pub fn tokenize(text: &str) -> Vec<String>
The terms of text: NFKC-normalised, lowercased runs of letters and digits of two characters or more, stop words dropped, stemmed.
text