M-Arjun/SpamShield

1import re2import string3 4def preprocess_text(text: str) -> str:5    """6    Lightweight preprocessing:7    - lowercase8    - URL normalization9    - remove excessive repeated characters10    - strip punctuation11    """12    if not text:13        return ""14    15    # Lowercase16    text = text.lower()17    18    # URL normalization19    text = re.sub(r'https?://\S+|www\.\S+', ' [URL] ', text)20    21    # Remove excessive repeated characters (e.g., "freeeeee" -> "free")22    text = re.sub(r'(.)\1{2,}', r'\1', text)23    24    # Handle spaced out characters (e.g., "F R E E" -> "FREE")25    # Only if they are single characters separated by spaces, and more than 2 in a row26    text = re.sub(r'\b(\w\s){2,}\w\b', lambda m: m.group().replace(' ', ''), text)27    28    # Strip punctuation29    text = text.translate(str.maketrans('', '', string.punctuation))30    31    # Remove extra whitespace32    text = re.sub(r'\s+', ' ', text).strip()33    34    return text35