3 years has passed, still no random test :/

Python translation. Contains 150 random tests and uses arrays instead of strings. Please let me know if I missed anything

Log:
Straße Straße, but not Straße-Straße
Expected: '['Strasse Strasse']', instead got: '['Straße Straße']'
Where does "Strasse" come from?

Dang! I'd pasted about five in the Example Test Cases, but must have exited the browser too fast. Will fix that later tonight/tomorrow a.m. Thanks for pointing out.

I think FArekkusu meant the tests that the users see ("Example Test Cases" in the kata editor).

Thanks. You mean no random tests, perhaps? There are about 10 test cases that pass.

Will add some random tests, thanks for pointing this out. Haven't used the kata editor in a long while, so I just followed the 'inserted' example Codewars provide, there didn't seem to be a link to more detailed documentation.

Please use appropriate data types. The return value should not be a string; it should be an array of numbers.

No random tests.

No sample tests.

Hi there and thanks for the comment.

An updated reply:
I had a look at what the user sees and, indeed, the text containing the duplicate words had not been saved, so there was nothing for people to work on. I've pasted it again and I hope that it's now been saved. I'd like to soft-wrap the sample text, but I'm not sure how to do this on Codewars, so I'll place it in the description too for people to read more easily.

If you do take another look, please let me know whether the edit is visible and, if the kata is now OK, whether 6 kyu is about right or not.

In real lifethere may be such problems indeed; I've read a couple of interesting articles on the shortcomings of JavaScript when it comes to unicode while researching my kata. However, testing my solution on a few online testers outside Codewars returned the correct results, exactly as expected.

At the moment I'm more worried that I may have underestimated the level of difficulty the kata presents. As you're right there at the top, with 1 kyu, you might be better placed to judge that.So if you and other experienced Codewars members think that the kata should be in the 5 kyu category, I'd change the classification.

There are no words starting with ẞ, but words may be in all caps, although for searching in ordinary texts it shouldn't really matter.
For real applications there may be a different problem: Unicode composed vs decomposed characters, like "ü" vs "ü".