service-finder/backend/scripts/seed_financial_ledger.py

#!/usr/bin/env python3
"""
Seed script: Populate audit.financial_ledger with 20 dummy records.
Uses psycopg2 (sync) to bypass asyncpg prepared statement enum validation issues.
"""
import uuid
import random
import os
from datetime import datetime, timedelta, timezone

import psycopg2

TRANSACTION_TYPES = [
    "SUBSCRIPTION", "COMMISSION", "REFUND", "MANUAL_ADJUSTMENT",
    "PURCHASE", "WITHDRAWAL", "BONUS", "FEE", "REWARD",
]

DB_ENTRY_TYPES = ["CREDIT", "DEBIT"]
DB_WALLET_TYPES = ["EARNED", "PURCHASED", "SERVICE_COINS", "VOUCHER"]
DB_STATUSES = ["PENDING", "SUCCESS", "FAILED", "REFUNDED"]

TEST_USER_IDS = [1, 2, 28, 29, 55, 75, 79, 85, 86, 88, 100, 102, 103, 104, 105, 106]

NOW = datetime.now(timezone.utc)


def get_dsn_params():
    """Extract connection parameters from DATABASE_URL."""
    db_url = os.environ.get("DATABASE_URL", "postgresql+asyncpg://service_finder_app:AppSafePass_2026@shared-postgres:5432/service_finder")
    db_url = db_url.replace("postgresql+asyncpg://", "postgresql://")
    # Parse URL
    parts = db_url.replace("postgresql://", "").split("@")
    user_pass = parts[0].split(":")
    host_db = parts[1].split("/")
    host_port = host_db[0].split(":")
    return {
        "user": user_pass[0],
        "password": user_pass[1],
        "host": host_port[0],
        "port": int(host_port[1]) if len(host_port) > 1 else 5432,
        "dbname": host_db[1].split("?")[0],
    }


def seed():
    params = get_dsn_params()
    conn = psycopg2.connect(**params)
    conn.autocommit = False

    try:
        with conn.cursor() as cur:
            # Count existing
            cur.execute("SELECT COUNT(*) FROM audit.financial_ledger")
            print(f"Existing records: {cur.fetchone()[0]}")

            # Get person_ids
            person_map = {}
            for uid in TEST_USER_IDS:
                cur.execute("SELECT person_id FROM identity.users WHERE id = %s", (uid,))
                pid = cur.fetchone()
                if pid and pid[0]:
                    person_map[uid] = pid[0]

            if not person_map:
                print("No users with person_id found!")
                return

            print(f"Users with person records: {list(person_map.keys())}")

            inserted = 0
            for i in range(20):
                uid = random.choice(list(person_map.keys()))
                pid = person_map[uid]
                entry_t = random.choice(DB_ENTRY_TYPES)
                wallet_t = random.choice(DB_WALLET_TYPES)
                stat = random.choice(DB_STATUSES)
                txn_t = random.choice(TRANSACTION_TYPES)
                cur_r = random.choice(["HUF", "EUR", "USD", "GBP"])
                amt = round(random.uniform(5.0, 5000.0), 2)
                days_ago = random.randint(0, 30)
                created_ts = NOW - timedelta(
                    days=random.randint(0, days_ago),
                    hours=random.randint(0, 23),
                    minutes=random.randint(0, 59),
                )
                ref = "TXN-" + uuid.uuid4().hex[:8].upper()
                desc = txn_t.lower().replace("_", " ").title() + f" for user {uid}"
                net_amt = round(amt / 1.27, 2) if random.random() > 0.3 else None
                tax_amt = round(amt - net_amt, 2) if net_amt else None
                balance = round(random.uniform(100, 50000), 2)
                inv_stat = "PAID" if stat == "completed" else stat.upper()
                txn_id = str(uuid.uuid4())

                # psycopg2 handles text → enum cast correctly with simple execute
                cur.execute(
                    """
                    INSERT INTO audit.financial_ledger
                        (user_id, person_id, amount, currency, transaction_type, details,
                         created_at, entry_type, balance_after, wallet_type, status,
                         transaction_id, net_amount, tax_amount, gross_amount, invoice_status)
                    VALUES (
                        %s, %s, %s, %s, %s,
                        %s::jsonb,
                        %s::timestamptz,
                        %s::audit.ledger_entry_type,
                        %s,
                        %s::audit.wallet_type,
                        %s::audit.ledger_status,
                        %s::uuid,
                        %s, %s, %s, %s
                    )
                    """,
                    (
                        uid, pid, amt, cur_r, txn_t,
                        f'{{"description": "{desc}", "reference": "{ref}"}}',
                        created_ts.isoformat(),
                        entry_t, balance, wallet_t, stat, txn_id,
                        net_amt, tax_amt, amt, inv_stat,
                    ),
                )
                inserted += 1

            conn.commit()
            print(f"✅ Inserted {inserted} records successfully!")

            cur.execute("SELECT COUNT(*) FROM audit.financial_ledger")
            print(f"Total records now: {cur.fetchone()[0]}")

            cur.execute("""
                SELECT
                    fl.entry_type::text,
                    fl.wallet_type::text,
                    fl.status::text,
                    COUNT(*)::int,
                    ROUND(AVG(fl.amount)::numeric, 2),
                    u.email
                FROM audit.financial_ledger fl
                LEFT JOIN identity.users u ON u.id = fl.user_id
                GROUP BY fl.entry_type, fl.wallet_type, fl.status, u.email
                ORDER BY fl.entry_type, fl.wallet_type
            """)
            rows = cur.fetchall()
            total_sum = sum(r[3] for r in rows) if rows else 0
            print(f"\n📊 Summary (total {total_sum} records):")
            print(f"  {'Entry':8} {'Wallet':15} {'Status':10} {'Cnt':5} {'Avg Amt':10} {'User'}")
            print(f"  {'-'*8} {'-'*15} {'-'*10} {'-'*5} {'-'*10} {'-'*30}")
            for r in rows:
                print(f"  {r[0]:8} {r[1]:15} {r[2]:10} {r[3]:5} {r[4]:10} {r[5] or 'N/A':30}")

    finally:
        conn.close()


if __name__ == "__main__":
    seed()