backend/app/ai/forecasting.py

from datetime import datetime, timedelta
import numpy as np
from sqlalchemy.orm import Session
from app.database.models import Account, Transaction, Goal, Investment, Subscription

def get_cashflow_metrics(db: Session, user_id: str, days: int = 90):
    """
    Computes daily average income and spending based on historical transactions.
    """
    # Fetch checking & savings accounts for user
    accounts = db.query(Account).filter(Account.user_id == user_id).all()
    account_ids = [acc.id for acc in accounts]
    
    if not account_ids:
        return 0.0, 0.0, 0.0  # Current balance, avg daily income, avg daily spending
        
    current_balance = sum(acc.balance for acc in accounts)
    
    # Fetch recent transactions
    cutoff = datetime.utcnow() - timedelta(days=days)
    txns = db.query(Transaction).filter(
        Transaction.account_id.in_(account_ids),
        Transaction.timestamp >= cutoff
    ).all()
    
    if not txns:
        return current_balance, 0.0, 0.0
        
    total_income = sum(t.amount for t in txns if t.type.lower() == "credit")
    total_spending = sum(t.amount for t in txns if t.type.lower() == "debit")
    
    avg_daily_income = total_income / days
    avg_daily_spending = total_spending / days
    
    return current_balance, avg_daily_income, avg_daily_spending

def predict_future_balance(db: Session, user_id: str, projection_days: int = 90):
    """
    Predicts future balances and returns trend description.
    """
    current_balance, daily_income, daily_spending = get_cashflow_metrics(db, user_id)
    
    net_daily = daily_income - daily_spending
    projected_balance = max(0.0, current_balance + (net_daily * projection_days))
    
    # Calculate percentage change
    if current_balance > 0:
        percent_change = (projected_balance - current_balance) / current_balance * 100
    else:
        percent_change = 0.0
        
    # Generate human-friendly description
    if percent_change < 0:
        insight = f"If current spending continues, your total balance may decrease by {abs(percent_change):.1f}% (down to ${projected_balance:,.2f}) in {projection_days} days."
    elif percent_change > 0:
        insight = f"Based on current trends, your total balance is projected to grow by {percent_change:.1f}% (up to ${projected_balance:,.2f}) in {projection_days} days."
    else:
        insight = "Your financial trajectory is steady with minor balance fluctuations."
        
    # Generate daily data points for charts
    chart_data = []
    for day in range(0, projection_days + 1, 5):
        val = max(0.0, current_balance + (net_daily * day))
        date_str = (datetime.utcnow() + timedelta(days=day)).strftime("%Y-%m-%d")
        chart_data.append({"date": date_str, "balance": round(val, 2)})
        
    return {
        "current_balance": round(current_balance, 2),
        "projected_balance": round(projected_balance, 2),
        "percent_change": round(percent_change, 2),
        "net_daily": round(net_daily, 2),
        "insight": insight,
        "chart_data": chart_data
    }

def forecast_savings_and_investments(db: Session, user_id: str, projection_months: int = 12):
    """
    Projects savings and investment growth.
    """
    accounts = db.query(Account).filter(Account.user_id == user_id).all()
    savings_balance = sum(acc.balance for acc in accounts if acc.type.lower() == "savings")
    checking_balance = sum(acc.balance for acc in accounts if acc.type.lower() == "checking")
    
    investments = db.query(Investment).filter(Investment.user_id == user_id).all()
    total_invested = sum(inv.current_value for inv in investments)
    
    # Subscriptions and recurring bills
    subs = db.query(Subscription).filter(Subscription.user_id == user_id, Subscription.active == True).all()
    monthly_sub_cost = sum(sub.amount if sub.billing_cycle.lower() == "monthly" else (sub.amount / 12) for sub in subs)
    
    # Let's assume standard default rates if not specified
    savings_apr = 0.04  # 4% APY
    investment_apr = 0.08  # 8% APY
    
