LGTM: various fixes for using string operator instead of numeric

[luzpaz]

No description provided.

[vocx-fc]

These changes test for None. It is fine to substitute the equality comparison == for is, however, it is a bit more Pythonic to simply test for the existence of the object. If the object is None, then it is equal to not object.

if something == None:
    proceed()

if something is None:  # better
    proceed()

if not something:  # best
    proceed()

Using object and not object tests for existence and non-existence.

if something:  # exists
    proceed()
 
if not something:  # does not exist
    proceed()

[yorikvanhavre]

For the story, the rationale I read is that something is None is faster than something == None as it compares pointers and not values

[joelgraff]

I can't help but comment on the is None vs not issue. I'm sure we're all familiar with it - that not is a test to see if something is "falsey". Thus, None, 0, False, [], {}, and '' all qualify.

The problem for me is that not is testing for two very different conditions and returning the same result. That is, 'False', 0, [], {} and '' are all about state. None is about existence. Failing to distinguish whether not something means it's is just empty / zero / false or means it doesn't exist at all usually ends up biting me.

Thus, I expressly avoid using not if I'm really testing for None. At very least, I test for it separately.

[vocx-fc]

Joel, I think in all cases mentioned above the test is for existence, which is why I think using not something is fine. There is no separate test for empty content (empty value). If this exists, proceed, if it doesn't exist, then don't proceed.