Question

受Matplotlib here中这个例子的启发，我想制作一个类似的例子.我遇到了以下问题:How can I relate the coordinates of the matplotlin graphics with my real world coordinates?

Example

用下图说明:

• 我需要inset_axis相当精确地位于红色点
• 我已经测试了所有四个角的left_inset_ax = fig.add_axes([1, 0, .05, .05], facecolor='k')个(参见黑色方块)
• 在有色区域内，我可以使用局部坐标(0. 1、0.. 1)，但我没有找到与外部坐标的关系
• phersilk我可以设置myCoordinates=True
• I have also tried this example here, but I this is hand taylored, whereas I need an automated solution without hard coded figures.

我有不同的types of 100 graphics.其中之一是下面的.它们是在polar 设置中生成的

• plt.plot(DF_histo_cumsum.index，DF_histo_cumsum，c =' k '，lw=0.7)，或通过
• b = ax0.bar(np.deg2rad(self.WD_cc)，Nbar[:，j]，Label=self.WD_cc，Color=self.ws_colors[j-1])

So they live in a different world of coordinates that is not easy to access for direct scaling.

Solutions and results

只要图表非常简单，给定的here看起来就有希望.我以它为基础添加更多功能，例如标题、x和y标签、等距非方形设置、不同数量的网格点.该方法是

• 生成2维数据

• Select 其中的几个inset_axis

• 转换为本地(0. 1)双向坐标

• 显示inset_axis

• 使用不同数量的网格点

   import numpy as np
   import matplotlib.pyplot as plt
  
    def scaleXY(xgr,ygr,xT,yT): 
       xgr = (xgr-xT.min())/(xT.max()-xT.min())
       ygr = (ygr-yT.min())/(yT.max()-yT.min())
       return xgr, ygr
  
   nx, ny = 150, 60                                            # choose: number of data
   ix = np.linspace(-5, 20, nx)*100
   iy = np.linspace(-1, 15, ny)*100
   xq, yq = np.meshgrid(ix,iy, indexing='ij')                   # generate data
   dx, dy = 50, 30                                              # choose
   xp,yp = xq[::dx,::dy], yq[::dx,::dy]                         # select a few of them
  
   xgr, ygr = scaleXY(xp,yp,xq,yq)                              # scale them to (0..1)
  
   with plt.style.context('fast'): 
           fig = plt.figure(figsize=(10,10)) 
           ax1 = fig.add_subplot(111)
  
           ax1.scatter(xq,yq, s=5)                               # plot the base data
           ax1.scatter(xp,yp, c='lime', s=300, alpha = 0.6)      # plot the testing points
  
           ins = ax1.inset_axes([xgr[2,1], ygr[2,1],  0.2,0.2])  # check the inset positioning
           ins = ax1.inset_axes([xgr[1,1], ygr[1,1],  0.2,0.2])  # check the inset positioning
           ins = ax1.inset_axes([xgr[0,0], ygr[0,0],  0.2,0.2])  # check the inset positioning
           ins = ax1.inset_axes([xgr[2,0], ygr[2,0],  0.2,0.2])  # check the inset positioning
  
           ax1.set_aspect('equal')
           title = 'Even more dangerous area'
           ax1.set_title(title,fontweight='bold', fontsize=17)
           ax1.set_xlabel('x-direction', fontsize=17)
           ax1.set_ylabel('y-direction', fontsize=17)        
  
       plt.show() 
  

作为results，可以看到图形的坐标与用户的坐标分开.这就引出了一个问题:How can I relate the coordinates of the matplotlin graphics with my real world coordinates?

I would appreciate any hints to other running solutions - thank you!