Superceded by latest fork :(

Approved

Use Python new test framework and increase number of random test from 1 to 100.

https://www.codewars.com/kumite/633edfecf98a8200158803d0?sel=633edfecf98a8200158803d0

That's still an issue, tests should be coherent between languages. Moreover positive number is ambiguous. It should be stated as strictly positive. And we don't know if we get numbers with leading zeros and if they need to be considered as a correct ones.

I'm not sure what you mean, it has probably been fixed since then.

str should not be used as argument name in Python.

.

I think job is done

If you copy and paste the block of code I suggested into the end of the Test Cases (not the Sample Test Cases), that should be enough.

I dont know how to ... i'm new to translating kata (actually first)..can you help me out??

You might want to include a randomized test similar to:

from random import choice,randint
import string

WEIGHTS = ((1, 15), (2, 16), (3, 17), (4, 18), (5, 19), (6, 20))
def weight():
    n = randint(1, 20)
    return next(w[0] for w in WEIGHTS if n <= w[1])

def encode_block(b):
    l = len(b)
    return b if l == 1 else "\\" + str(l) + b

test_blocks = [''.join(choice(string.ascii_lowercase) for _ in range(weight())) for _ in range(randint(200, 300))]
test_string = ''.join([encode_block(block) for block in test_blocks])

Test.assert_equals(decode(test_string), test_blocks);

I weighted the distribution of characters, but you could probably simplify that.

Python translation

Ah! I did change it, but it got reverted when the Julia translation was forked for merge conflict. OK, changing the description again.

Thanks for the heads-up. See above.

That may be a description issue really; it could say "backslash followed by a positive number". The JavaScript ( original ) reference solution only recognises numbers that don't start with, or are, 0, and no testing at all is done with such numbers. I did the Haskell reference solution according to "positive", I did not modify the description ( no I don't know why I didn't .. ), but I did include fixed and possibly random tests for this behaviour.