Description
Why:
Numpydoc ought to support annotations matching the typing functionality of numpy.typing and the built-in scalar types.
What:
I am using NumPy 2.3.0, and numpydoc 1.8.0.
I was writing some typed code using the NDArray dtype specifications, and found numpydoc would raise PR06 for the 'integer' keyword found in the scalar in-built numpy dtype integer.
import numpy as np, numpy.typing as npt
a : npt.NDArray[np.floating | np.integer]
When writing docstrings for such typing, numpydoc throws a PR06 error, due to the inclusion of the integer string.
Parameter "a" type should use "int" instead of "integer"
What I expected:
In the context of using numpy.typing and np.integer, I didn't expect to see numpydoc raise an error for a valid numpy in-build scalar datatype.
How we might solve this:
Either numpy / numpy.typing ought to be using the same rules (i.e. int instead of integer, float instead of floating etc) to represent the in-built dtypes, or (more likely) numpydoc should accommodate this labeling?
I also noticed we added an exclusion case for PR06 via skipping over the "{" character, for option sets defined by braces/curly-brackets in this issue / PR.