Postgres Seeding: Manual Scripts vs BugiaData

# seed_manual.py (excerpt from a 200+ line script)
import random
from faker import Faker
from sqlalchemy import text

fake = Faker("en_US")

def seed_users(conn, count):
    user_ids = []
    for _ in range(count):
        row = conn.execute(
            text("""
                INSERT INTO users (id, name, email, created_at)
                VALUES (gen_random_uuid(), :name, :email, NOW())
                RETURNING id
            """),
            {"name": fake.name(), "email": fake.unique.email()},
        ).fetchone()
        user_ids.append(str(row[0]))
    return user_ids

def seed_orders(conn, user_ids, count):
    order_ids = []
    for _ in range(count):
        uid = random.choice(user_ids) if user_ids else None
        if uid is None:
            # often missed edge case in pipelines
            raise RuntimeError("No users available for orders.user_id FK")
        row = conn.execute(
            text("""
                INSERT INTO orders (id, user_id, status, total_cents, created_at)
                VALUES (gen_random_uuid(), :user_id, :status, :total, NOW())
                RETURNING id
            """),
            {
                "user_id": uid,
                "status": random.choice(["pending", "paid", "refunded"]),
                "total": random.randint(1000, 200000),
            },
        ).fetchone()
        order_ids.append(str(row[0]))
    return order_ids

# ... 130+ more lines:
# - seed_order_items needs orders pre-seeded
# - seed_refunds depends on paid orders only
# - ad-hoc retry loops for FK violations
# - custom cleanup and transaction guards
# - flaky CI due to order-of-operations drift

{
  "count": 50,
  "schema": {
    "users": { "columns": {
      "id": {"type":"uuid"},
      "email": {"type":"email"},
      "name": {"type":"name"}
    }},
    "orders": { "columns": {
      "id": {"type":"uuid"},
      "user_id": {"type":"foreign_key","reference":"users.id"},
      "status": {"type":"random_element"},
      "total_cents": {"type":"random_int"}
    }},
    "order_items": { "columns": {
      "id": {"type":"uuid"},
      "order_id": {"type":"foreign_key","reference":"orders.id"},
      "sku": {"type":"bothify"}
    }}
  }
}