敏感信息脱敏工具类

desensitize.py

import re

def desensitize_text(input_text):

    # 替换 手机/证件号/车架号（10到20位大小写字母+数字）
    id_pattern = re.compile(r'([a-zA-Z0-9]{2})[a-zA-Z0-9]*([a-zA-Z0-9]{2})')
    input_text = re.sub(id_pattern, r'\1****\2', input_text)

    # 地址脱敏, 文本包含以下任意三个, 就对这几个字前面两个字符用*代替
    addr_words = ["省","市","区","县", "镇","街", "号",  "座", "楼","路"]
    if sum(input_text.count(word) for word in addr_words) >= 3:
        for word in addr_words:
            addr_pattern = re.compile(r'(?<!\*\*)\w{2,}'+word)
            matches = addr_pattern.search(input_text)
            while None!=matches:
                input_text = input_text[:matches.start()]+'**' + word +input_text[matches.end():]
                matches = addr_pattern.search(input_text)
    return input_text

# from faker import Faker
# fake = Faker('zh_CN')
# print(desensitize_text("我的手机号是 13812345678，身份证号是 320123198012345678，驾驶证号是 AB1234567890cdEF12345。"))
# print(desensitize_text("我的车架号:LJ1E6A2D1AG085640, 车辆订单号:8100045677713437, 车牌号是沪ABB2271"))
# print(desensitize_text("安徽省合肥市蜀山区秀康达城锦园 1 号楼101"))
# for i in range(5):
#     print(desensitize_text(fake.address()))