Closed ankailou closed 4 years ago
Hi @ankailou, thank you for letting me know about this problem and sorry for my late response. I confirm that this is an issue with the rule-based detection algorithm. I will work on a fix as soon as possible.
I reopen this one because the fix I provided in the past is more or less an ugly hack. I'll have to find a more sophisticated solution to this kind of problem.
When I build a detector from
.fromAllBuiltInSpokenLanguages()
, it detects the follow text as Greek instead of English:For reference, this text was extracted from an academic paper with some greek letters in the formulas, but not enough to merit this text to be classified as greek.