python代码实现数据脱敏技术的完整指南

import re

def mask_phone(phone: str) -> str:
    """手机号脱敏：保留前3位和后4位，中间用*替换"""
    if not re.match(r'^1[3-9]\d{9}$', phone):  # 简单校验手机号格式
        raise ValueError("手机号格式错误")
    return re.sub(r'(\d{3})\d{4}(\d{4})', r'\1****\2', phone)

# 测试
print(mask_phone("13812345678"))  # 输出：138****5678

def mask_id_card(id_card: str) -> str:
    """身份证号脱敏：保留前6位和后4位，中间用*替换"""
    if not re.match(r'^\d{17}[\dXx]$', id_card):  # 简单校验身份证格式
        raise ValueError("身份证号格式错误")
    return re.sub(r'^(\d{6})\d+(\d{4})$', r'\1********\2', id_card)

# 测试
print(mask_id_card("110101199001011234"))  # 输出：110101********1234

def mask_email(email: str) -> str:
    """邮箱脱敏：保留首位字符和域名，中间用*替换"""
    if not re.match(r'^[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}$', email):
        raise ValueError("邮箱格式错误")
    name, domain = email.split('@')
    if len(name) == 1:  # 若用户名只有1位，直接保留
        return f"{name}@****.{domain.split('.')[-1]}"
    return f"{name[0]}****@{domain}"

# 测试
print(mask_email("abcdef@example.com"))  # 输出：a****@example.com
print(mask_email("a@test.com"))  # 输出：a@****.com

def mask_address(address: str) -> str:
    """地址脱敏：保留省/市，隐藏详细地址（按第一个“区/县”后的内容替换）"""
    # 简单规则：保留到第一个“区”“县”“市”后的内容
    pattern = re.compile(r'^(.+?[市区县])(.*)$')
    match = pattern.match(address)
    if match:
        return f"{match.group(1)}***"
    return "***"  # 若无法匹配，返回完全脱敏

# 测试
print(mask_address("北京市海淀区中关村大街1号"))  # 输出：北京市海淀区***
print(mask_address("上海市浦东新区张江高科技园区"))  # 输出：上海市浦东新区***

import random

# 常见姓氏（前100位）和名字常用字
FAMILY_NAMES = ["王", "李", "张", "刘", "陈", "杨", "赵", "黄", "周", "吴", 
                "徐", "孙", "胡", "朱", "高", "林", "何", "郭", "马", "罗"]
GIVEN_NAMES = ["伟", "芳", "娜", "秀英", "敏", "静", "强", "磊", "军", "洋", 
               "勇", "艳", "杰", "娟", "丽", "涛", "明", "超", "秀兰", "霞"]

def generate_chinese_name() -> str:
    """随机生成中文姓名（姓氏+1-2字名字）"""
    family_name = random.choice(FAMILY_NAMES)
    given_name = random.choice(GIVEN_NAMES)
    # 随机选择1字或2字名字（从GIVEN_NAMES中取前1位或全部）
    given_name = given_name[0] if random.random() < 0.3 else given_name
    return family_name + given_name

# 测试
for _ in range(5):
    print(generate_chinese_name())  # 可能输出：王磊、李娜、张伟、刘敏、陈杰

from datetime import datetime, timedelta

def generate_fake_id_card() -> str:
    """生成假身份证号（格式正确，地区码和出生日期随机）"""
    # 随机地区码（前6位，这里用常见地区码示例）
    area_codes = ["110101", "110102", "310101", "310104", "440106", "440305"]
    area_code = random.choice(area_codes)
    
    # 随机出生日期（1950-2000年）
    start_date = datetime(1950, 1, 1)
    end_date = datetime(2000, 12, 31)
    delta = end_date - start_date
    random_days = random.randint(0, delta.days)
    birth_date = start_date + timedelta(days=random_days)
    birth_str = birth_date.strftime("%Y%m%d")
    
    # 顺序码（3位，奇数为男，偶数为女，这里随机）
    sequence_code = f"{random.randint(0, 999):03d}"
    
    # 校验码（简化处理，随机数字或X）
    check_code = random.choice("0123456789X")
    
    return area_code + birth_str + sequence_code + check_code

# 测试
for _ in range(3):
    print(generate_fake_id_card())  # 可能输出：11010119850312456X、310101199211237890

def generate_fake_email() -> str:
    """生成随机邮箱"""
    # 用户名：6-12位字母+数字
    username_length = random.randint(6, 12)
    username = ''.join(random.choices("abcdefghijklmnopqrstuvwxyz0123456789", k=username_length))
    # 常见域名
    domains = ["gmail.com", "qq.com", "163.com", "outlook.com", "yahoo.com"]
    domain = random.choice(domains)
    return f"{username}@{domain}"

# 测试
for _ in range(3):
    print(generate_fake_email())  # 可能输出：x7y3z9w@qq.com、abc123def@163.com

def generate_fake_ipv4() -> str:
    """生成随机IPv4地址"""
    ip_segments = [str(random.randint(0, 255)) for _ in range(4)]
    return ".".join(ip_segments)

# 测试
for _ in range(3):
    print(generate_fake_ipv4())  # 可能输出：192.168.1.100、10.0.0.1、255.255.255.255

class DataMasker:
    @staticmethod
    def mask_phone(phone): ...  # 手机号脱敏
    @staticmethod
    def mask_id_card(id_card): ...  # 身份证脱敏
    @staticmethod
    def generate_chinese_name(): ...  # 生成姓名
    # 其他方法...