[Refactor] Simplify signal generation in BollingerMeanReversion, BuyAndHold, DualMAStrategy, and DualThrustStrategy

Author: KarhouTam

src/quanteval/strategies/bollinger_mean_reversion.py

@@ -52,19 +52,10 @@ def generate_signals(self, data: pd.DataFrame) -> pd.Series:
         upper = middle + (num_std * std)
         lower = middle - (num_std * std)
 
-        # Initialize signal with NaN
-        signal = pd.Series(np.nan, index=data.index)
-
-        # Entry signal: Price touches lower band (buy signal)
+        # Generate signals
+        signal = pd.Series(np.nan, index=data.index, name='Signal')
         signal[data['Close'] < lower] = 1
-
-        # Exit signal: Price touches upper band (sell signal)
         signal[data['Close'] > upper] = 0
 
-        # Forward fill to maintain position between signals
-        signal = signal.ffill()
-
-        # Fill initial NaNs with 0 (no position)
-        signal = signal.fillna(0)
+        return signal.ffill().fillna(0)
 
-        return signal

src/quanteval/strategies/buy_hold.py

@@ -21,6 +21,8 @@ class BuyAndHoldStrategy(Strategy):
         >>> results = bt.run()
     """
 
+    name = 'BuyAndHold'
+
     def __init__(self):
         super().__init__(name='BuyAndHold')
 

src/quanteval/strategies/dual_ma.py

@@ -60,7 +60,9 @@ def generate_signals(self, data: pd.DataFrame) -> pd.Series:
             fast_ma = data['Close'].rolling(window=fast_window).mean()
             slow_ma = data['Close'].rolling(window=slow_window).mean()
 
-        # Generate signal: 1 if fast > slow, else 0
-        signal = np.where(fast_ma > slow_ma, 1, 0)
+        # Generate signals
+        signal = pd.Series(np.nan, index=data.index, name='Signal')
+        signal[fast_ma > slow_ma] = 1
+        signal[fast_ma <= slow_ma] = 0
 
-        return pd.Series(signal, index=data.index, name='Signal')
+        return signal.ffill().fillna(0)

src/quanteval/strategies/dual_thrust.py

@@ -49,8 +49,7 @@ def generate_signals(self, data: pd.DataFrame) -> pd.Series:
         Returns:
             Series with values:
                 1: Long position (price > upper threshold)
-                -1: Short position (price < lower threshold)
-                0: No position
+                0: No position (price < lower threshold)
         """
         k1 = self.params['k1']
         k2 = self.params['k2']
@@ -61,15 +60,16 @@ def generate_signals(self, data: pd.DataFrame) -> pd.Series:
         hc = data['Close'].shift(1).rolling(window=n).max()
         ll = data['Low'].shift(1).rolling(window=n).min()
         lc = data['Close'].shift(1).rolling(window=n).min()
+
         diff = np.maximum(hh - lc, hc - ll)
 
         # Calculate thresholds
-        upper_threshold = data['Open'] + k1 * diff
-        lower_threshold = data['Open'] - k2 * diff
+        upper = data['Open'] + k1 * diff
+        lower = data['Open'] - k2 * diff
 
         # Generate signals
-        signals = pd.Series(0, index=data.index)
-        signals[data['High'] > upper_threshold] = 1  # Buy signal
-        signals[data['Low'] < lower_threshold] = 0  # Sell signal
+        signal = pd.Series(np.nan, index=data.index, name='Signal')
+        signal[data['Close'] > upper] = 1
+        signal[data['Close'] < lower] = 0
 
-        return signals
+        return signal.ffill().fillna(0)