Search term to key in lexical model: not working both ways by default

The removing of diacritics etc. to a normalized (regularized) string doesn’t appear to work both ways by default, on predictive text. Specifically, it didn’t work on my input side. Some vowels in my keyboard language have accents as tone marks. So when I typed words without tone marks, words with tone marks from my word list were suggested too, which is correct.
But on the other hand, when I purposefully typed words with tone marks, nothing was found at all, indicating, that the input wasn’t being normalized.
I had to add the search function from https://help.keyman.com/developer/13.0/guides/lexical-models/advanced/search-term-to-key manually to the model.ts file and after that it now works well.

Hi @Friedo, apologies for the long delay in replying to this. We now have a fix for this issue ready to publish in the next 14.0 alpha release of Keyman and are in the process of pushing the fix back to version 13.0.

1 Like

Hi Marc,
No worries. Thanks for fixing this!

1 Like

You’re welcome. Actually, @eddieantonio did the work here so all thanks to him :slight_smile:

1 Like