PERF: Series.any/all

[jbrockmendel]

• closes Series.all much slower than Series.values.all #26032 (Replace xxxx with the GitHub issue number)
• Tests added and passed if fixing a bug or adding a new feature
• All code checks passed.
• Added type annotations to new arguments/methods/functions.
• Added an entry in the latest doc/source/whatsnew/vX.X.X.rst file if fixing a bug or adding a new feature.

Example based on OP from #26032

import numpy as np
import pandas as pd
ser = pd.Series(np.random.randint(0, 2, 100000)).astype(bool)

In [2]: %timeit s.any(skipna=True)
5.79 µs ± 719 ns per loop (mean ± std. dev. of 7 runs, 100,000 loops each)  # <- PR
8.44 µs ± 473 ns per loop (mean ± std. dev. of 7 runs, 100,000 loops each)  # <- main

We'll be hard-pressed to get trim any more overhead from this, so I'm declaring this as closing #26032.

[mroeschke]

This seems only really needed for the Series case so can it be moved & simplified there?

[jbrockmendel]

doing it in series.py requires importing a bunch of currently-private items that the linter complains about. Also before long I'd like todo this with more of the reductions in both Series/DataFrame.

[mroeschke]

Thanks @jbrockmendel