Just now I released version 0.1.0 of LanguageIdent.
I renamed the langPredict-NVDA-Addon to LanguageIdent. I think its a better fit. I also switched the AI-model from fast-text-langident to langid-python-module. It’s easier to use and has the great feature, that allows to limit to identified languages.
In this release aside the new name the main feature for the user is a whitelist. You can specify, which languages are considered for language identification. For example if I specify “de, en” and the model would think Italian “it” is the best match, it still will return either “de” or “en”.
I’ll focus on documentation and quality next. I’m looking for a 1.0 release this year. I’m also considering supporting Intel 64Bit and ARM 64 Bit. I have a Macbook Air M1 and plan to try running Windows ARM 64.