from faker import Faker import random from datetime import datetime # 初始化Faker生成器（简体中文） fake = Faker("zh_CN") # 定义固定值 TEAM_ID = 106 PLAN_ID = 131 # 常见民族列表 ETHNIC_GROUPS = [ "汉", "壮", "满", "回", "苗", "维吾尔", "土家", "彝", "蒙古", "藏", "布依", "侗", "瑶", "朝鲜", "白", "哈尼", "哈萨克", "黎", "傣", "畲", "傈僳", "仡佬", "东乡", "高山", "拉祜", "水", "佤", "纳西", "羌", "土", "仫佬", "锡伯", "柯尔克孜", "达斡尔", "景颇", "毛南", "撒拉", "布朗", "塔吉克", "阿昌", "普米", "鄂温克", "怒", "京", "基诺", "德昂", "保安", "俄罗斯", "裕固", "乌兹别克", "门巴", "鄂伦春", "独龙", "塔塔尔", "赫哲", "珞巴", ] # 疾病列表（健康状态） HEALTH_CONDITIONS = [ "无", "近视", "弱视", "鼻炎", "哮喘", "过敏体质", "蛀牙", "肥胖", "发育迟缓", "ADHD", "缺铁性贫血", "维生素缺乏", "易感冒体质", "偏瘦", "过敏(花粉)", "过敏(尘螨)", "过敏(牛奶)", "消化不良", "多动症", "抽动症", "癫痫", "小儿麻痹", "乙肝病毒携带", ] # 艺术特长 ART_TALENTS = [ "无", "钢琴", "小提琴", "绘画", "书法", "舞蹈", "声乐", "戏剧表演", "摄影", "编程", "动画制作", "模型制作", "手工艺术", ] # 体育特长 SPORTS_TALENTS = [ "无", "游泳", "篮球", "足球", "乒乓球", "羽毛球", "田径", "武术", "跆拳道", "空手道", "轮滑", "滑板", "自行车", "健美操", ] # 科技特长 TECH_TALENTS = [ "无", "机器人编程", "电子制作", "3D打印", "人工智能", "网页设计", "科学实验", "天文观测", "植物培养", "动物观察", "航模制作", ] # 职业列表 OCCUPATIONS = [ "医生", "教师", "工程师", "程序员", "设计师", "建筑师", "销售经理", "会计", "律师", "公务员", "记者", "警察", "消防员", "厨师", "商人", "自由职业", "个体经营者", "银行职员", "人力资源", "市场营销", ] # 公司类型 WORKPLACES = [ "国有企业", "民营企业", "外资企业", "合资企业", "政府机关", "事业单位", "学校", "医院", "设计事务所", "律师事务所", "科技公司", "互联网公司", "金融机构", "建筑公司", "零售企业", ] # 小学名称 PRIMARY_SCHOOLS = [ "第一实验小学", "第二实验小学", "育才小学", "阳光小学", "希望小学", "和平小学", "实验小学分校", "新世纪小学", "东方小学", "先锋小学", "星光小学", "明德小学", ] def generate_id_num(birthday): """生成符合规则的身份证号""" # 生成前6位（地区码，使用真实的地区码） region_codes = [ "110101", "110105", "110106", "110107", "310112", "330102", "440106", "440304", ] # 出生日期码（8位） birth_code = birthday.replace("-", "") # 顺序码（3位） sequence_code = str(random.randint(101, 998)) # 生成校验码（1位） base = ( region_codes[random.randint(0, len(region_codes) - 1)] + birth_code + sequence_code ) weight = [7, 9, 10, 5, 8, 4, 2, 1, 6, 3, 7, 9, 10, 5, 8, 4, 2] checksum_map = { 0: "1", 1: "0", 2: "X", 3: "9", 4: "8", 5: "7", 6: "6", 7: "5", 8: "4", 9: "3", 10: "2", } total = 0 for i in range(17): total += int(base[i]) * weight[i] checksum = checksum_map[total % 11] return base + checksum def generate_student_data(): """生成单个学生数据""" # 基础信息 sex = random.choice(["男", "女"]) # 生日 (2010-2015年出生) birthday = fake.date_between(start_date="-15y", end_date="-8y").strftime("%Y-%m-%d") # 生成符合规则的身份证号 id_num = generate_id_num(birthday) # 计算年龄（仅用于展示） birth_year = int(birthday[:4]) current_year = datetime.now().year age = current_year - birth_year # 生成学生信息 data = { "teamId": TEAM_ID, "stuName": fake.name(), # 学生姓名 "idNum": id_num, # 身份证号 "formerName": random.choices( ["无", fake.last_name() + fake.first_name()], weights=[0.8, 0.2] )[ 0 ], # 曾用名 "sex": sex, # 性别 "nation": random.choice(ETHNIC_GROUPS), # 民族 "birthday": birthday, # 生日 "finishSchool": random.choice(PRIMARY_SCHOOLS), # 毕业学校 "file": None, # 文件 "hkadr": fake.province() + "/" + fake.city() + "/" + fake.district(), # 户口所在地 "homeAddress": fake.address(), # 家庭住址 "health": random.choice(HEALTH_CONDITIONS), # 健康状况 "artisticSpecialty": random.choice(ART_TALENTS), # 艺术特长 "artisticAchievements": "无", # 艺术成就 "technologicalSpecialty": random.choice(TECH_TALENTS), # 科技特长 "technologicalAchievements": "无", # 科技成就 "sportsSpecialty": random.choice(SPORTS_TALENTS), # 体育特长 "sportsAchievements": "无", # 体育成就 "planId": PLAN_ID, # 计划ID "imageUrl": "", # 图片URL "schoolCode": "G" + id_num[:-1], # 学籍号 "age": age, # 年龄（仅用于展示） "email": fake.email() if random.random() > 0.7 else "", # 邮箱 "height": str(random.randint(130, 170)), # 身高 "weight": str(random.randint(25, 60)), # 体重 "measurements": f"{random.randint(60, 90)},{random.randint(60, 90)},{random.randint(60, 90)}", # 三围 } # 生成监护人信息 guardian_prefix = fake.last_name() # 监护人1 - 通常为父母之一 relationship_choices = ["父亲", "母亲", "爷爷", "奶奶", "外公", "外婆"] relationship = random.choice(relationship_choices[:2]) # 主要选择父母 data.update( { "guardian1Name": guardian_prefix + ("先生" if relationship == "父亲" else "女士"), # 监护人1姓名 "guardian1Relationship": relationship, # 关系 "guardian1IdCardNumber": generate_id_num( fake.date_between(start_date="-55y", end_date="-25y").strftime( "%Y-%m-%d" ) ), # 身份证号 "guardian1Contact": fake.phone_number(), # 联系方式 "guardian1Occupation": random.choice(OCCUPATIONS), # 职业 "guardian1Workplace": random.choice(WORKPLACES), # 工作单位 } ) # 监护人2 - 可能是另一位家长或其他人 if relationship == "父亲": guardian2_relationship = "母亲" else: guardian2_relationship = random.choices( ["父亲", "爷爷", "奶奶", "外公", "外婆", "其他亲属"], weights=[0.7, 0.1, 0.1, 0.05, 0.05, 0.1], )[0] data.update( { "guardian2Name": guardian_prefix + ( "先生" if guardian2_relationship in ["父亲", "爷爷", "外公"] else "女士" ), # 监护人2姓名 "guardian2Relationship": guardian2_relationship, # 关系 "guardian2IdCardNumber": generate_id_num( fake.date_between(start_date="-55y", end_date="-25y").strftime( "%Y-%m-%d" ) ), # 身份证号 "guardian2Contact": ( fake.phone_number() if guardian2_relationship != "无" else "" ), # 联系方式 "guardian2Occupation": ( random.choice(OCCUPATIONS) if guardian2_relationship != "无" else "" ), # 职业 "guardian2Workplace": ( random.choice(WORKPLACES) if guardian2_relationship != "无" else "" ), # 工作单位 } ) return data def generate_students(): student = generate_student_data() return student generate_students()