Use int instead of numbers.integral in docstring

[NicolasHug]

In general we want to avoid having type-annotation-like types within the docstring. Type annotations are meant to be read by a machine, not by humans, and over time they will clutter the docs.

In this instance, int is a more common and familiar term than numbers.Integral, and will help keep the doc more user-friendly. I don't believe there is any loss of accuracy keeping int in the docstring.

[ahmadsharif1]

Does int convey the message to the user that numpy integers are okay to pass? I believe that was @scotts intent here.

[NicolasHug]

I would argue that:

• We don't need to convey that point
• numbers.Integral doesn't convey that point either.

I really think numpy integers are an edge case that isn't in the mind of users. I've been using numpy for 10 years and I still don't know what these numpy types are. When int is the documented type, users expect duck typing to prevail, and expect int-like objects to work.

[scotts]

Deferring to @NicolasHug judgement here. Looks good to me!