📜  Python中的 Matplotlib.axes.Axes.pickable()

📅  最后修改于: 2022-05-13 01:54:59.077000             🧑  作者: Mango

Python中的 Matplotlib.axes.Axes.pickable()

Matplotlib是Python中的一个库,它是 NumPy 库的数值数学扩展。 Axes 类包含大部分图形元素:Axis、Tick、Line2D、Text、Polygon 等,并设置坐标系。 Axes 的实例通过回调属性支持回调。

matplotlib.axes.Axes.pickable()函数

matplotlib 库的 axes 模块中的Axes.pickable()函数用于返回艺术家是否可挑选。

下面的示例说明了 matplotlib.axes 中的 matplotlib.axes.Axes.pickable()函数:

示例 1:

# Implementation of matplotlib function
import numpy as np
np.random.seed(19680801)
import matplotlib.pyplot as plt
   
volume = np.random.rayleigh(27, size = 40)
amount = np.random.poisson(10, size = 40)
ranking = np.random.normal(size = 40)
price = np.random.uniform(1, 10, size = 40)
   
fig, ax = plt.subplots()
   
scatter = ax.scatter(volume * 2, amount * 3,
                     c = ranking * 3, 
                     s = 0.3*(price * 3)**2,
                     vmin = -4, vmax = 4, 
                     cmap = "Spectral")
  
legend1 = ax.legend(*scatter.legend_elements(num = 5),
                    loc ="upper left",
                    title ="Ranking")
  
ax.add_artist(legend1)
  
ax.text(60, 30, "Value return : " + str(ax.pickable()), 
        fontweight ="bold", 
        fontsize = 18)
   
fig.suptitle('matplotlib.axes.Axes.pickable() function\
Example', fontweight ="bold")
  
plt.show()

输出:

示例 2:

import numpy as np
import matplotlib.pyplot as plt
import matplotlib.cbook as cbook
   
np.random.seed(10**7)
data = np.random.lognormal(size =(10, 4),
                           mean = 4.5,
                           sigma = 4.75)
  
labels = ['G1', 'G2', 'G3', 'G4']
   
result = cbook.boxplot_stats(data, 
                             labels = labels, 
                             bootstrap = 1000)
   
for n in range(len(result)):
    result[n]['med'] = np.median(data)
    result[n]['mean'] *= 0.1
  
fig, axes1 = plt.subplots()
axes1.bxp(result)
  
axes1.text(2, 30000,
           "Value return : " + str(axes1.pickable()), 
           fontweight ="bold")
   
fig.suptitle('matplotlib.axes.Axes.pickable() function \
Example', fontweight ="bold")
  
plt.show()

输出: