How to specify that parameter can equal the string 'integer'?

[MarcoGorelli]

Say I have the following:

def foo(bar: str) -> None:
    """
    Parameters
    ----------
    bar : {'integer', 'float'}
        Some description.
    """
    if bar not in {'integer', 'float'}:
        raise ValueError('bar is neither "integer" nor "float"')

The parameter bar is a str, and can only take on the values 'integer' and 'float'. However, if I run numpydoc on it, it reports:

$ python -m numpydoc t.foo --validate
t.foo:SS01:No summary found (a short summary in a single line should be present at the beginning of the docstring)
t.foo:ES01:No extended summary found
t.foo:PR06:Parameter "bar" type should use "int" instead of "integer"
t.foo:SA01:See Also section not found
t.foo:EX01:No examples section found

The first 2 and last 2 errors are expected, but the middle one (PR06) feels like a false positive.

Noticed when trying to validate pandas.to_numeric (https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.to_numeric.html), which has a parameter downcast which can be equal to the string 'integer'

[rossbar]

You're right, this is a bug in the numpydoc validation which fails to account for the fact that these values are part of a set of options.

[MarcoGorelli]

Nice, thanks for fixing this!