backend/finance/analytics.py

import numpy as np
from datetime import datetime, timedelta
from collections import defaultdict

class FinancialAnalytics:
    def __init__(self, incomes, expenses):
        self.incomes = sorted(incomes, key=lambda x: x['date'])
        self.expenses = sorted(expenses, key=lambda x: x['date'])
        self.all_txns = sorted(incomes + expenses, key=lambda x: x['date'])
        
        # Basic Aggregates
        self.total_income = sum(float(i.get('amount', 0)) for i in self.incomes)
        self.total_expense = sum(float(e.get('amount', 0)) for e in self.expenses)
        self.net_savings = self.total_income - self.total_expense
        
    def _get_monthly_aggregates(self, txns):
        monthly_map = defaultdict(float)
        for t in txns:
            m = t['date'].strftime('%Y-%m')
            monthly_map[m] += float(t.get('amount', 0))
        return monthly_map

    def calculate_executive_overview(self):
        savings_rate = (self.net_savings / self.total_income * 100) if self.total_income > 0 else 0
        
        # Cash Burn Rate (Monthly Avg Expense)
        monthly_expenses = self._get_monthly_aggregates(self.expenses)
        burn_rate = np.mean(list(monthly_expenses.values())) if monthly_expenses else 0
        
        # Stability Index (Simple Volatility measure)
        if monthly_expenses:
            std_dev = np.std(list(monthly_expenses.values()))
            mean_exp = np.mean(list(monthly_expenses.values()))
            # Lower CV (Coefficient of Variation) is better. 
            # Score 100 - (CV * 100). If CV > 1, score 0.
            cv = std_dev / mean_exp if mean_exp > 0 else 0
            stability_score = max(0, min(100, 100 - (cv * 100)))
        else:
            stability_score = 0
            
        return {
            "total_income": self.total_income,
            "total_expenses": self.total_expense,
            "net_cash_flow": self.net_savings,
            "savings_rate": savings_rate,
            "burn_rate": burn_rate,
            "stability_score": stability_score,
            "status": "Surplus" if self.net_savings >= 0 else "Deficit"
        }

    def analyze_income_structure(self):
        sources = defaultdict(float)
        for i in self.incomes:
            sources[i.get('category', 'Uncategorized')] += float(i.get('amount', 0))
            
        # HHI (Herfindahl-Hirschman Index) for Concentration
        total_sq = sum(v**2 for v in sources.values())
        hhi = (total_sq / (self.total_income**2)) if self.total_income > 0 else 0
        
        monthly_incomes = self._get_monthly_aggregates(self.incomes)
        consistency = 0
        if monthly_incomes:
            std_dev = np.std(list(monthly_incomes.values()))
            mean_inc = np.mean(list(monthly_incomes.values()))
            cv = std_dev / mean_inc if mean_inc > 0 else 0
            consistency = max(0, min(100, 100 - (cv * 50))) # Penalty for volatility

        return {
            "sources": dict(sources),
            "concentration_score": hhi * 100, # 100 = Single source (High Risk), 0 = Diverse
            "consistency_score": consistency,
            "monthly_variance": np.var(list(monthly_incomes.values())) if monthly_incomes else 0
        }

    def analyze_expense_structure(self):
        categories = defaultdict(float)
        for e in self.expenses:
            categories[e.get('category', 'Uncategorized')] += float(e.get('amount', 0))
            
        fixed_proxy = ['rent', 'mortgage', 'insurance', 'subscription', 'utilities', 'internet']
        fixed_total = sum(v for k,v in categories.items() if any(f in k.lower() for f in fixed_proxy))
        
        return {
            "breakdown": dict(categories),
            "fixed_ratio": (fixed_total / self.total_expense * 100) if self.total_expense > 0 else 0,
            "discretionary_ratio": 100 - ((fixed_total / self.total_expense * 100) if self.total_expense > 0 else 0)
        }

    def calculate_financial_health_score(self):
        # 0-100 Composite Score
        exec_data = self.calculate_executive_overview()
        inc_data = self.analyze_income_structure()
        
        score = 0
        
        # 1. Savings Rate (Max 40)
        s_rate = exec_data['savings_rate']
        score += min(40, s_rate * 2) # 20% savings = 40 pts
        
