I'm using several large List Entities in Microsoft LUIS. To import these lists in different LUIS apps, and also for a more convenient workflow, I want to import these lists as JSON files. In general, this works but for some lists I get the following error message:
Unfortunately, this error message doesn't provide any clue to problematic entries. So, I tried to make smaller lists and eventually found the problem. I had these two items in the list: "Athena" and "Aþena". According to wikipedia, þ is an Icelandic letter which is pronouced similarly to th, and was therefore often replaced in the past. However, those two words lead to a 409 error if entered manually.
Of course, I can just remove one of the entries, and also handle the þ letter in a script, but I would like to know if there is a list of letters that are treated in the same way, which may lead to the same issue. I didn't find anything in the documentation. If you deal with many large entity lists that may grow in the future, it is not very practical to track down those cases individually.
Edit Also I found that the letter æ is interpreted as ae.