aszerW
908b28473f
## 架构设计 - 三层架构：core(抽象接口) → shared(通用实现) → tests(验证测试) - 5个核心抽象基类：StrategyBase, FactorBase, SignalGenerator, Executor, DataFetcher - 零侵入：与现有框架并行开发，不修改生产代码 ## 已完成 ✓ 核心接口层（5个ABC类） ✓ 通用因子层（MomentumFactor完全复制现有逻辑） ✓ 对比验证测试（新旧因子输出差异=0，测试通过） ## 验证结果 - 最大差异: 0.000000e+00 - 平均差异: 0.000000e+00 - 容差: < 1e-10 ## 下一步 - 阶段3: 信号层迁移（TopNSelector, DynamicThreshold, RebalanceController） - 阶段4: 执行层迁移（BacktestRunner） - 阶段5: 数据层迁移（DataFetcher实现） - 阶段6: 完整策略对比验证 ## 设计原则 - 按需抽象，不预先设计 - 职责分离，避免框架膨胀 - 测试驱动，每个组件必须有对比测试 - 渐进式迁移，验证通过再替换
105 lines
3.2 KiB
Python
105 lines
3.2 KiB
Python
"""
|
||
动量因子（通用版本）
|
||
|
||
使用加权线性回归：得分 = 年化收益率 × R²
|
||
|
||
与现有 MomentumFactor 对比验证：
|
||
- 输入相同 → 输出应该相同
|
||
"""
|
||
|
||
import pandas as pd
|
||
import numpy as np
|
||
import math
|
||
from framework_v2.core import FactorBase
|
||
|
||
|
||
class MomentumFactor(FactorBase):
|
||
"""
|
||
动量因子
|
||
|
||
计算加权线性回归动量得分：
|
||
得分 = 年化收益率 × R²
|
||
|
||
参数：
|
||
- n_days: 动量窗口（默认25）
|
||
- weighted: 是否加权（默认True）
|
||
- crash_filter: 是否启用崩盘过滤（默认True）
|
||
"""
|
||
|
||
name = "momentum"
|
||
category = "momentum"
|
||
|
||
def __init__(
|
||
self,
|
||
n_days: int = 25,
|
||
weighted: bool = True,
|
||
crash_filter: bool = True
|
||
):
|
||
super().__init__(n_days=n_days, weighted=weighted, crash_filter=crash_filter)
|
||
self.n_days = n_days
|
||
self.weighted = weighted
|
||
self.crash_filter = crash_filter
|
||
|
||
def compute(self, data: pd.DataFrame) -> pd.Series:
|
||
"""计算动量因子值"""
|
||
if 'close' not in data.columns:
|
||
raise ValueError("data must contain 'close' column")
|
||
|
||
prices = data['close']
|
||
|
||
if self.weighted:
|
||
factor_values = prices.rolling(self.n_days).apply(
|
||
lambda x: self._weighted_momentum_score(x.values),
|
||
raw=False
|
||
)
|
||
else:
|
||
factor_values = prices.pct_change(self.n_days)
|
||
|
||
if self.crash_filter:
|
||
factor_values = self._apply_crash_filter(prices, factor_values)
|
||
|
||
return factor_values
|
||
|
||
def _weighted_momentum_score(self, prices: np.ndarray) -> float:
|
||
"""计算加权动量得分（完全复制现有逻辑）"""
|
||
if len(prices) < 5:
|
||
return 0.0
|
||
|
||
# 价格下界 clip，防止 log(0) 或 log(负数)
|
||
prices = np.clip(prices, 0.01, None)
|
||
y = np.log(prices)
|
||
|
||
# 异常值检测
|
||
if np.any(np.isnan(y)) or np.any(np.isinf(y)):
|
||
return 0.0
|
||
|
||
x = np.arange(len(y))
|
||
weights = np.linspace(1, 2, len(y))
|
||
|
||
slope, intercept = np.polyfit(x, y, 1, w=weights)
|
||
annualized_returns = math.exp(slope * 250) - 1
|
||
|
||
y_pred = slope * x + intercept
|
||
ss_res = np.sum(weights * (y - y_pred) ** 2)
|
||
ss_tot = np.sum(weights * (y - np.average(y, weights=weights)) ** 2)
|
||
r2 = 1 - ss_res / ss_tot if ss_tot > 0 else 0
|
||
|
||
return annualized_returns * r2
|
||
|
||
def _apply_crash_filter(self, prices: pd.Series, factor_values: pd.Series) -> pd.Series:
|
||
"""崩盘过滤：连续3天跌>5%清零（完全复制现有逻辑）"""
|
||
result = factor_values.copy()
|
||
|
||
for i in range(3, len(prices)):
|
||
r1 = prices.iloc[i] / prices.iloc[i-1]
|
||
r2 = prices.iloc[i-1] / prices.iloc[i-2]
|
||
r3 = prices.iloc[i-2] / prices.iloc[i-3]
|
||
|
||
con1 = min(r1, r2, r3) < 0.95
|
||
con2 = (r1 < 1) and (r2 < 1) and (r3 < 1) and (prices.iloc[i] / prices.iloc[i-3] < 0.95)
|
||
|
||
if con1 or con2:
|
||
result.iloc[i] = 0.0
|
||
|
||
return result
|