Duplicate.

I was baffled by this signature of the translate method until I realized "Python" here means Python 2.

Most readable of the one-liner solutions I see here.

Fair enough. I love Python, my career revolves around it, but sometimes I just want a little more rigor in the type system! Props to Python 3 for disallowing ordinal comparisons to None.

That's because of Python's duck typing – bool is a subclass of int.

print(False == 0 and True == 1) # Prints "True"

And since the description asks to return 0 when the numbers are equal and False is equivalent to 0, this is acceptable.

well, that'd be pretty anti-pythonic, tho... ;)

I'm guessing by "false extending 0" you mean to point out that False == 0? Sure, but take note also that False is not 0.

My point was just that the description (i.e. acceptance criteria) states "when a is close to b, return 0" (type int) while this returns False (type bool). Python's weak typing in this area makes it easy to be casual. But if you added type annotation to this function and enforced it (say, with mypy or enforce), returning False would be an error.

but false extending 0, if the expected answer is 0, that's still correct.

proof: this passes this fixed test: test.assert_equals(close_compare(5, 5), 0)

This returns False if a and b are equal. (I posted a comment in discourse also, since this suggests the tests need tightening.)

That's not the problem though; the main problem is that it's not even half-finished (just look at the tests and import statements). Fixing it would require pretty much writing up the whole test.

It's on the list: https://github.com/Codewars/codewars.com/wiki/List-of-Affected-Kata The question is whether the packages will be installed or not.

Clojure: it's unfinished and should be revoked/removed.

Tests require result to have a specific ordering but it's not mentioned anywhere. Either it should be specified, or the test should sort the user result before comparison, or the function should return a set instead.

@pbx,

Instead of just showing it in the description (and then forcing everyone to write their own), it might be a good idea to add the digits dictionary to Preloaded since every solution has the same dictionary. That way people could choose whether to use a preloaded dict of digit/letter pairs or to write their own (I'm guessing many people would use the preloaded dict).

LETTERS = {
    '2': 'ABC',
    '3': 'DEF',
    '4': 'GHI',
    '5': 'JKL',
    '6': 'MNO',
    '7': 'PQRS',
    '8': 'TUV',
    '9': 'WXYZ'
}

If you do add that to Preloaded, just make sure to tell people about the name of it in the description too. You could use separate markdown code sections for each language (so the Python dict only shows up in the description for Python not Clojure and vice versa).

Thanks!