Repository navigation
antialiased image check seems wrong when used on zoomed image #15517
Description
Bug report
Bug summary
The new antialiased image code appears to not handle zoomed images correctly.
attn @jklymak
Code for reproduction
from pylab import *
plt.rcdefaults()
fig, axs = plt.subplots(1, 2)
axs[0].imshow(np.random.rand(10, 10)) # now defaults to interpolation="antialiased"
axs[1].imshow(np.random.rand(1000, 1000))
axs[1].set(xlim=(-.5, 9.5), ylim=(-.5, 9.5))
plt.show()Actual outcome
Run with mpl master (or anything after #13724 was merged).
Expected outcome
The right image should also be rendered using "nearest" interpolation.
Matplotlib version
- Operating system: linux
- Matplotlib version: master
- Matplotlib backend (
print(matplotlib.get_backend())): qt5agg - Python version: 37
- Jupyter version (if applicable): no
- Other libraries:
Activity
jklymak commentedon Oct 25, 2019
Fair enough, its because we use the actual data shape, not the visible data shape:
I'm not quite sure how we can toggle between the two... I thought we had zooming working, but maybe I'm wrong...
anntzer commentedon Oct 25, 2019
Indeed, I somehow remember that we did check zooming, but heh...
jklymak commentedon Oct 25, 2019
The y transform isn't being applied properly. I'm just doing
but I'm pretty sure shape[0] is not being transformed correctly to
dispyjklymak commentedon Oct 25, 2019
Ooops, I think I fixed it, but need to add this as a test.