如何获取 Matplotlib 补丁的坐标并使用它来添加新轴？

Question

我使用 fig = plt.figure() 和 ax = fig.add_subplot(my_arguments) 创建了一个图形和坐标轴。然后我使用 matplotlib.patches 添加了一些补丁。我通过使用 matplotlib.transforms.Affine2D() 在数据坐标中平移和旋转来转换每个补丁，然后通过将 ax.transData() 添加到我的 Affine2D 的末尾来将转换后的坐标转换为显示坐标转换。

这是代码的简化版本：

import matplotlib as mpl
import matplotlib.patches as patches
from matplotlib.transforms import Bbox

fig = plt.figure()
ax = fig.add_subplot(111)
# plot anything here
ax.plot(range(10), 'ro')
my_patches = []
# in my code there many patches and therefore the line
# below is actually a list comprehension for each one
my_patches.append(
    patches.Rectangle( (1, 2), 10, 20, transform=mpl.transforms.Affine2D() \
    .translate(1, 1) \
    .rotate_deg_around(1, 2, 35)
    + ax.transData, fill=False, color='blue')
    )
# now add a new axis using the coordinates of the patch
patch = my_patches[0]
# get the coords of the lower left corner of the patch
left, bottom = patch.get_xy()
# get its width and height
width, height = patch.get_width(), patch.get_height()
# create a Bbox instance using the coords of the patch
bbox = Bbox.from_bounds(left, bottom, width, height)
# transform from data coords to display coords
disp_coords = ax.transData.transform(bbox)
# transform from display coords to figure coords
fig_coords = fig.transFigure.inverted().transform(disp_coords)
# new axis
ax2 = fig.add_axes(Bbox(fig_coords))
# plot anything else here
ax2.plot(range(10), 'bo')

但是，附加轴并未添加到图形中与补丁的变换坐标相同的位置（尽管它们很接近）。我错过了什么吗？

Answer 1

我不确定这段代码的用途是什么，所以这可能不是您想要的。但是为了让轴框出现在坐标 (1,2) 处，您可能应该先绘制 canvas，然后再使用从补丁中获得的坐标。

...
fig.canvas.draw()
left, bottom = patch.get_xy()
...

如何获取 Matplotlib 补丁的坐标并使用它来添加新轴？

How can I get the coordinates of a Matplotlib patch and use it to add a new axis?

python

patch

transformation

matplotlib