您可以将数据以原始形式提供给Matplotlib，并为每个轴指定尺度以指定转换：对于x轴，内置的对数尺度就足够了;对于y轴，您可以创建自定义尺度。在这里，我通过定义y_forward和y_inverse来实现这一点。

import numpy as np
import matplotlib.pyplot as plt
from matplotlib.ticker import ScalarFormatter, MaxNLocator

#data
failure = np.array([168, 202, 190, 197, 169, 214, 201, 219, 206, 198, 190, 183, 206, 218, 214, 206, 213, 202, 209, 206])

# Define forward and inverse transformations for the y-axis
def y_forward(val):
    return np.log(-np.log(1 - val))

def y_inverse(val):
    return 1 - np.exp(-np.exp(val))

#get plotting position for pprbablity distribution plots
def plottingPositionOriginal(failure, a=0.3):
    x = np.sort(failure)
    n = len(x)
    F = []
    for i in range(1, len(failure)+1):
        F.append((i - a) / (n + 1 - 2 * a))
    
    y = np.array(F)
    return x, y

# transform the y axis so the plot appears linear
def plottingPositionFit(failure):
    t, F = plottingPositionOriginal(failure)
    x = np.log(t)
 
    y = y_forward(F)
    beta, c = np.polyfit(x, y, 1)
    y_fit = beta * x + c
    F_fit = y_inverse(y_fit)

    return t, F, F_fit

X, Y, y_hat = plottingPositionFit(failure)

fig, ax = plt.subplots()
ax.scatter(X, Y) 
ax.set_xscale("log")
ax.xaxis.set_minor_formatter(ScalarFormatter())
ax.set_yscale("function", functions=(y_forward, y_inverse))
ax.yaxis.set_major_locator(MaxNLocator(10))

ax.plot(X, y_hat) 

plt.show()