MNT: Fix handling of ints in hsv_to_rgb() - #32058

PR summary

On NumPy 2.0, hsv_to_rgb() raises ValueError: Unable to avoid copy while creating an array as requested for integer input, for example:

import numpy as np
import matplotlib.colors as mcolors
mcolors.hsv_to_rgb(np.array([0, 1, 1]))  # ValueError on NumPy 2.0

The function built its working array with np.array(hsv, copy=False, dtype=np.promote_types(...)). When the input dtype differs from the target (as with integer input), NumPy 2.0 needs a copy to perform the cast and copy=False forbids it, so the call raises. The "Don't work on ints" comment shows integer input is meant to be promoted to float, so this is a regression rather than intended behavior.

This is the same NumPy 2.0 regression that #30815 fixed in the twin function rgb_to_hsv(), but the same fix was never applied to hsv_to_rgb(). This mirrors it: use np.asarray() for the dtype promotion and expand a 1D input to 2D explicitly. Behavior for float and 2D input is unchanged. Adds test_hsv_to_rgb_int().

AI Disclosure

AI (Claude Code) was used to locate the missing twin fix. I drafted the change, wrote the regression test, reviewed and verified the change, including reproducing the crash and confirming the fix returns the correct colors.

PR checklist

• [N/A] "closes #0000" is in the body of the PR description (no separate issue; this is the counterpart to MNT: Fix handling of ints in rgb_to_hsv() #30815)
• new and changed code is tested
• [N/A] Plotting related features are demonstrated in an example
• [N/A] New Features and API Changes are noted with a directive and release note
• Documentation complies with general and docstring guidelines