Files
atc-agents/deploy/airflow/scripts/generate_mysql_employee_data.py
mo 944635ffe1 feat(masking): add PII columns to generators + backfill (Faker-style, no deps)
Source generators (postgres/mysql/mongodb) now emit realistic PII
(name/email/phone/ip/iban/address/national_id/dob). deploy/mask_pii_setup.py
ALTERs the source tables and backfills a bounded sample so Debezium CDC and
OpenMetadata PII auto-classification see real sensitive values.
2026-06-27 02:50:28 +02:00

104 lines
4.1 KiB
Python

#!/usr/bin/env python3
"""
Generate fake employee event data for MySQL.
Light + configurable via GEN_ROWS. Now also populates realistic PII columns
(employee_name/email/phone, national_id, home_address, date_of_birth).
"""
import mysql.connector
import random
from datetime import datetime, timedelta
import uuid
import os
DB_HOST = "10.0.21.51"
DB_PORT = "3306"
DB_NAME = "hr"
DB_USER = "mo"
DB_PASSWORD = "Dell2026!"
TARGET_ROWS = int(os.getenv("GEN_ROWS", "5000"))
BATCH_SIZE = min(10000, max(500, TARGET_ROWS))
DEPARTMENTS = ["HR", "Operations", "Sales", "Marketing", "Finance", "IT", "Engineering", "Legal"]
ROLE_NAMES = ["Analyst", "Lead", "Manager", "Consultant", "Director", "Engineer", "Specialist", "Coordinator"]
REGIONS = ["EU", "APAC", "LATAM", "NA", "EMEA"]
EVENT_TYPES = ["TRANSFER", "PROMOTION", "TERMINATION", "HIRED", "SALARY_CHANGE", "DEPARTMENT_CHANGE"]
FIRST = ["Sophie", "Liam", "Emma", "Noah", "Julia", "Lucas", "Mila", "Daan", "Anna", "Sem",
"Eva", "Finn", "Tess", "Bram", "Lotte", "Max", "Sara", "Thijs", "Nina", "Ruben"]
LAST = ["de Vries", "Jansen", "Bakker", "Visser", "Smit", "Meijer", "Mulder", "Bos",
"Vos", "Peters", "Hendriks", "Dijkstra", "Kok", "Willems", "Maas", "Koning"]
STREETS = ["Kerkstraat", "Dorpsweg", "Molenpad", "Stationsplein", "Industrieweg", "Lindelaan",
"Beukenhof", "Havenstraat", "Parkweg", "Schoolstraat"]
CITIES = ["Amsterdam", "Rotterdam", "Utrecht", "Eindhoven", "Den Haag", "Groningen", "Tilburg", "Almere"]
DOMAINS = ["example.com", "mail.nl", "acme-corp.eu", "company.eu", "contact.io"]
def _name():
return f"{random.choice(FIRST)} {random.choice(LAST)}"
def _pii():
nm = _name()
return (
nm,
f"{nm.lower().replace(' ', '.')}{random.randint(1, 999)}@{random.choice(DOMAINS)}",
f"+31 6 {random.randint(10, 99)} {random.randint(100000, 999999)}",
str(random.randint(100000000, 999999999)),
f"{random.choice(STREETS)} {random.randint(1, 320)}, {random.randint(1000, 9999)} {random.choice(CITIES)}",
f"{random.randint(1960, 2002)}-{random.randint(1, 12):02d}-{random.randint(1, 28):02d}",
)
def generate_fake_employee_event():
employee_id = random.randint(1, 100000)
department = random.choice(DEPARTMENTS)
role_name = random.choice(ROLE_NAMES)
region = random.choice(REGIONS)
event_type = random.choice(EVENT_TYPES)
days_ago = random.randint(0, 730)
event_ts = datetime.now() - timedelta(days=days_ago, hours=random.randint(0, 23),
minutes=random.randint(0, 59))
salary_change = round(random.uniform(1000.0, 20000.0), 2) if random.random() > 0.3 else None
notes = str(uuid.uuid4()) * 10
name, email, phone, nid, address, dob = _pii()
return (employee_id, department, role_name, region, event_type, salary_change, event_ts, notes,
name, email, phone, nid, address, dob)
INSERT_SQL = """
INSERT INTO employee_events (employee_id, department, role_name, region,
event_type, salary_change, event_ts, notes,
employee_name, employee_email, employee_phone,
national_id, home_address, date_of_birth)
VALUES (%s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s)
"""
def main():
print(f"Connecting to MySQL at {DB_HOST}:{DB_PORT}...")
conn = mysql.connector.connect(host=DB_HOST, port=DB_PORT, database=DB_NAME, user=DB_USER, password=DB_PASSWORD)
cursor = conn.cursor()
print(f"Generating {TARGET_ROWS} employee events (with PII)... batch={BATCH_SIZE}")
total_generated = 0
batch = []
for i in range(TARGET_ROWS):
batch.append(generate_fake_employee_event())
if len(batch) >= BATCH_SIZE:
cursor.executemany(INSERT_SQL, batch)
conn.commit()
total_generated += len(batch)
batch = []
if batch:
cursor.executemany(INSERT_SQL, batch)
conn.commit()
total_generated += len(batch)
cursor.close()
conn.close()
print(f"Completed! Generated {total_generated} employee events.")
if __name__ == "__main__":
main()