indian-txn-classifier / code /pii_shield.py
SahilGoel's picture
Upload code/pii_shield.py with huggingface_hub
8e34db8 verified
Raw
History Blame Contribute Delete
2.67 kB
"""
pii_shield.py — Lightweight PII Masking
=======================================
Masks sensitive data before logging or sending to external services.
Pure regex, zero dependencies — works in 512MB environments.
Detects & masks:
- PAN numbers (ABCDE1234F pattern)
- Aadhaar numbers (12 digits)
- Bank account numbers (9-18 digits near keywords)
- IFSC codes
- Mobile numbers
- Email addresses
"""
import re
import logging
logger = logging.getLogger(__name__)
PII_PATTERNS = [
{
"name": "PAN",
"pattern": r'\b([A-Z]{5}\d{4}[A-Z])\b',
"mask": "XXXXX0000X",
},
{
"name": "Aadhaar",
"pattern": r'\b(\d{4}[\s\-]?\d{4}[\s\-]?\d{4})\b',
"mask": "XXXX-XXXX-XXXX",
},
{
"name": "Mobile",
"pattern": r'(?:\+91[\s\-]?)?\b([6-9]\d{9})\b',
"mask": "XXXXXXXXXX",
},
{
"name": "Email",
"pattern": r'\b([a-zA-Z0-9._%+\-]+@[a-zA-Z0-9.\-]+\.[a-zA-Z]{2,})\b',
"mask": "***@***.***",
},
{
"name": "IFSC",
"pattern": r'\b([A-Z]{4}0[A-Z0-9]{6})\b',
"mask": "XXXX0XXXXXX",
},
{
"name": "BankAccount",
"pattern": r'(?:account[\s\-:]*(?:no|number|num)?[\s\-:]*)\b(\d{9,18})\b',
"mask": "XXXXXXXXXX",
"case_insensitive": True,
},
]
def mask_pii(text: str) -> str:
"""Mask all PII patterns in the given text. Returns sanitized text."""
masked = text
for p in PII_PATTERNS:
flags = re.IGNORECASE if p.get("case_insensitive") else 0
pattern = p["pattern"]
mask = p["mask"]
count = 0
def replace(m):
nonlocal count
count += 1
return mask
masked = re.sub(pattern, replace, masked, flags=flags)
if count > 0:
logger.debug(f"PII Shield: masked {count} {p['name']}(s)")
return masked
def has_pii(text: str) -> bool:
"""Check if text contains any PII. Returns True if sensitive data detected."""
for p in PII_PATTERNS:
flags = re.IGNORECASE if p.get("case_insensitive") else 0
if re.search(p["pattern"], text, flags):
return True
return False
def shield_transaction(description: str) -> str:
"""Shield a single transaction description — masks PAN/Aadhaar but preserves merchant info."""
# Only mask PAN and Aadhaar — keep bank account numbers visible for classification
for pattern_name in ["PAN", "Aadhaar"]:
for p in PII_PATTERNS:
if p["name"] == pattern_name:
description = re.sub(p["pattern"], p["mask"], description)
return description