Activity

1. lclanzi commented on Jan 14, 2020

   

   lclanzi

   on Jan 14, 2020

   Author

   More actions

   I provide an example. The first figure (etaGainH.png) is the one I save from python with plt.savefig('etaGainH.png'). The second figure is the one I get when I use plt.savefig('etaGainH.eps').

   
   etaGainH-eps-converted-to.pdf

   This is the code I'm using:

   fig, ax = plt.subplots(1,1,figsize=(5.3,4.3))
   X,Y = np.meshgrid(hstar,invparam,indexing='ij')
   cmap = mpl.cm.get_cmap('viridis') ;\
   plt.pcolormesh(X,Y,etaGain*100,cmap=cmap,shading='gouraud');\
   cbar = plt.colorbar()
   cbar.set_label(r'$\mathcal{G}_{\eta}\;[\%]$',fontsize = 14)
   plt.ylim([0.5,1.5])
   plt.xlim([0.16,0.4])
   plt.savefig('etaGainH.eps',format='eps',bbox_inches='tight',dpi=300)

   Note that if I use shading='flat' instead of shading='gouraud', this is what I get

   
   etaGainHFlat-eps-converted-to.pdf

   Code

   fig, ax = plt.subplots(1,1,figsize=(5.3,4.3))
   X,Y = np.meshgrid(hstar,invparam,indexing='ij')
   cmap = mpl.cm.get_cmap('viridis') ;\
   plt.pcolormesh(X,Y,etaGain*100,cmap=cmap,shading='flat');\
   cbar = plt.colorbar()
   cbar.set_label(r'$\mathcal{G}_{\eta}\;[\%]$',fontsize = 14)
   plt.ylim([0.5,1.5])
   plt.xlim([0.16,0.4])
   plt.savefig('etaGainH.eps',format='eps',bbox_inches='tight',dpi=300)

   The problem seems to be related to this particular shading option.
   Hope that the issue is more clear.
2. anntzer

   added

   backend: ps

   status: needs confirmation

   on Jan 14, 2020
3. anntzer commented on Jan 14, 2020

   

   anntzer

   on Jan 14, 2020

   Contributor

   More actions

   Locally this works for me, and #1993 suggests that it indeed should.
4. lclanzi commented on Jan 14, 2020

   

   lclanzi

   on Jan 14, 2020

   Author

   More actions

   Hi, thank you for your reply.
   I did not explain well what the problem is in the first post.
   The code

   import matplotlib.pyplot as plt
   import numpy as np
   plt.pcolormesh(np.random.randn(10,10), shading='gouraud')
   plt.gcf().savefig('test.eps')

   works fine. However, when I save the image with .eps, some pixels change. You can see the difference in my second post. This problem does not exist when I use shading='flat' or when I save the image as .png.
5. anntzer commented on Jan 14, 2020

   

   anntzer

   on Jan 14, 2020

   Contributor

   More actions

   Your example is not reproducible: we don't have the hstar, invparam variables.
6. lclanzi commented on Jan 14, 2020

   

   lclanzi

   on Jan 14, 2020

   Author

   More actions

   I'm sorry, you are right.
   Here it is a reproducible example.

   For images with shading='gouraud':

   import matplotlib
   import matplotlib.pyplot as plt
   import numpy as np
   
   x = np.linspace(0,1,25)
   y = np.linspace(0,1,25)
   X,Y = np.meshgrid(x,y,indexing='ij')
   z = np.sin(X*2*np.pi) + np.cos(Y*2*np.pi)
   cmap = plt.get_cmap('viridis')
   plt.subplots(1,1,figsize=(5.3,4.3))
   plt.pcolormesh(x, y, z, cmap=cmap, shading='gouraud')
   cbar = plt.colorbar()
   plt.title('Gouraud')
   plt.savefig('TestGou.eps',bbox_inches='tight',dpi=300)

   Output:
   
   TestGou-eps-converted-to.pdf

   For images with shading='flat':

   import matplotlib
   import matplotlib.pyplot as plt
   import numpy as np
   
   x = np.linspace(0,1,25)
   y = np.linspace(0,1,25)
   X,Y = np.meshgrid(x,y,indexing='ij')
   z = np.sin(X*2*np.pi) + np.cos(Y*2*np.pi)
   cmap = plt.get_cmap('viridis')
   plt.subplots(1,1,figsize=(5.3,4.3))
   plt.pcolormesh(x, y, z, cmap=cmap, shading='flat')
   cbar = plt.colorbar()
   plt.title('Flat')
   plt.savefig('TestFlat.eps',bbox_inches='tight',dpi=300)

   Output:
   
   TestFlat-eps-converted-to.pdf

   If you compare the image TestGou.png and TestGou-eps-converted-to.pdf, they are different. On the other hand, TestFlat-eps-converted-to.pdf and TestFlat.png are identical.
   I would like to know where I'm doing something wrong. Thank you in advance for your help.
7. anntzer

   removed

   status: needs confirmation

   on Jan 14, 2020
8. anntzer commented on Jan 14, 2020

   

   anntzer

   on Jan 14, 2020

   Contributor

   More actions

   Yes, I can confirm the difference now.
9. QuLogic commented on Apr 21, 2023

   

   QuLogic

   on Apr 21, 2023

   Member

   More actions

   I'm not sure of the exact portion of the change that did it, but this was fixed by #23294.
10. QuLogic

    closed this as completedon Apr 21, 2023
11. QuLogic

    added this to the v3.6.0 milestone on Apr 21, 2023