DOC: Clarify that np.char comparison functions always return ndarrays

[BvB93]

The docstring of np.char.equal(), and all other comparison functions in the np.char module, incorrectly claim that a boolean is returned if two scalars are compared. This is incorrect, as the respective function(s) always return an ndarray (0d in the case of scalar comparisons).

This pull requests thus corrects the docstrings of aforementioned functions.

Examples

In [1]: import numpy as np

In [2]: np.__version__
Out[2]: '1.20.0.dev0+961b56f'

In [3]: np.char.equal('a', 'a')
Out[3]: array(True)  # That's definitely not a bool

In [4]: np.char.equal(['a', 'b'], ['a', 'b'])
Out[4]: array([ True,  True])

[rossbar]

LGTM - the current docs do seem to be incorrect.

[mattip]

Thanks @BvB93