from sus.clean import clean
from sus.lemmatize_text import lemmatize_text
import re


def preprocess(text: str) -> str:
    cleaned = clean(text, convert_to_ascii=True)
    lemmas = [
        re.sub(r'\d+', 'NUM', lemma)
        for lemma in lemmatize_text(cleaned)
    ]
    return " ".join(lemmas)