hishn12/Email_Classification_without_LLM

1import re2from typing import List, Dict3 4class PIIDetector:5    """Detects PII using regex only (no spaCy or LLMs)."""6 7    def __init__(self):8        self.patterns = {9            'full_name': r'\b[A-Z][a-z]+(?:\s[A-Z][a-z]+)+\b',10            'email': r'\b[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Za-z]{2,}\b',11            # 'phone_number': r'\b(?:\+91[\-\s]?|0)?[6-9]\d{9}\b',12            'dob': r'\b\d{2}[/-]\d{2}[/-]\d{4}\b',13            'aadhar_num': r'\b\d{4}[\s-]?\d{4}[\s-]?\d{4}\b',14            'credit_debit_no': r'\b(?:\d[ -]*?){13,16}\b',15            'cvv_no': r'\b\d{3}\b',  # Just 3-digit (standalone), improved16            'expiry_no': r'\b(0[1-9]|1[0-2])\/([0-9]{2,4})\b'  # MM/YY or MM/YYYY17        }18 19    def detect(self, text: str) -> Dict[str, List[str]]:20        findings = {}21 22        for pii_type, pattern in self.patterns.items():23            matches = re.findall(pattern, text)24 25            values = []26            for m in matches:27                if isinstance(m, tuple):28                    m = "/".join(m)29                if isinstance(m, str) and m.strip():  # ✅ prevent empty strings30                    values.append(m)31 32            if values:33                findings[pii_type] = values34 35        return findings36