'Cause PEP-8 never goes out of style

[alex-rakowski]

pep 8 - Comparisons to singletons compliance

# goes from 
if arg == True:
    pass
if False == arg:
    pass
# to
if arg is True:
    pass
if arg is False:
    pass

[sezelt]

For builtin Python booleans, testing == False and is False have the same behavior, but for other types this is not the case. In particular for numpy arrays, these do entirely different things: is compares the identity of the array object itself, while == is an elementwise value comparison. For example:

In [1]: import numpy as np

In [2]: ar = np.random.rand(3,3)

In [3]: ar
Out[3]:
array([[0.12607551, 0.56005212, 0.42254926],
       [0.48698435, 0.5986023 , 0.75294198],
       [0.82027795, 0.58230241, 0.92049683]])

In [4]: b = ar > 0.5

In [5]: b
Out[5]:
array([[False,  True, False],
       [False,  True,  True],
       [ True,  True,  True]])

In [6]: b is False
Out[6]: False

In [7]: b == False
Out[7]:
array([[ True, False,  True],
       [ True, False, False],
       [False, False, False]])

There is also one place where I can't tell the original meaning of the code, but from context it looks like it should actually be an assignment (=) rather than an equality test (see review comment).

[alex-rakowski]

Thanks @sezelt

[sezelt]

Thanks for the changes. I think we are just waiting on clarification of that one line whose original meaning was unclear

[sezelt]

Just needs clarification on the line in fit.py

[bsavitzky]

So this is just changing every boolean == to an is? Not sure I'm into it :p

[sezelt]

From PEP-8:

Comparisons to singletons like None should always be done with is or is not, never the equality operators.

See also this Stackoverflow:
https://stackoverflow.com/questions/27276610/boolean-identity-true-vs-is-true

This thread actually shows an interesting example where a numpy function that returns a scalar boolean does not return a Python singleton, and thus would behave differently under this change.

[alex-rakowski]

It changes == into is for comparisons to singletons, which is the pep-8 style.

Comparisons to singletons like None should always be done with is or is not, never the equality operators.
Also, beware of writing if x when you really mean if x is not None – e.g. when testing whether a variable or argument that defaults to None was set to some other value. The other value might have a type (such as a container) that could be false in a boolean context!

@sezelt caught the examples I incorrectly changed, which were used in numpy masks.

[alex-rakowski]

@sezelt @bsavitzky Changed to assignment should be good to go