如何以优雅的Pytest方式避免测试代码重复？

针对你提到的两个测试用例核心逻辑重复、仅路径标签和断言预期不同的问题，可以用以下几种Pytest原生方式优化：

方案一：类级参数化（推荐，最简洁） #

直接用@pytest.mark.parametrize装饰测试类，把差异项（测试列表键、断言预期值）作为参数传入，一个测试类即可覆盖两种场景：

import pytest
import yaml

VERSION_TAG = "your_version_tag"
THRESHOLD_COMPREHENSION = [0.05, 0.1]
WINDOW_SIZE_COMPREHENSION = [10, 20]

@pytest.fixture(scope='session')
def load_config() -> dict:
    with open(r"test_plan_1.yaml") as f:
        return yaml.safe_load(f)

class TestBase:
    def calculate_mape_range(self, test_path_1: str, test_path_2: str, window_size: int, threshold: float) -> int:
        # 这里实现你的MAPE计算逻辑
        return 0

# 用类参数化区分健康/故障测试场景
@pytest.mark.parametrize(
    "test_list_key, expected_failures",
    [
        ("healthy_test_list", 0),    # 健康场景预期0次连续失败
        ("faulty_test_list", 3),     # 故障场景预期非0（根据实际需求调整）
    ]
)
@pytest.mark.parametrize("threshold", THRESHOLD_COMPREHENSION)
@pytest.mark.parametrize("window_size", WINDOW_SIZE_COMPREHENSION)
class TestMape:
    def test_mape_validation(self, threshold: float, window_size: int, load_config: dict, test_list_key: str, expected_failures: int):
        # 从配置中获取对应测试文件路径
        test_files = load_config['test_plan']['test_ids'][VERSION_TAG]['tools']['test_file_ids'][test_list_key]
        # 调用公共计算方法
        consecutive_failures = TestBase().calculate_mape_range(
            test_files[0],
            test_files[1],
            window_size,
            threshold
        )
        # 断言结果
        assert consecutive_failures == expected_failures

方案二：子类继承+类属性（保留原类结构） #

如果需要保留原有的TestMapeHealthy和TestMapeFaulty类结构，可以把公共测试逻辑抽离到父类，子类通过类属性指定差异项：

import pytest
import yaml

VERSION_TAG = "your_version_tag"
THRESHOLD_COMPREHENSION = [0.05, 0.1]
WINDOW_SIZE_COMPREHENSION = [10, 20]

@pytest.fixture(scope='session')
def load_config() -> dict:
    with open(r"test_plan_1.yaml") as f:
        return yaml.safe_load(f)

class TestBase:
    def calculate_mape_range(self, test_path_1: str, test_path_2: str, window_size: int, threshold: float) -> int:
        # 实现MAPE计算逻辑
        return 0

# 父类定义公共测试逻辑
class TestMapeBase(TestBase):
    # 子类需要覆盖这两个属性
    test_list_key = ""
    expected_failures = 0

    @pytest.mark.parametrize("threshold", THRESHOLD_COMPREHENSION)
    @pytest.mark.parametrize("window_size", WINDOW_SIZE_COMPREHENSION)
    def test_mape(self, threshold: float, window_size: int, load_config: dict):
        test_files = load_config['test_plan']['test_ids'][VERSION_TAG]['tools']['test_file_ids'][self.test_list_key]
        consecutive_failures = self.calculate_mape_range(
            test_files[0],
            test_files[1],
            window_size,
            threshold
        )
        assert consecutive_failures == self.expected_failures

# 健康场景测试类
class TestMapeHealthy(TestMapeBase):
    test_list_key = "healthy_test_list"
    expected_failures = 0

# 故障场景测试类
class TestMapeFaulty(TestMapeBase):
    test_list_key = "faulty_test_list"
    expected_failures = 3  # 根据实际需求调整预期值

额外优化点 #

• 去掉原代码中的try块：Pytest会自动捕获断言异常并生成详细的测试报告，手动try反而会掩盖错误细节。
• 简化配置路径读取：可以把重复的配置路径抽成fixture，进一步减少重复代码：

@pytest.fixture(scope='session')
def test_file_ids(load_config):
    return load_config['test_plan']['test_ids'][VERSION_TAG]['tools']['test_file_ids']

之后在测试方法中直接使用test_file_ids[test_list_key]即可。

内容的提问来源于stack exchange，提问作者Efrat.shp