Perfplot raised a "TypeError: bench() got an unexpected keyword argument 'logx'". How to fix?

Question

在 SO 上搜索 numpy 数组混合数据类型填充后，我发现了一个不错的小 numpy 数组填充性能测试器 perfplot。当发布 code answer from Nico Schlömer was ran, I saw a dip in the performance chart. So I changed the perflot.show(..snippet..) to perflot.bench(..snippet..) as suggest here 并得到以下错误：

  File "X:\ScriYpts\Z.py", line 40, in <module>
    xlabel='length(a)'
TypeError: bench() got an unexpected keyword argument 'logx'

如何修复？

我的代码：

#!/usr/bin/env python
# -*- coding: utf-8 -*-

import numpy as np
import perfplot


def fill(n):
    a = np.empty(n)
    a.fill(val)
    return a


def colon(n):
    a = np.empty(n)
    a[:] = val
    return a


def full(n):
    return np.full(n, val)


def ones_times(n):
    return val * np.ones(n)


def mlist(n):
    return np.array(n * [val])

val = 42.0


out = perfplot.bench(
    setup=lambda n: n,
    kernels=[fill, colon, full, ones_times, mlist],
    n_range=[2**k for k in range(20)],
    logx=True,
    logy=True,
    xlabel='length(a)'
    )

out.show()

Answer 1

深入了解 perfplot main.py 后，我发现没有 logx' 和 logy **kwargs 可用。

我的解决方案：

out = perfplot.bench(
    setup=lambda n: n,
    kernels=[fill, colon, full, ones_times, mlist],
    n_range=[2**k for k in range(20)],
#    logx=True,  # disabled here
#    logy=True,  # disabled here
    xlabel='length(a)'
    )

out.show(logx=True, logy=True)  # both "log" **kwargs added here in `show()`

由于某些未知原因，xlabel='length(a)' 在 .show() 中不被直接接受为 **kwarg。将它留在 .bench().

中是有效的

Perfplot raised a "TypeError: bench() got an unexpected keyword argument 'logx'". How to fix?

Perfplot raised a "TypeError: bench() got an unexpected keyword argument 'logx'". How to fix?

python

numpy

python-3.x

perfplot