在现有的Cairo图像表面/上下文上覆盖matplotlib命令

Question

我有一个从SVG文件创建的图表，并使用cairosvg库加载。

我想通过直接matplotlib命令以编程方式定义额外的重叠图来进一步增强此图。

我正在尝试在内存中执行此操作而不写出文件并重新加载。仅保存最终文件。

但是，将现有的cairo与matplotlib渲染的cairo 组合起来很难实现。我认为this was a potential solution，但不确定。

简化示例：

import matplotlib
matplotlib.use('Cairo') # Make sure here that our figures will use Cairo!
from cairosvg import surface
from cairosvg import parser
import matplotlib.pyplot as plt
import matplotlib.image as img 
import numpy as np

SVG='\
<svg version="1.1" xmlns="http://www.w3.org/2000/svg" width="500" height="500">\
<rect id="arena" x="10" y="10" width="480" height="480" fill="#ff5555" fill-opacity="0.5"/>\
</svg>\
'

svg = parser.Tree(bytestring=SVG)

surf = surface.PNGSurface(svg,None,1)
#surf.cairo.write_to_png("test-svg.png") # Can do this to see svg render to file
surf.finish() # Required?
ctx = surf.context # Can get the context here fine..

fig = plt.imshow(ctx) # <----- MAGIC REQUIRED HERE ABOUTS
#fig = plt.figure() # This works like a normal plot - but no svg

ax = fig.gca() # Create some overlay data
x = np.array(np.random.randn(100))
y = np.array(np.random.randn(100))
ax.plot(x,y)

plt.savefig("test-final.png") # Save svg & overlay plot

编辑：也在尝试以下内容，基于前面的链接。尽管如此，仍然没有快乐。

from matplotlib.artist import Artist

class SurfArtist(Artist):
  def __init__(self,surf):
    Artist.__init__(self)
    self._surf = surf

  def draw(self,renderer):
    from matplotlib.backends.backend_cairo import RendererCairo
    if not isinstance(renderer, RendererCairo):
      raise TypeError("backend not supported")
    ctx = renderer.gc.ctx if hasattr(renderer, "gc") else renderer.ctx
    self._surf.context = ctx
    self._surf.cairo.show_page()

...

fig = plt.figure()
ax = fig.gca()
surf_artist = SurfArtist(surf)
surf_artist.set_zorder(float('inf'))
ax.artists.append(surf_artist)

Answer 1

下面的代码应该可以正常工作，首先将表面转换为numpy数组，如https://github.com/Kozea/CairoSVG/issues/57

matplotlib.use('Cairo') # Make sure here that our figures will use Cairo!
from cairosvg import surface
from cairosvg import parser
import matplotlib.pyplot as plt
import matplotlib.image as img 
import numpy as np

# function to convert to array
def surface_to_npim(surface):
    """ Transforms a Cairo surface into a numpy array. """
    im = +np.frombuffer(surface.get_data(), np.uint8)
    H,W = surface.get_height(), surface.get_width()
    im.shape = (H,W, 4) # for RGBA
    return im[:,:,:3]


SVG='\
<svg version="1.1" xmlns="http://www.w3.org/2000/svg" width="500" height="500">\
<rect id="arena" x="10" y="10" width="480" height="480" fill="#ff5555" fill-opacity="0.5"/>\
</svg>\
'

svg = parser.Tree(bytestring=SVG)

surf = surface.PNGSurface(svg,None,1)
#surf.cairo.write_to_png("test-svg.png") # Can do this to see svg render to file
surf.finish() # Required?


# convert svg and plot
img = surface_to_npim(surf.cairo)
plt.imshow(img)

# Create some overlay data
x = np.array(np.random.randn(100))
y = np.array(np.random.randn(100))
plt.plot(x,y)

plt.savefig("test-final.png") # Save svg & overlay plot