sigma-terminal 2.0.2__py3-none-any.whl → 3.2.0__py3-none-any.whl

sigma/backtest.py

@@ -0,0 +1,372 @@
+"""Backtesting engine for Sigma."""
+
+import json
+from datetime import datetime, timedelta
+from typing import Any, Optional
+
+import numpy as np
+import pandas as pd
+import yfinance as yf
+
+
+# ============================================================================
+# BACKTEST STRATEGIES
+# ============================================================================
+
+STRATEGIES = {
+    "sma_crossover": {
+        "name": "SMA Crossover",
+        "description": "Buy when short SMA crosses above long SMA, sell when crosses below",
+        "params": {"short_window": 20, "long_window": 50},
+    },
+    "rsi_mean_reversion": {
+        "name": "RSI Mean Reversion",
+        "description": "Buy when RSI < 30 (oversold), sell when RSI > 70 (overbought)",
+        "params": {"rsi_period": 14, "oversold": 30, "overbought": 70},
+    },
+    "macd_momentum": {
+        "name": "MACD Momentum",
+        "description": "Buy on MACD bullish crossover, sell on bearish crossover",
+        "params": {"fast": 12, "slow": 26, "signal": 9},
+    },
+    "bollinger_bands": {
+        "name": "Bollinger Bands",
+        "description": "Buy when price touches lower band, sell at upper band",
+        "params": {"window": 20, "num_std": 2},
+    },
+    "dual_momentum": {
+        "name": "Dual Momentum",
+        "description": "Combines absolute and relative momentum",
+        "params": {"lookback": 12},
+    },
+    "breakout": {
+        "name": "Breakout",
+        "description": "Buy on new highs, sell on new lows",
+        "params": {"lookback": 20},
+    },
+}
+
+
+def run_backtest(
+    symbol: str,
+    strategy: str,
+    period: str = "2y",
+    initial_capital: float = 100000,
+    params: Optional[dict] = None,
+) -> dict:
+    """Run a backtest for a given strategy."""
+    
+    if strategy not in STRATEGIES:
+        return {"error": f"Unknown strategy: {strategy}", "available": list(STRATEGIES.keys())}
+    
+    try:
+        # Get historical data
+        ticker = yf.Ticker(symbol.upper())
+        hist = ticker.history(period=period)
+        
+        if hist.empty or len(hist) < 50:
+            return {"error": "Insufficient data for backtest", "symbol": symbol}
+        
+        # Get strategy params
+        strat_info = STRATEGIES[strategy]
+        strat_params = params or strat_info["params"]
+        
+        # Generate signals based on strategy
+        if strategy == "sma_crossover":
+            signals = _sma_crossover_signals(hist, **strat_params)
+        elif strategy == "rsi_mean_reversion":
+            signals = _rsi_signals(hist, **strat_params)
+        elif strategy == "macd_momentum":
+            signals = _macd_signals(hist, **strat_params)
+        elif strategy == "bollinger_bands":
+            signals = _bollinger_signals(hist, **strat_params)
+        elif strategy == "dual_momentum":
+            signals = _momentum_signals(hist, **strat_params)
+        elif strategy == "breakout":
+            signals = _breakout_signals(hist, **strat_params)
+        else:
+            signals = pd.Series(0, index=hist.index)
+        
+        # Run simulation
+        results = _simulate_trades(hist, signals, initial_capital)
+        
+        # Calculate metrics
+        metrics = _calculate_metrics(results, hist, initial_capital)
+        
+        return {
+            "symbol": symbol.upper(),
+            "strategy": strat_info["name"],
+            "strategy_description": strat_info["description"],
+            "period": period,
+            "initial_capital": initial_capital,
+            "parameters": strat_params,
+            "performance": metrics["performance"],
+            "risk": metrics["risk"],
+            "trades": metrics["trades"],
+            "monthly_returns": metrics["monthly_returns"],
+        }
+        
+    except Exception as e:
+        return {"error": str(e), "symbol": symbol}
+
+
+def _sma_crossover_signals(hist: pd.DataFrame, short_window: int, long_window: int) -> pd.Series:
+    """Generate SMA crossover signals."""
+    short_sma = hist["Close"].rolling(short_window).mean()
+    long_sma = hist["Close"].rolling(long_window).mean()
+    
+    signals = pd.Series(0, index=hist.index)
+    signals[short_sma > long_sma] = 1  # Long
+    signals[short_sma < long_sma] = -1  # Exit
+    
+    return signals
+
+
+def _rsi_signals(hist: pd.DataFrame, rsi_period: int, oversold: int, overbought: int) -> pd.Series:
+    """Generate RSI mean reversion signals."""