    # We assume the user saves 10% of their net income monthly (derived from transaction history)
    _, daily_income, daily_spending = get_cashflow_metrics(db, user_id)
    monthly_income = daily_income * 30.4
    monthly_spending = daily_spending * 30.4
    net_monthly = max(0.0, monthly_income - monthly_spending)
    monthly_savings_addition = net_monthly * 0.5  # Put 50% of net into savings
    monthly_investment_addition = net_monthly * 0.3  # Put 30% of net into investments
    
    savings_data = []
    investment_data = []
    debt_data = []
    
    current_savings = savings_balance
    current_inv = total_invested
    
    # Let's model a baseline debt if the user has a Goal of type "debt" or a general dummy debt
    # We will look for Goals containing "debt" or "loan"
    goals = db.query(Goal).filter(Goal.user_id == user_id).all()
    debt_goal = next((g for g in goals if "debt" in g.title.lower() or "loan" in g.title.lower()), None)
    
    total_debt = 5000.0  # Default initial simulated debt if none found
    if debt_goal:
        total_debt = max(0.0, debt_goal.target_amount - debt_goal.current_amount)
    
    monthly_debt_payment = min(total_debt, max(150.0, net_monthly * 0.1)) # Assume 10% of net or at least $150
    
    for month in range(0, projection_months + 1):
        # Compounding savings
        if month > 0:
            current_savings = (current_savings + monthly_savings_addition) * (1 + savings_apr / 12)
            current_inv = (current_inv + monthly_investment_addition) * (1 + investment_apr / 12)
            total_debt = max(0.0, total_debt - monthly_debt_payment)
            
        label = f"Month {month}"
        savings_data.append({"month": label, "amount": round(current_savings, 2)})
        investment_data.append({"month": label, "amount": round(current_inv, 2)})
        debt_data.append({"month": label, "amount": round(total_debt, 2)})
        
    return {
        "projection_months": projection_months,
        "monthly_savings_addition": round(monthly_savings_addition, 2),
        "monthly_investment_addition": round(monthly_investment_addition, 2),
        "savings_growth": savings_data,
        "investment_growth": investment_data,
        "debt_decline": debt_data,
        "total_projected_savings": round(current_savings, 2),
        "total_projected_investments": round(current_inv, 2),
        "total_remaining_debt": round(total_debt, 2)
    }

def simulate_future_scenarios(db: Session, user_id: str, projection_months: int = 6):
    """
    Simulates three scenarios: Status Quo, Frugal (cut spending 20%), and Luxury (increase spending 15%).
    """
    current_balance, daily_income, daily_spending = get_cashflow_metrics(db, user_id)
    monthly_income = daily_income * 30.4
    monthly_spending = daily_spending * 30.4
    
    scenarios = {
        "status_quo": {"spend_mult": 1.0, "name": "Status Quo (Current spending)"},
        "frugal": {"spend_mult": 0.8, "name": "Frugal Mode (Cut non-essentials by 20%)"},
        "lifestyle_inflation": {"spend_mult": 1.15, "name": "Lifestyle Inflation (+15% spending)"}
    }
    
    results = {}
    for key, config in scenarios.items():
        mult = config["spend_mult"]
        projected_spend = monthly_spending * mult
        net_monthly = monthly_income - projected_spend
        
        balance_trend = []
        balance = current_balance
        for m in range(0, projection_months + 1):
            if m > 0:
                balance = max(0.0, balance + net_monthly)
            balance_trend.append({"month": f"M{m}", "balance": round(balance, 2)})
            
        results[key] = {
            "name": config["name"],
            "monthly_income": round(monthly_income, 2),
            "monthly_spending": round(projected_spend, 2),
            "net_monthly": round(net_monthly, 2),
            "balance_projection": balance_trend,
            "final_balance": round(balance, 2),
            "savings_change_pct": round(((balance - current_balance) / current_balance * 100) if current_balance > 0 else 0.0, 2)
        }
        
    return results