feat: 添加股票价格的拉普拉斯可视化和基于FFT的基本指标计算

Laplace-display.py

@@ -0,0 +1,86 @@
+import yfinance as yf
+import numpy as np
+import matplotlib.pyplot as plt
+from scipy import ndimage
+import pandas as pd
+
+# Define stock symbols
+stocks = ['AAPL', 'MSFT', 'GOOGL', 'META', 'NVDA', 'TSLA', 'AMD', 'INTC']
+
+# Download data
+# Use one year of daily data
+end_date = pd.Timestamp.today()
+start_date = end_date - pd.DateOffset(years=1)
+
+# Create empty DataFrame to store prices
+df = pd.DataFrame()
+
+# Download data for each stock
+for symbol in stocks:
+    stock = yf.download(symbol, start=start_date, end=end_date, progress=False)
+    df[symbol] = stock['Close']
+
+# Fill any missing values using forward fill
+df = df.fillna(method='ffill')
+
+# Normalize prices to start from 100 for better comparison
+normalized_prices = df.div(df.iloc[0]) * 100
+
+# Convert to numpy array for Laplacian calculation
+prices_array = normalized_prices.values.T  # Shape: (n_stocks, n_times)
+
+# Calculate Laplacian
+laplacian = ndimage.laplace(prices_array)
+
+# Create visualization
+fig, (ax1, ax2, ax3) = plt.subplots(3, 1, figsize=(15, 15))
+
+# Plot normalized prices
+for i, symbol in enumerate(stocks):
+    ax1.plot(normalized_prices.index, normalized_prices[symbol], 
+             label=symbol, alpha=0.7)
+ax1.set_title('Normalized Stock Prices (Starting at 100)')
+ax1.set_xlabel('Date')
+ax1.set_ylabel('Normalized Price')
+ax1.legend(bbox_to_anchor=(1.05, 1), loc='upper left')
+ax1.grid(True)
+
+# Plot prices as a heatmap
+im2 = ax2.imshow(prices_array, aspect='auto', cmap='viridis',
+                 extent=[0, len(df), 0, len(stocks)])
+ax2.set_title('Price Heatmap')
+ax2.set_xlabel('Trading Days')
+ax2.set_ylabel('Stock')
+ax2.set_yticks(np.arange(len(stocks)) + 0.5)
+ax2.set_yticklabels(stocks)
+plt.colorbar(im2, ax=ax2, label='Normalized Price')
+
+# Plot Laplacian
+im3 = ax3.imshow(laplacian, aspect='auto', cmap='coolwarm',
+                 extent=[0, len(df), 0, len(stocks)])
+ax3.set_title('Laplacian of Price Surface')
+ax3.set_xlabel('Trading Days')
+ax3.set_ylabel('Stock')
+ax3.set_yticks(np.arange(len(stocks)) + 0.5)
+ax3.set_yticklabels(stocks)
+plt.colorbar(im3, ax=ax3, label='Laplacian Value')
+
+plt.tight_layout()
+plt.show()
+
+# Print some basic statistics about the Laplacian values
+print("\nLaplacian Statistics:")
+print(f"Mean Laplacian value: {np.mean(laplacian):.4f}")
+print(f"Standard deviation: {np.std(laplacian):.4f}")
+print(f"Min value: {np.min(laplacian):.4f}")
+print(f"Max value: {np.max(laplacian):.4f}")
+
+# Find the most significant anomalies
+threshold = 2 * np.std(laplacian)
+anomalies = np.where(np.abs(laplacian) > threshold)
+print("\nSignificant price anomalies (beyond 2 standard deviations):")
+for stock_idx, time_idx in zip(*anomalies):
+    date = df.index[time_idx]
+    stock = stocks[stock_idx]
+    laplacian_value = laplacian[stock_idx, time_idx]
+    print(f"Stock: {stock}, Date: {date.strftime('%Y-%m-%d')}, Laplacian: {laplacian_value:.4f}")

ftt/ftt.py → fft/fft.py

@@ -37,15 +37,6 @@ def load_data(self, data):
         self.df['log_return'] = np.log(self.df['Close'] / self.df['Close'].shift(1))
         self.df = self.df.dropna()
 
