numpy.bool_ is a type, not the same thing as Python booleans:

>>> np.bool_
<type 'numpy.bool_'>

An instance of the type then cannot be the same object as True

The pep recommendation doesn't mean that nothing is allowed to compare ==
to singletons, just that if you're checking for the literal singleton True
object you should use "is" to do that. And np.bool_(True) is a different
object than True -- it has to be, it's a different class! -- so "is" will
return False no matter what pep 8 says.

See also: 1 == True, 1 is True
On Jun 5, 2015 12:53 AM, "Ludwig Schmidt-Hackenberg" <
notifications@github.com> wrote:

Even though numpy.bool_(True) == True evaluates to True. Doesn't this
violate the PEP8
http://legacy.python.org/dev/peps/pep-0008/#other-recommendations
recommendation that `Comparisons to singletons like None should always be
done with is or is not, never the equality operators.?

Sorry, if this is a dupe. I searched for bool comarison and pep8 and could
not find an relevant issue.
numpy.version: 1.9.0.dev-297f54b

—
Reply to this email directly or view it on GitHub
#5942.

Sorry for the bad example, better would have been:
print numpy.ones((1), dtype=bool)[0] --> True
print numpy.ones((1), dtype=bool)[0] is True --> False
print numpy.ones((1), dtype=bool)[0] == True --> True

Thanks for the explanation.

Ah, yes, it doesn't help that the numpy types like np.bool_ often disguise
themselves as being python built-in types... Which is unfortunate. Arguably
it should repr as "np.bool_(True)". I guess you could open a bug or
discussion about that on the mailing list if you want.
On Jun 5, 2015 1:46 AM, "Ludwig Schmidt-Hackenberg" <
notifications@github.com> wrote:

Sorry for the bad example, better would have been:
print numpy.ones((1), dtype=bool)[0] --> True
print numpy.ones((1), dtype=bool)[0] is True --> False
print numpy.ones((1), dtype=bool)[0] == True --> True

Thanks for the explanation.

—
Reply to this email directly or view it on GitHub
#5942 (comment).

Actually, the PEP8 recommendation for True and False is that you don't compare them against anything, not with ==, and especially not with is, but use them directly. The following is the relevant paragraph:

Don't compare boolean values to True or False using == .

Yes: if greeting:
No: if greeting == True:
Worse: if greeting is True:

Changing the repr sounds sensible except that doctests may be an problem...