这是有可能的,但需要花费大量的工作。一种可能的解决方案如下所示。它涉及绘制6个不同的热图,并调整间距使其看起来还可以。然后还需要同步颜色缩放并手动设置色条。
import matplotlib
import matplotlib.pyplot as plt
import numpy as np; np.random.seed(0)
import seaborn as sns; sns.set()
data = np.random.rand(10, 12)
asp = data.shape[0]/float(data.shape[1])
figw = 8
figh = figw*asp
cmap = plt.cm.copper
norm = matplotlib.colors.Normalize(vmin= data.min(), vmax= data.max())
gridspec_kw = {"height_ratios":[9,1], "width_ratios" : [4,5,3]}
heatmapkws = dict(square=False, cbar=False, cmap = cmap, linewidths=1.0, vmin= data.min(), vmax= data.max() )
tickskw = dict(xticklabels=False, yticklabels=False)
left = 0.07; right=0.87
bottom = 0.1; top = 0.9
fig, axes = plt.subplots(ncols=3, nrows=2, figsize=(figw, figh), gridspec_kw=gridspec_kw)
plt.subplots_adjust(left=left, right=right,bottom=bottom, top=top, wspace=0.1, hspace=0.1*asp )
sns.heatmap(data[:9,0:4], ax=axes[0,0], xticklabels=False, yticklabels=True, **heatmapkws)
sns.heatmap(data[:9,4:9], ax=axes[0,1], xticklabels=False, yticklabels=False, **heatmapkws)
sns.heatmap(data[:9,9:12], ax=axes[0,2],xticklabels=False, yticklabels=False, **heatmapkws)
sns.heatmap(data[9:,:4], ax=axes[1,0], xticklabels=True, yticklabels=True, **heatmapkws)
sns.heatmap(data[9:,4:9], ax=axes[1,1], xticklabels=True, yticklabels=False, **heatmapkws)
sns.heatmap(data[9:,9:12], ax=axes[1,2], xticklabels=True, yticklabels=False,**heatmapkws)
axes[1,0].set_yticklabels([9])
axes[1,1].set_xticklabels([4,5,6,7,8])
axes[1,2].set_xticklabels([9,10,11])
cax = fig.add_axes([0.9,0.1,0.03,0.8])
sm = matplotlib.cm.ScalarMappable(cmap=cmap, norm=norm)
sm.set_array([])
fig.colorbar(sm, cax=cax)
plt.show()
for
循环而不是复制调用sns.heatmap
。) - mwaskom