Closed kostyfisik closed 6 years ago
We use https://github.com/optimaize/language-detector in our core and it - like most ngram-based approaches - doesn't work well with short texts.
The problem seems to have disappeared
Yes, it should be much better now thanks to: https://forum.languagetool.org/t/improved-language-detection/3250
Inserting "Масло масляное" in Russian lead to detection of English, which is wrong. In the meanwhile the russian error is correctly detected.