Email Normalizer

#!/usr/bin/env python3
"""
Email Normalizer
Standardize email addresses across your lists
"""
import re

def normalize_email(email):
    """Normalize a single email address"""
    email = email.strip().lower()

    # Split local and domain parts
    if '@' not in email:
        return None

    local, domain = email.rsplit('@', 1)

    # Gmail-specific: remove dots and plus addressing
    if domain in ['gmail.com', 'googlemail.com']:
        local = local.replace('.', '')
        if '+' in local:
            local = local.split('+')[0]
        domain = 'gmail.com'  # Normalize googlemail

    # Common domain fixes
    domain_fixes = {
        'gmial.com': 'gmail.com',
        'gmai.com': 'gmail.com',
        'hotmal.com': 'hotmail.com',
        'yaho.com': 'yahoo.com',
    }
    domain = domain_fixes.get(domain, domain)

    # Remove leading/trailing spaces and dots
    local = local.strip('.')
    domain = domain.strip('.')

    return f"{local}@{domain}"

def normalize_list(emails):
    """Normalize a list of emails"""
    normalized = []
    for email in emails:
        result = normalize_email(email)
        if result:
            normalized.append(result)
    return list(set(normalized))  # Remove duplicates

if __name__ == '__main__':
    import sys
    emails = open(sys.argv[1]).read().splitlines()
    normalized = normalize_list(emails)
    print(f"Original: {len(emails)}, Normalized: {len(normalized)}")