执行时间（性能）装饰器的单元测试-解网

问：

我有一个执行时间（性能）装饰器，它计算给定函数的平均执行时间并使用值打印结果：float

from time import perf_counter, sleep

def benchmark_decorator(func):

    def wrapper(*args, **kwargs):
        results = list()
        n_repeats = 3
        for i in range(n_repeats):
            time_start = perf_counter()
            result = func(*args, **kwargs)
            time_end = perf_counter()
            time_duration = time_end - time_start
            results.append(time_duration)
            print(f'>run {i+1} took {time_duration} seconds')
        avg_duration = sum(results) / n_repeats
        print(f'Took {avg_duration} seconds on average')
        return result
    
    return wrapper

如何测试此装饰器是否正确计算平均时间？例如，让我们检查一下这个简单的函数：

@benchmark_decorator
def func(seconds):
    sleep(seconds)

func(1)

我们有这样的输出：

>run 1 took 1.0005083000287414 seconds
>run 2 took 1.0000609999988228 seconds
>run 3 took 1.000284199952148 seconds
Took 1.0002844999932374 seconds on average

所以我每次得到的结果都略有不同，不能只用一个数字来断言这个结果。此外，我不能只用 example，因为函数可以非常快，我们将得到非常低的时间值：.3f

@benchmark_decorator
def func():
    pass

func()

给出以下输出：

>run 1 took 4.400033503770828e-06 seconds
>run 2 took 1.500011421740055e-06 seconds
>run 3 took 5.00003807246685e-07 seconds
Took 2.133349577585856e-06 seconds on average

如果我们用这个值四舍五入，我们将得到零。那么如何正确地测试这个略有不同的时间呢？.3f

python 单元测试日期时间时间浮点

import pytest

@benchmark_decorator
def wrap_sleep(seconds):
    sleep(seconds)

def test_benchmark_decorator(capsys):
    wrap_sleep(1)
    captured = capsys.readouterr()
    average = float(captured.out.rsplit(" ", 4)[1])
    # Rounding
    assert average == pytest.approx(1, abs=1e-3)
    # Relative 2% error
    assert average == pytest.approx(1, rel=0.02)

执行时间（性能）装饰器的单元测试

Unit tests for execution time (perfomance) decorator

评论

评论