标签的用途是什么＆＃39; matplotlib箱形图中的属性？

Question

我使用matplotlib的箱形图并尝试使用标签为箱面分配颜色;使用类似于以下内容的代码：

from matplotlib import pyplot as plt
import numpy as np

data = {}
data['a'] = np.arange(12)+1
data['b'] = np.arange(14)+1
data['c'] = np.arange(8)+1

color_dict = {'trt_a':'red', 'trt_b':'blue', 'trt_c':'green'}
controls = ['trt_a', 'trt_b', 'trt_c']

fig, ax = plt.subplots()

boxplot_dict = ax.boxplot([data[x] for x in ['a', 'b', 'c']], \
    positions = [1, 1.5, 2], labels = controls, \
    patch_artist = True, widths = 0.25)

for b in boxplot_dict['boxes']:
    lab = b.get_label()
    print("Label property of box ", i, " is ", lab)
    b.set_facecolor(color_dict[lab]) # <- This is what I would like to be able to do

ax.set_ylim([0,16])
plt.show()

我理解使用labels作为输入参数来命名不同的箱图。但是，我注意到使用get_label方法访问时框的label属性未设置为任何内容。

那么label属性（不是输入参数）有什么用？

Answer 1

labels中的ax.boxplot()参数设置轴的xticklabels，而不是框对象的标签（将在ax.legend()中使用）。

相反，您可以从ax.get_xticklabels访问所需的标签。要在代码中替换的相关代码段为：

for b in boxplot_dict['boxes']:
    lab = ax.get_xticklabels()[i].get_text()
    print("Label property of box {0} is {1}".format(i, lab))
    b.set_facecolor(color_dict[lab]) 
    i += 1

完整的代码是：

from matplotlib import pyplot as plt
import numpy as np

data = {}
data['a'] = np.arange(12)+1
data['b'] = np.arange(14)+1
data['c'] = np.arange(8)+1

color_dict = {'trt_a':'orange', 'trt_b':'blue', 'trt_c':'green'}
controls = ['trt_a', 'trt_b', 'trt_c']

fig, ax = plt.subplots()

boxplot_dict = ax.boxplot(
    [data[x] for x in ['a', 'b', 'c']],
    positions = [1, 1.5, 2],
    labels = controls, 
    patch_artist = True,
    widths = 0.25)

i=0
for b in boxplot_dict['boxes']:
    lab = ax.get_xticklabels()[i].get_text()
    print("Label property of box {0} is {1}".format(i, lab))
    b.set_facecolor(color_dict[lab]) 
    i += 1

ax.set_ylim([0,16])
plt.show()

替代示例，设置框对象的label属性

from matplotlib import pyplot as plt
import numpy as np

data = {}
data['a'] = np.arange(12)+1
data['b'] = np.arange(14)+1
data['c'] = np.arange(8)+1

color_dict = {'trt_a':'orange', 'trt_b':'blue', 'trt_c':'green'}
controls = ['trt_a', 'trt_b', 'trt_c']

fig, ax = plt.subplots()

boxplot_dict = ax.boxplot(
    [data[x] for x in ['a', 'b', 'c']],
    positions = [1, 1.5, 2],
    labels = controls, 
    # patch_artist = True,  # Legend isn't working with patch objects.. 
    widths = 0.25)

i=0
for b in boxplot_dict['boxes']:
    lab = ax.get_xticklabels()[i].get_text()
    print("Label property of box {0} is {1}".format(i, lab))
    b.set_color(color_dict[lab])
    b.set_label(lab)
    i += 1

ax.set_ylim([0,16])
ax.legend()
plt.show()