You're welcome, Marc!

Do you know how lang_detect generates the probability?

it computes probabilities from features of spelling by leveraging Naive Bayes with character-based n-gram.

You can have an overview of the algorithm from page n°18 of the following document

https://www.slideshare.net/shuyo/language-detection-library-for-java

I hope this helps.

Indeed, exploration is the best teacher 👨‍🏫