validation] exclude with a comment, but happy to solve it differently if you would rather not carry an exclusion for this.

No behaviour changes. pytest --pyargs astropy.uncertainty passes (148 passed, 3 skipped), and numpydoc lint is clean with both numpydoc 1.10.0 (the version pinned in .pre-commit-config.yaml) and 1.11.0.

AI Disclosure

If AI tools were used to develop this pull request, describe the tools including specific model and version, how they were used, and what content is AI generated. Otherwise enter "N/A".

The docstring edits were drafted with Claude Opus 5 via Claude Code. I have gone over the full diff myself, and verified locally that python -m numpydoc.hooks.validate_docstrings astropy/uncertainty/*.py reports nothing and that the subpackage test suite still passes.

• I certify that I am human and that I take full responsibility for this pull request including all interactions with reviewers.

Merge method

• By checking this box, the PR author has requested that maintainers do NOT use the "Squash and Merge" button. Maintainers should respect this when possible; however, the final decision is at the discretion of the maintainer that merges the PR.