        # 2. Stability (Max 20)
        score += (exec_data['stability_score'] * 0.2)
        
        # 3. Income Consistency (Max 20)
        score += (inc_data['consistency_score'] * 0.2)
        
        # 4. Solvency/Liquidity (Max 20) - Simplified as positive cash flow
        if exec_data['net_cash_flow'] > 0:
            score += 20
            
        return min(100, max(0, score))

    def analyze_temporal_patterns(self):
        # Day of Week Analysis (0=Mon, 6=Sun)
        dow_map = defaultdict(float)
        for e in self.expenses:
            dow = e['date'].weekday()
            dow_map[dow] += float(e.get('amount', 0))
            
        days = ['Mon', 'Tue', 'Wed', 'Thu', 'Fri', 'Sat', 'Sun']
        return [
            {'day': days[i], 'amount': round(dow_map[i], 2)} 
            for i in range(7)
        ]

    def detect_recurring_payments(self):
        # Simple algorithm: Group by Merchant + Amount (rounded)
        groups = defaultdict(list)
        for e in self.expenses:
            key = (e.get('title', '').strip().lower(), round(float(e.get('amount', 0)), 2))
            groups[key].append(e['date'])
            
        recurring = []
        for (merchant, amount), dates in groups.items():
            if len(dates) >= 2:
                # Check if dates are roughly 28-31 days apart
                dates.sort()
                intervals = [(dates[i+1] - dates[i]).days for i in range(len(dates)-1)]
                avg_interval = np.mean(intervals)
                if 25 <= avg_interval <= 35:
                    recurring.append({
                        'merchant': merchant.title(),
                        'amount': amount,
                        'frequency': 'Monthly',
                        'next_due': (dates[-1] + timedelta(days=30)).strftime('%Y-%m-%d')
                    })
        return recurring

    def analyze_capex(self):
        # Purchases > $500
        capex = []
        for e in self.expenses:
            amt = float(e.get('amount', 0))
            if amt > 500:
                capex.append({
                    'date': e['date'].strftime('%Y-%m-%d'),
                    'item': e.get('title', 'Large Purchase'),
                    'amount': amt,
                    'category': e.get('category', 'Uncategorized')
                })
        return sorted(capex, key=lambda x: x['date'], reverse=True)

    def estimate_tax_liability(self):
        # Simple Progressive Tax Estimation (Annualized)
        if not self.incomes: return {'estimated_tax': 0, 'bracket': '0%'}
        
        # Annualize based on available data range or just assume simple run rate
        annual_income = self.total_income # Simplified for now, assuming data IS annual or close to it
        
        # 2025 Standard Brackets (Simplified Single Filer)
        tax = 0
        income = annual_income
        brackets = [
            (11600, 0.10),
            (47150, 0.12),
            (100525, 0.22),
            (191950, 0.24),
            (243725, 0.32),
            (609350, 0.35),
            (float('inf'), 0.37)
        ]
        
        previous_limit = 0
        for limit, rate in brackets:
            if income > previous_limit:
                taxable_in_bracket = min(income, limit) - previous_limit
                tax += taxable_in_bracket * rate
                previous_limit = limit
            else:
                break
                
        effective_rate = (tax / annual_income * 100) if annual_income > 0 else 0
        return {
            'estimated_tax': round(tax, 2),
            'effective_rate': round(effective_rate, 1),
            'taxable_income': round(annual_income, 2)
        }

    def calculate_runway(self):
        # Months to Zero
        burn_rate = self.calculate_executive_overview()['burn_rate']
        
        # Assuming Net Savings IS the current liquid assets for this context
        # In reality, we'd query a "Balance" but we only have transaction history here
        # So we use net_savings as a proxy for "Cash on Hand" generated during this period
        assets = self.net_savings 
        
        if burn_rate <= 0: return "Infinite"
        if assets <= 0: return "0 Months"
        
        months = assets / burn_rate
        return round(months, 1)

    def analyze_monthly_trends(self):
        # Last 12 Months Trends (Precise iteration)
        today = datetime.now()
        trends = []
        
        # Start from 11 months ago to current month
        for i in range(11, -1, -1):
            year = today.year
            month = today.month - i
            
            # Adjust for year wrapping
            while month <= 0:
                month += 12
                year -= 1
            
            month_str = f"{year}-{month:02d}"
            
            inc = sum(float(x.get('amount', 0)) for x in self.incomes if x['date'].strftime('%Y-%m') == month_str)
            exp = sum(float(x.get('amount', 0)) for x in self.expenses if x['date'].strftime('%Y-%m') == month_str)
            
            trends.append({
                "month": month_str,
                "income": round(inc, 2),
                "expense": round(exp, 2),
                "net": round(inc - exp, 2)
            })
        return trends

    def analyze_category_details(self):
        # Detailed Expense Breakdown
        if self.total_expense == 0: return []
        
        cats = defaultdict(float)
        for e in self.expenses:
            cats[e.get('category', 'Uncategorized')] += float(e.get('amount', 0))
            
        sorted_cats = sorted(cats.items(), key=lambda x: x[1], reverse=True)
        
        return [
            {
                "category": k,
                "amount": round(v, 2),
                "percentage": round((v / self.total_expense) * 100, 1)
            }
            for k, v in sorted_cats
        ]

    def generate_recommendations(self):
        # Rule-Based Advise
        recs = []
        
        # 1. Savings Rule
        if self.total_income > 0:
            savings_rate = (self.net_savings / self.total_income) * 100
            if savings_rate < 20:
                recs.append(f"Increase savings rate. Current is {round(savings_rate, 1)}%, aim for >20%.")
            else:
                recs.append(f"Great savings rate ({round(savings_rate, 1)}%). Consider investing the surplus.")
        
