liaozhaorun
c098631487
- 新增 LocalFactorEvaluator 类封装 FactorEngine，提供 (M,T) 矩阵输出 - evaluate_factors_with_evaluator() 支持新评估方式 - ValidationPipeline 优先使用 evaluator 计算信号 - 新增测试文件验证功能
131 lines
4.1 KiB
Python
131 lines
4.1 KiB
Python
"""Tests for Factorminer pipeline integration with LocalFactorEvaluator."""
|
|
|
|
from __future__ import annotations
|
|
|
|
from typing import Dict, List, Optional, Tuple
|
|
|
|
import numpy as np
|
|
import pytest
|
|
|
|
from src.factorminer.core.factor_library import FactorLibrary
|
|
from src.factorminer.core.library_io import import_from_paper
|
|
from src.factorminer.evaluation.local_engine import LocalFactorEvaluator
|
|
from src.factorminer.evaluation.pipeline import (
|
|
PipelineConfig,
|
|
ValidationPipeline,
|
|
run_evaluation_pipeline,
|
|
)
|
|
from src.factorminer.evaluation.runtime import (
|
|
evaluate_factors_with_evaluator,
|
|
)
|
|
|
|
|
|
class TestLocalFactorEvaluatorIntegration:
|
|
"""测试 LocalFactorEvaluator 与评估管线的集成。"""
|
|
|
|
@pytest.fixture
|
|
def evaluator(self) -> LocalFactorEvaluator:
|
|
"""创建评估器 fixture。"""
|
|
return LocalFactorEvaluator(
|
|
start_date="20200101",
|
|
end_date="20200131",
|
|
stock_codes=None,
|
|
)
|
|
|
|
@pytest.fixture
|
|
def returns_matrix(self) -> np.ndarray:
|
|
"""创建模拟收益率矩阵 fixture。"""
|
|
M, T = 100, 20
|
|
rng = np.random.default_rng(42)
|
|
return rng.standard_normal((M, T))
|
|
|
|
@pytest.fixture
|
|
def splits(self) -> Dict[str, object]:
|
|
"""创建模拟分割 fixture。"""
|
|
|
|
class MockSplit:
|
|
def __init__(self, indices: np.ndarray, returns: np.ndarray):
|
|
self.indices = indices
|
|
self.returns = returns
|
|
self.target_returns = {}
|
|
|
|
T = 20
|
|
indices = np.arange(T)
|
|
rng = np.random.default_rng(42)
|
|
returns = rng.standard_normal((100, T))
|
|
|
|
return {
|
|
"train": MockSplit(indices[:15], returns[:, :15]),
|
|
"val": MockSplit(indices[15:], returns[:, 15:]),
|
|
}
|
|
|
|
def test_evaluate_factors_with_evaluator_deprecated_path(
|
|
self,
|
|
evaluator: LocalFactorEvaluator,
|
|
returns_matrix: np.ndarray,
|
|
splits: Dict[str, object],
|
|
) -> None:
|
|
"""测试 evaluate_factors_with_evaluator 在有 evaluator 时的行为。"""
|
|
|
|
# 模拟一个因子对象
|
|
class MockFactor:
|
|
def __init__(self, id: str, name: str, formula: str, category: str):
|
|
self.id = id
|
|
self.name = name
|
|
self.formula = formula
|
|
self.category = category
|
|
|
|
factors = [
|
|
MockFactor("f1", "close", "close", "price"),
|
|
MockFactor("f2", "# TODO: unsupported", "unsupported", "test"),
|
|
]
|
|
|
|
try:
|
|
artifacts = evaluate_factors_with_evaluator(
|
|
factors=factors,
|
|
evaluator=evaluator,
|
|
returns=returns_matrix,
|
|
splits=splits,
|
|
)
|
|
# 验证返回结果结构
|
|
assert len(artifacts) == 2
|
|
assert artifacts[0].name == "close"
|
|
assert artifacts[1].name == "# TODO: unsupported"
|
|
# unsupported 因子应该被标记
|
|
assert artifacts[1].error == "Unsupported operator in formula"
|
|
except Exception as e:
|
|
# FactorEngine 可能因为数据不存在而失败
|
|
pytest.skip(f"FactorEngine 数据不存在: {e}")
|
|
|
|
def test_evaluate_factors_fallback_legacy(
|
|
self,
|
|
returns_matrix: np.ndarray,
|
|
splits: Dict[str, object],
|
|
) -> None:
|
|
"""测试 evaluator=None 时回退到 legacy 方式。"""
|
|
|
|
class MockFactor:
|
|
def __init__(self, id: str, name: str, formula: str, category: str):
|
|
self.id = id
|
|
self.name = name
|
|
self.formula = formula
|
|
self.category = category
|
|
|
|
factors = [
|
|
MockFactor("f1", "test", "close", "price"),
|
|
]
|
|
|
|
# evaluator=None 应该回退到 legacy
|
|
artifacts = evaluate_factors_with_evaluator(
|
|
factors=factors,
|
|
evaluator=None,
|
|
returns=returns_matrix,
|
|
splits=splits,
|
|
)
|
|
# Legacy 方式会尝试 compute_tree_signals 但 data_dict 为空
|
|
assert len(artifacts) == 1
|
|
|
|
|
|
if __name__ == "__main__":
|
|
pytest.main([__file__, "-v"])
|