Inconsistent plotting results with PyVistaQT versus Jupyter (Trame) backends

[robertsoakes]

Inconsistent plotting results with PyVistaQT versus Jupyter (Trame) backends

[larsoner]

Can you check to see if pv.Plotter itself (rather than QtInteractor / BackgroundPlotter) also has this problem?

[robertsoakes]

@larsoner I'll try and stage a test for this tomorrow.

I'll be honest, I'm not completely sure how, though. Is there a way to pass a pv.Plotter instance to QT -- I guess I could probably just do a screen capture. The Jupyter server backend seems to work, but volume rendering is not supported in the current client renderer, and I wasn't able to get the deprecated renderers to display anything.

I spent some additional time poking at things, and the issue seems to be partly due to PyVista ignoring data in the vtkImageData direction matrix. I wasn't able to figure out how to apply the rotation matrix to the image data, but the results for a unit matrix and the directional matrix look exactly the same.

This is what the Jupyter Trame backend plots without the direction matrix:

After having poked more at the Trame backend, it looks like PyVista is just passing the underlying VTK classes through? (Trame is super cool, BTW. Really looking forward to seeing where it goes.)

[larsoner]

I'll be honest, I'm not completely sure how, though. Is there a way to pass a pv.Plotter instance to QT

No, I'm suggesting you avoid Qt entirely for the pv.Plotter test. Just use that interface with whatever window it pops up and see if it has the same problem or not. It will help identify if it's really a Qt-specific problem or not.