Predicate functions shouldn't convert its implementation based on
logic expressions (combinations of `not`, `and` and `or`).

I.e. The following should be correct:

    def is_NOT_divisible_by_4(n):
        return n % 4

    assert is_NOT_divisible_by_4(10)

But as the TestCases for this exercise are using `assertIs(..., True)
instead of `assertTrue(...)`, implementations cointing on the numbers
behaviour on boolean contexts will Fail.

What this mean, is that for the given definition of
`is_NOT_divisible_by_4`, the folloging assertion will Fail:

    assert is_NOT_divisible_by_4(10) == True

I think it is meaningless to check for `True` as it is naive to write:

    if predicate(x) == True:
        do_something(x)

Instead of:

    if predicate(x):
        do_something(x)