Damn, 350 lines of code and you didn't even win in performance. My 'higher level' approach (100 lines) runs in 850ms.

Do you remember how long it took you to make this kata?

'alr approved some time ago'

Thank you for the suggestion. As previously said ideally it should test for every possible order but that would make for complicated tests.
The assignment is what it is, if instructions are followed and the derivation rules are followed in the way they are described then you will end up with a solution that passes the tests.

Not clear

Perhaps a bit late but FYI to resolve merge conflicts in the Description, one has to fork the original translation and get the fork approved instead. Hope this helps :)

In the Python test for an empty array [], the expected answer is 0 but it should be None.

This is an offical code not mine, as you can see in the comment.

Replacing array with just a number return 1 constantly!

Yeah. Thats why I raised an issue because the description says it shouldn't pass but the tests do not check for this :-)

Are there urls like www.www.etc.?

My older kata, the code length limit 170 is too tolerant at this moment. perhaps 130 will be nice ;-)
Thanks your solution. Cheers ;-)

I'm still having issues with merging your translation. Could you re-publish it? (Again...)