当layout='constrained'时，有没有办法调整子图之间的空间？

Question

我正在尝试删除两个特定轴之间的垂直空间；然而

plt.figure(layout='constrained')

会阻止

matplotlib.pyplot.subplots_adjust(hspace=0)

。

与

layout='constrained'

import matplotlib.pyplot as plt
fig = plt.figure(figsize=(10, 8),layout='constrained')
# fig = plt.figure(figsize=(10, 8))
subfigs = fig.subfigures(2, 1)
axsUp = subfigs[0].subplots(1, 3)
subfigsnest = subfigs[1].subfigures(1, 2, width_ratios=[1, 2])
ax = subfigsnest[0].subplots(1)
axsnest = subfigsnest[1].subplots(2, 1, sharex=True)
subfigsnest[1].subplots_adjust(hspace=0)
plt.show()

• 一般都有次要情节；但ax5和ax6之间有差距

没有

layout='constrained'

import matplotlib.pyplot as plt
# fig = plt.figure(figsize=(10, 8),layout='constrained')
fig = plt.figure(figsize=(10, 8))
subfigs = fig.subfigures(2, 1)
axsUp = subfigs[0].subplots(1, 3)
subfigsnest = subfigs[1].subfigures(1, 2, width_ratios=[1, 2])
ax = subfigsnest[0].subplots(1)
axsnest = subfigsnest[1].subplots(2, 1, sharex=True)
subfigsnest[1].subplots_adjust(hspace=0)
plt.show()

• 所需的ax5和ax6

将

subplots_adjust(hspace=0)

与

layout='constrained'

一起使用不起作用，并且会产生警告 UserWarning：该图使用了与 subplots_adjust 和/或ight_layout 不兼容的布局引擎；不调用 subplots_adjust。.

Answer 1

虽然这个答案不使用

layout='constrained'

，但以下内容可能对大多数人（包括OP）有用。

我建议使用 嵌套

GridSpec

进行设置，其中仅为第二行设置

hspace=0

。这确实确保了所有轴整齐对齐，同时仍然允许

ax5

和

ax6

粘合在一起共享一个公共 x 轴。

import matplotlib.pyplot as plt
from matplotlib.gridspec import GridSpec

fig = plt.figure(figsize=(10, 8))
gs0 = GridSpec(2, 1, figure=fig)  # set up two rows

gs1 = gs0[0].subgridspec(1, 3)    # set up 1st row such that it has 3 columns
ax1 = fig.add_subplot(gs1[0, 0])  # create ax1
ax2 = fig.add_subplot(gs1[0, 1])  # create ax2
ax3 = fig.add_subplot(gs1[0, 2])  # create ax3

gs2 = gs0[1].subgridspec(2, 3, hspace=0)       # set up second row such that it has 2 sub-rows and 3 sub-columns
ax4 = fig.add_subplot(gs2[:, 0])               # create ax4 such that it spans all sub-rows but only 1 sub-column
ax5 = fig.add_subplot(gs2[0, 1:])              # create ax5 in the 1st sub-row but spanning the last 2 sub-columns 
ax6 = fig.add_subplot(gs2[1, 1:], sharex=ax5)  # create ax6 in the 2nd sub-row but spanning the last 2 sub-columns
ax5.tick_params(labelbottom=False)             # remove ax5 x-axis ticks
ax5.set_yticks(ax5.get_yticks()[1:])           # make sure 1st ax5 ticklabel does not overlab with last ax6 ticklabel
ax6.set_yticks(ax6.get_yticks()[:-1])          # make sure last ax6 ticklabel does not overlab with 1st ax5 ticklabel

for i, ax in enumerate([ax1, ax2, ax3, ax4, ax5, ax6]):
    ax.text(0.5, 0.5, f"ax{i+1}", transform=ax.transAxes, ha='center', va='center', fontsize='xx-large')