+    delta = hist["Close"].diff()
+    gain = (delta.where(delta > 0, 0)).rolling(rsi_period).mean()
+    loss = (-delta.where(delta < 0, 0)).rolling(rsi_period).mean()
+    rs = gain / loss
+    rsi = 100 - (100 / (1 + rs))
+    
+    signals = pd.Series(0, index=hist.index)
+    signals[rsi < oversold] = 1  # Buy oversold
+    signals[rsi > overbought] = -1  # Sell overbought
+    
+    return signals
+
+
+def _macd_signals(hist: pd.DataFrame, fast: int, slow: int, signal: int) -> pd.Series:
+    """Generate MACD momentum signals."""
+    ema_fast = hist["Close"].ewm(span=fast).mean()
+    ema_slow = hist["Close"].ewm(span=slow).mean()
+    macd = ema_fast - ema_slow
+    macd_signal = macd.ewm(span=signal).mean()
+    
+    signals = pd.Series(0, index=hist.index)
+    signals[macd > macd_signal] = 1  # Bullish
+    signals[macd < macd_signal] = -1  # Bearish
+    
+    return signals
+
+
+def _bollinger_signals(hist: pd.DataFrame, window: int, num_std: int) -> pd.Series:
+    """Generate Bollinger Bands signals."""
+    mid = hist["Close"].rolling(window).mean()
+    std = hist["Close"].rolling(window).std()
+    upper = mid + (std * num_std)
+    lower = mid - (std * num_std)
+    
+    signals = pd.Series(0, index=hist.index)
+    signals[hist["Close"] < lower] = 1  # Buy at lower band
+    signals[hist["Close"] > upper] = -1  # Sell at upper band
+    
+    return signals
+
+
+def _momentum_signals(hist: pd.DataFrame, lookback: int) -> pd.Series:
+    """Generate dual momentum signals."""
+    returns = hist["Close"].pct_change(lookback * 21)  # Monthly lookback
+    
+    signals = pd.Series(0, index=hist.index)
+    signals[returns > 0] = 1  # Positive momentum
+    signals[returns < 0] = -1  # Negative momentum
+    
+    return signals
+
+
+def _breakout_signals(hist: pd.DataFrame, lookback: int) -> pd.Series:
+    """Generate breakout signals."""
+    high_roll = hist["High"].rolling(lookback).max()
+    low_roll = hist["Low"].rolling(lookback).min()
+    
+    signals = pd.Series(0, index=hist.index)
+    signals[hist["Close"] >= high_roll] = 1  # New high breakout
+    signals[hist["Close"] <= low_roll] = -1  # New low breakdown
+    
+    return signals
+
+
+def _simulate_trades(hist: pd.DataFrame, signals: pd.Series, initial_capital: float) -> dict:
+    """Simulate trades based on signals."""
+    capital = initial_capital
+    position = 0
+    shares = 0
+    trades = []
+    equity_curve = [initial_capital]
+    
+    for i in range(1, len(hist)):
+        date = hist.index[i]
+        price = hist["Close"].iloc[i]
+        signal = signals.iloc[i]
+        prev_signal = signals.iloc[i-1]
+        
+        # Buy signal
+        if signal == 1 and prev_signal != 1 and position == 0:
+            shares = int(capital * 0.95 / price)  # Use 95% of capital
+            if shares > 0:
+                cost = shares * price
+                capital -= cost
+                position = 1
+                trades.append({
+                    "date": str(date.date()),
+                    "action": "BUY",
+                    "price": round(price, 2),
+                    "shares": shares,
+                    "value": round(cost, 2),
+                })
+        
+        # Sell signal
+        elif signal == -1 and position == 1:
+            proceeds = shares * price
+            pnl = proceeds - (trades[-1]["value"] if trades else 0)
+            capital += proceeds
+            trades.append({
+                "date": str(date.date()),
+                "action": "SELL",
+                "price": round(price, 2),
+                "shares": shares,
+                "value": round(proceeds, 2),
+                "pnl": round(pnl, 2),
+            })
+            shares = 0
+            position = 0
+        
+        # Track equity
+        equity = capital + (shares * price if position == 1 else 0)
+        equity_curve.append(equity)
+    
+    return {
+        "trades": trades,
+        "equity_curve": equity_curve,
+        "final_equity": equity_curve[-1],
+        "final_position": position,
+        "final_shares": shares,
+    }
+
+
+def _calculate_metrics(results: dict, hist: pd.DataFrame, initial_capital: float) -> dict:
+    """Calculate backtest metrics."""
