python - Plotnine 组 facet_grid 标签

Question

使用 plotnine 的facet_grid功能时有没有办法对标签进行分组？

例如，采用以下 DataFrame（我制作了示例，以便您可以重现我的问题。实际上，我的 DataFrame 要大得多，但结构几乎相同）：

df_test = pd.DataFrame(data= {'rate_1': [0.0,0.1,0.2,0.3,0.0,0.1,0.2,0.3,0.3,0.2,0.1,0.0],
                              'rate_2': [0.0,0.1,0.2,0.3,0.0,0.1,0.2,0.3,0.3,0.2,0.1,0.0],
                              'rate_3':[0.0,0.1,0.2,0.3,0.0,0.1,0.2,0.3,0.3,0.2,0.1,0.0],
                              'rate_4': [0.0,0.1,0.2,0.3,0.1,0.2,0.3,0.0,0.2,0.3,0.0,0.1],
                              'samples': [50000, 100000, 50000, 100000,50000, 100000, 50000, 100000,50000, 100000, 50000, 100000],
                              'model': ['model 1', 'model 2', 'model 3', 'model 4','model 1', 'model 2', 'model 3', 'model 4','model 1', 'model 2', 'model 3', 'model 4'],
                              'metric': [0.5,0.4,0.3,0.2,0.5,0.4,0.3,0.2,1,0.3,0.4,0.3]})

我创建了以下函数来根据速率、样本量和模型绘制绝对度量：

# Plotting
my_plot = ggplot(data=df_test, mapping=aes('rate_1', 'rate_2', fill='metric'))
my_plot += geom_tile()
my_plot += theme(axis_text_x=element_text(rotation=90), figure_size=(16, 8), strip_background_x=element_text(width=1.))
my_plot += scale_x_continuous(breaks=[.0, .1, .2, .3, .4, .5])
my_plot += scale_y_continuous(breaks=[.0, .1, .2, .3, .4, .5])
my_plot += facet_grid('rate_3 + samples  ~ model + rate_4', labeller=label_value)
my_plot += scale_fill_gradient2(midpoint=0, low='blue', mid="white", high="red")
my_plot

结果图如下所示：

在标记 facet_grid 时，有没有办法对模型名称（模型 1、模型 2、模型 3、模型 4）和 rate_4 进行分组？我正在寻找一个结果，使得“列”的名称如下所示（我使用 Excel 进行说明）：

谢谢！

Answer 1

多亏了这个提示，我能够通过操纵 matplotlib 文本和背景对象来“分组”标签：

# Plotting
my_plot = ggplot(data=df_test, mapping=aes('rate_1', 'rate_2', fill='metric'))
my_plot += geom_tile()
my_plot += theme(axis_text_x=element_text(rotation=90), figure_size=(16, 8), strip_background_x=element_text(width=1.))
my_plot += scale_x_continuous(breaks=[.0, .1, .2, .3, .4, .5])
my_plot += scale_y_continuous(breaks=[.0, .1, .2, .3, .4, .5])
my_plot += facet_grid('rate_3 + samples  ~ model + rate_4', labeller=label_value)
my_plot += scale_fill_gradient2(midpoint=0, low='blue', mid="white", high="red")

fig = my_plot.draw()
fig._themeable['strip_text_x'][0].set_text('\n 0.0')
fig._themeable['strip_text_x'][1].set_text('model 1 \n 0.10')
fig._themeable['strip_text_x'][2].set_text('\n 0.20')
fig._themeable['strip_background_x'][0].set_width(3.0)

fig._themeable['strip_text_x'][3].set_text('\n 0.00')
fig._themeable['strip_text_x'][4].set_text('model 2 \n 0.10')
fig._themeable['strip_text_x'][5].set_text('\n 0.20')
fig._themeable['strip_background_x'][3].set_width(3.0)

fig._themeable['strip_text_x'][6].set_text('\n 0.00')
fig._themeable['strip_text_x'][7].set_text('model 3 \n 0.10')
fig._themeable['strip_text_x'][8].set_text('\n 0.20')
fig._themeable['strip_background_x'][6].set_width(3.0)

fig._themeable['strip_text_x'][9].set_text('\n 0.00')
fig._themeable['strip_text_x'][10].set_text('model 4 \n 0.10')
fig._themeable['strip_text_x'][11].set_text('\n 0.20')
fig._themeable['strip_background_x'][9].set_width(3.0)

plt.show()

结果如下所示：