backend/app/scripts/seed.py

import os
import sys
import uuid
import random
from datetime import datetime, timedelta

# Add parent directory to path so we can import from app
sys.path.append(os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__)))))

from app.database.database import SessionLocal, engine, SQLALCHEMY_DATABASE_URL
from app.database.models import (
    Base, User, Account, Transaction, Subscription,
    Goal, Investment, AIInsight, FraudLog, Notification, AnalyticsSnapshot
)

# Create tables
Base.metadata.create_all(bind=engine)

def seed_data():
    db = SessionLocal()

    print(f"Seeding into: {SQLALCHEMY_DATABASE_URL}")

    # Check if we already have users
    if db.query(User).count() > 0:
        print("Database already seeded.")
        db.close()
        return

    print("Seeding database...")

    personas = ["Saver", "Investor", "Impulsive Spender", "Minimalist", "Risk Taker"]

    merchants = ["Swiggy", "Amazon", "Netflix", "Uber", "Fuel", "Salary", "SIP",
                 "Starbucks", "Apple", "Walmart"]
    categories = ["Food", "Shopping", "Entertainment", "Transport", "Income",
                  "Investment", "Groceries", "Tech", "Utilities"]

    for persona in personas:
        try:
            user = User(
                email=f"{persona.lower().replace(' ', '_')}@example.com",
                password_hash="hashed_password",
                profile_data={"name": f"{persona} User", "phone": "+1234567890"},
                financial_personality=persona,
                ai_personalization_settings={"theme": "dark", "notifications": "all"}
            )
            db.add(user)
            db.flush()  # get user.id without committing

            # ── Accounts ──────────────────────────────────────────────────────
            checking = Account(
                user_id=user.id, type="checking",
                balance=random.uniform(1000.0, 10000.0), currency="USD"
            )
            savings = Account(
                user_id=user.id, type="savings",
                balance=random.uniform(5000.0, 50000.0), currency="USD"
            )
            db.add_all([checking, savings])
            db.flush()

            # ── Subscriptions ─────────────────────────────────────────────────
            # Active subscription (high usage)
            db.add(Subscription(
                user_id=user.id, merchant="Netflix", amount=15.99,
                billing_cycle="monthly", active=True,
                ai_usage_detection={"usage_frequency": "high", "recommendation": "keep"}
            ))
            # Unused subscription — triggers unused detection in subscriptions.py
            db.add(Subscription(
                user_id=user.id, merchant="Spotify", amount=9.99,
                billing_cycle="monthly", active=True,
                ai_usage_detection={"usage_frequency": "low", "recommendation": "cancel"}
            ))
            # Duplicate subscription — triggers duplicate detection in subscriptions.py
            # (second Netflix entry for the same user)
            db.add(Subscription(
                user_id=user.id, merchant="Netflix", amount=15.99,
                billing_cycle="monthly", active=True,
                ai_usage_detection={"usage_frequency": "medium", "recommendation": "review"}
            ))

            # ── Goals ─────────────────────────────────────────────────────────
            db.add(Goal(
                user_id=user.id, title="Emergency Fund",
                target_amount=10000.0,
                current_amount=random.uniform(1000.0, 5000.0),
                target_date=datetime.utcnow() + timedelta(days=365),
                ai_generated_plan={"monthly_saving_required": 500.0, "risk": "low"}
            ))

            # ── Investments ───────────────────────────────────────────────────
            db.add(Investment(
                user_id=user.id, asset_name="S&P 500", type="stock",
                amount_invested=random.uniform(1000.0, 10000.0),
                current_value=random.uniform(1100.0, 12000.0),
                portfolio_allocation=50.0,
                ai_risk_analysis={"risk_level": "medium", "recommendation": "hold"}
            ))

            # ── Transactions ──────────────────────────────────────────────────
            start_date = datetime.utcnow() - timedelta(days=90)

            # Monthly salary (3 months)
            for i in range(3):
                tx_date = start_date + timedelta(days=i * 30)
                db.add(Transaction(
                    account_id=checking.id, amount=5000.0, type="credit",
                    category="Income", timestamp=tx_date, merchant="Salary",
                    tags=["salary", "income"],
                    ai_generated_metadata={"is_recurring": True, "confidence": 0.99},
                    spending_emotion_label="neutral"
                ))

            # Regular expense transactions
            for _ in range(30):
                tx_date = start_date + timedelta(days=random.randint(0, 89))
                amount = random.uniform(10.0, 500.0)
                merchant = random.choice(merchants)

                if merchant == "Salary":
                    continue

                # Persona-based spending adjustments
                if user.financial_personality == "Saver" and amount > 200:
                    amount = random.uniform(10.0, 100.0)
                elif user.financial_personality == "Impulsive Spender":
                    amount = random.uniform(50.0, 800.0)

                tx = Transaction(
                    account_id=checking.id, amount=amount, type="debit",
                    category=random.choice(categories),
                    timestamp=tx_date, merchant=merchant,
                    tags=["expense"],
                    ai_generated_metadata={"category_confidence": 0.9},
                    spending_emotion_label=random.choice(["happy", "regret", "neutral", "essential"])
                )
                db.add(tx)
                db.flush()

                # Seed a fraud log for ~5% of transactions
                if random.random() < 0.05:
                    db.add(FraudLog(
                        transaction_id=tx.id,
                        risk_score=random.uniform(0.7, 0.99),
                        suspicious_activity_details="Unusual location and high amount for this merchant.",
                        status="pending"
                    ))

            # Late-night transaction — ensures behavior.py late-night detection fires
            late_night_date = start_date + timedelta(days=random.randint(1, 80),
                                                      hours=23, minutes=random.randint(0, 59))
            db.add(Transaction(
                account_id=checking.id,
                amount=random.uniform(50.0, 300.0),
                type="debit",
                category="Entertainment",
                timestamp=late_night_date,
                merchant="Online Store",
                tags=["late-night", "impulse"],
                ai_generated_metadata={"category_confidence": 0.85},
                spending_emotion_label="regret"
            ))

            # ── Supporting records ────────────────────────────────────────────
            db.add(AIInsight(
                user_id=user.id, type="cashflow",
                content=f"You are spending 20% more on {random.choice(categories)} this month."
            ))
            db.add(Notification(
                user_id=user.id, title="Weekly Summary",
                message="Your weekly financial summary is ready.", type="insight"
            ))
            db.add(AnalyticsSnapshot(
                user_id=user.id, date=datetime.utcnow(),
                total_balance=checking.balance + savings.balance,
                total_spending=2000.0, total_savings=3000.0,
                financial_score=random.uniform(60.0, 95.0),
                trends_json={"spending_trend": "down", "savings_trend": "up"}
            ))

            db.commit()
            print(f"  ✓ Seeded user: {persona}")

        except Exception as e:
            db.rollback()
            print(f"  ✗ Failed to seed user '{persona}': {e}")

    db.close()
    print("Database seeded successfully!")

if __name__ == "__main__":
    seed_data()