+    equity = np.array(results["equity_curve"])
+    trades = results["trades"]
+    
+    # Performance metrics
+    total_return = (equity[-1] / initial_capital - 1) * 100
+    
+    # Calculate daily returns
+    daily_returns = np.diff(equity) / equity[:-1]
+    
+    # Annualized metrics
+    trading_days = len(equity) - 1
+    years = trading_days / 252
+    annual_return = ((equity[-1] / initial_capital) ** (1/years) - 1) * 100 if years > 0 else 0
+    
+    # Risk metrics
+    volatility = np.std(daily_returns) * np.sqrt(252) * 100 if len(daily_returns) > 0 else 0
+    
+    # Sharpe ratio (assuming 0% risk-free rate)
+    sharpe = (np.mean(daily_returns) * 252) / (np.std(daily_returns) * np.sqrt(252)) if np.std(daily_returns) > 0 else 0
+    
+    # Max drawdown
+    peak = np.maximum.accumulate(equity)
+    drawdown = (peak - equity) / peak
+    max_drawdown = np.max(drawdown) * 100
+    
+    # Sortino ratio (downside deviation)
+    negative_returns = daily_returns[daily_returns < 0]
+    downside_std = np.std(negative_returns) * np.sqrt(252) if len(negative_returns) > 0 else 0
+    sortino = (np.mean(daily_returns) * 252) / downside_std if downside_std > 0 else 0
+    
+    # Calmar ratio
+    calmar = annual_return / max_drawdown if max_drawdown > 0 else 0
+    
+    # Trade statistics
+    winning_trades = [t for t in trades if t.get("action") == "SELL" and t.get("pnl", 0) > 0]
+    losing_trades = [t for t in trades if t.get("action") == "SELL" and t.get("pnl", 0) <= 0]
+    
+    num_trades = len([t for t in trades if t.get("action") == "SELL"])
+    win_rate = len(winning_trades) / num_trades * 100 if num_trades > 0 else 0
+    
+    avg_win = np.mean([t["pnl"] for t in winning_trades]) if winning_trades else 0
+    avg_loss = np.mean([t["pnl"] for t in losing_trades]) if losing_trades else 0
+    
+    profit_factor = abs(sum(t["pnl"] for t in winning_trades) / sum(t["pnl"] for t in losing_trades)) if losing_trades and sum(t["pnl"] for t in losing_trades) != 0 else 0
+    
+    # Buy and hold comparison
+    buy_hold_return = (hist["Close"].iloc[-1] / hist["Close"].iloc[0] - 1) * 100
+    alpha = total_return - buy_hold_return
+    
+    # Monthly returns
+    monthly_returns = []
+    if len(equity) > 21:
+        for i in range(21, len(equity), 21):
+            month_return = (equity[i] / equity[i-21] - 1) * 100
+            month_date = hist.index[min(i, len(hist)-1)]
+            monthly_returns.append({
+                "month": month_date.strftime("%Y-%m"),
+                "return": round(month_return, 2),
+            })
+    
+    return {
+        "performance": {
+            "initial_capital": f"${initial_capital:,.0f}",
+            "final_equity": f"${results['final_equity']:,.0f}",
+            "total_return": f"{total_return:.2f}%",
+            "annual_return": f"{annual_return:.2f}%",
+            "buy_hold_return": f"{buy_hold_return:.2f}%",
+            "alpha": f"{alpha:.2f}%",
+        },
+        "risk": {
+            "volatility": f"{volatility:.2f}%",
+            "max_drawdown": f"{max_drawdown:.2f}%",
+            "sharpe_ratio": f"{sharpe:.2f}",
+            "sortino_ratio": f"{sortino:.2f}",
+            "calmar_ratio": f"{calmar:.2f}",
+        },
+        "trades": {
+            "total_trades": num_trades,
+            "win_rate": f"{win_rate:.1f}%",
+            "profit_factor": f"{profit_factor:.2f}",
+            "avg_win": f"${avg_win:,.0f}",
+            "avg_loss": f"${avg_loss:,.0f}",
+            "trade_list": trades[-10:],  # Last 10 trades
+        },
+        "monthly_returns": monthly_returns[-12:],  # Last 12 months
+    }
+
+
+def get_available_strategies() -> dict:
+    """Get list of available strategies."""
+    return {
+        name: {
+            "name": info["name"],
+            "description": info["description"],
+            "default_params": info["params"],
+        }
+        for name, info in STRATEGIES.items()
+    }
+
+
+# Tool definition for LLM
+BACKTEST_TOOL = {
+    "type": "function",
+    "function": {
+        "name": "run_backtest",
+        "description": "Run a backtest simulation with a trading strategy",
+        "parameters": {
+            "type": "object",
+            "properties": {
+                "symbol": {"type": "string", "description": "Stock ticker symbol"},
+                "strategy": {
+                    "type": "string",
+                    "enum": list(STRATEGIES.keys()),
+                    "description": "Trading strategy to test"
+                },
+                "period": {"type": "string", "description": "Backtest period (e.g., 2y)", "default": "2y"},
+                "initial_capital": {"type": "number", "description": "Starting capital", "default": 100000},
+            },
+            "required": ["symbol", "strategy"]
+        }
+    }
+}