        # 2. Expense Rules
        monthly_trends = self.analyze_monthly_trends()
        if monthly_trends:
            avg_exp = np.mean([m['expense'] for m in monthly_trends])
            if monthly_trends[-1]['expense'] > avg_exp * 1.2:
                recs.append("Recent month expenses are 20% higher than average. Review discretionary spending.")

        # 3. Category Rules
        cats = self.analyze_category_details()
        for c in cats:
            if c['category'].lower() in ['dining', 'entertainment', 'shopping'] and c['percentage'] > 15:
                recs.append(f"High spending in {c['category']} ({c['percentage']}%). Try to reduce to <10%.")
                
        # 4. Burn Rate
        burn = self.calculate_executive_overview()['burn_rate']
        if burn > 0 and self.net_savings < burn * 3:
             recs.append("Emergency fund is low. Build a buffer of at least 3 months of expenses.")

        return recs

    def calculate_financial_velocity(self):
        # Real-time pacing analysis
        # ANCHOR FIX: Use latest available data date as "today" for velocity calculations
        if self.expenses:
            today = max(e['date'] for e in self.expenses)
        elif self.all_txns:
            today = max(t['date'] for t in self.all_txns)
        else:
            today = datetime.now()
            
        current_month_str = today.strftime('%Y-%m')
        
        # Filter for current month expenses
        current_month_expenses = [
            float(e.get('amount', 0)) for e in self.expenses 
            if e['date'].strftime('%Y-%m') == current_month_str
        ]
        
        spent_so_far = sum(current_month_expenses)
        days_passed = max(1, today.day)
        daily_burn = spent_so_far / days_passed
        
        if today.month == 12:
            next_month = today.replace(year=today.year + 1, month=1, day=1)
        else:
            next_month = today.replace(month=today.month + 1, day=1)
            
        days_in_month = (next_month - timedelta(days=1)).day
        projected_total = daily_burn * days_in_month
        
        # Calculate trailing 3-month average for comparison
        past_3_months = []
        for i in range(1, 4):
            d = today - timedelta(days=i*30)
            m_str = d.strftime('%Y-%m')
            m_exp = sum(float(e.get('amount', 0)) for e in self.expenses if e['date'].strftime('%Y-%m') == m_str)
            if m_exp > 0: past_3_months.append(m_exp)
            
        avg_monthly_spend = np.mean(past_3_months) if past_3_months else (projected_total * 1.1)
        
        velocity_status = "On Track"
        if projected_total > avg_monthly_spend * 1.15:
            velocity_status = "Accelerating (High Spend)"
        elif projected_total < avg_monthly_spend * 0.85:
            velocity_status = "Decelerating (Saving)"
            
        return {
            "spent_so_far": round(spent_so_far, 2),
            "daily_burn": round(daily_burn, 2),
            "projected_total": round(projected_total, 2),
            "benchmark_avg": round(avg_monthly_spend, 2),
            "status": velocity_status,
            "days_remaining": days_in_month - days_passed
        }

    def get_recent_transactions(self, limit=5):
        # Get purely recent activity for "Live" feel
        # self.all_txns is already sorted by date (asc), so take from end
        recent = sorted(self.all_txns, key=lambda x: x['date'], reverse=True)[:limit]
        
        formatted = []
        for t in recent:
            formatted.append({
                "date": t['date'].strftime('%Y-%m-%d'),
                "title": t.get('title', 'Unknown'),
                "amount": float(t.get('amount', 0)),
                "type": "expense" if t in self.expenses else "income",
                "category": t.get('category', 'Uncategorized')
            })
        return formatted

    def generate_full_report(self):
        return {
            "executive": self.calculate_executive_overview(),
            "income_analysis": self.analyze_income_structure(),
            "expense_analysis": self.analyze_expense_structure(),
            "health_score": self.calculate_financial_health_score(),
            "temporal": self.analyze_temporal_patterns(),
            "recurring": self.detect_recurring_payments(),
            "capex": self.analyze_capex(),
            "tax": self.estimate_tax_liability(),
            "runway": self.calculate_runway(),
            "monthly_trends": self.analyze_monthly_trends(),
            "category_details": self.analyze_category_details(),
            "recommendations": self.generate_recommendations(),
            "velocity": self.calculate_financial_velocity(),
            "recent_activity": self.get_recent_transactions()
        }