-    def compute_basic_metrics(self):
-        """计算基本指标"""
-        # 计算移动平均
-        self.df['MA21'] = self.df['Close'].rolling(window=21).mean()
-        self.df['MA63'] = self.df['Close'].rolling(window=63).mean()
-        self.df['MA252'] = self.df['Close'].rolling(window=252).mean()
-
-        # 计算波动率
-        self.df['vol_21'] = self.df['log_return'].rolling(window=21).std() * np.sqrt(252)
 
     def perform_fft(self, filter_threshold=None):
         """执行傅里叶变换，可选择性地过滤高频成分"""
@@ -208,34 +199,53 @@ def analyze_with_different_filters(self, periods=[5, 21, 63]):
         plt.legend()
         plt.grid(True)
         return plt
-
+
+    def compute_basic_metrics(self):
+        """计算基本指标，使用FFT滤波替代移动平均"""
+        # 计算21日、63日和252日滤波后的价格序列
+        self.df['FFT21'] = self.filter_high_frequency(cutoff_period=21)
+        self.df['FFT63'] = self.filter_high_frequency(cutoff_period=63)
+        self.df['FFT252'] = self.filter_high_frequency(cutoff_period=252)
+
+        """计算基本指标"""
+        # 计算移动平均
+        self.df['MA21'] = self.df['Close'].rolling(window=21).mean()
+        self.df['MA63'] = self.df['Close'].rolling(window=63).mean()
+        self.df['MA252'] = self.df['Close'].rolling(window=252).mean()
+
+        # 计算波动率
+        self.df['vol_21'] = self.df['log_return'].rolling(window=21).std() * np.sqrt(252)
+
     def get_trading_signals(self):
-        """生成交易信号"""
-        # 基于多个指标综合生成信号
+        """基于FFT滤波生成交易信号"""
         signals = pd.DataFrame(index=self.df.index)
 
-        # 1. 趋势信号（基于移动平均）
-        signals['trend'] = np.where(self.df['MA21'] > self.df['MA63'], 1, -1)
+        # 使用FFT滤波后的价格序列判断趋势
+        signals['trend'] = np.where(self.df['FFT21'] > self.df['FFT63'], 1, -1)
 
-        # 2. 波动率信号
+        # 波动率信号
         vol_mean = self.df['vol_21'].mean()
         signals['volatility'] = np.where(self.df['vol_21'] > vol_mean, 'high', 'low')
 
-        # 3. 相位信号（基于希尔伯特变换）
+        # 相位信号
         analytic_signal = hilbert(self.df['log_return'].values)
         phase = np.angle(analytic_signal)
         phase_diff = np.diff(phase)
-        phase_diff = np.append(phase_diff, phase_diff[-1])  # 补充最后一个值
+        phase_diff = np.append(phase_diff, phase_diff[-1])
         signals['phase'] = np.where(phase_diff > 0, 1, -1)
 
         return signals
 
     def print_analysis_summary(self):
         """打印分析摘要"""
-        # 确保先执行FFT分析和寻找显著周期
+        # 执行FFT分析和寻找显著周期
         self.perform_fft()
         self.find_significant_periods()
 
+        # 计算各周期滤波
+        fft21 = self.filter_high_frequency(cutoff_period=21)
+        fft63 = self.filter_high_frequency(cutoff_period=63)
+
         print("\n=== 股票分析摘要 ===")
 
         # 基本统计
@@ -252,7 +262,7 @@ def print_analysis_summary(self):
 
         # 趋势分析
         print("\n3. 趋势分析:")
-        current_trend = "上升" if self.df['MA21'].iloc[-1] > self.df['MA63'].iloc[-1] else "下降"
+        current_trend = "上升" if fft21.iloc[-1] > fft63.iloc[-1] else "下降"
         print(f"当前趋势: {current_trend}")
 
         # 市场状态