Coincident topology issues with pythreejs jupyter_backend

Describe the bug, what’s wrong, and what you expected.

Hello,

I have an issue when visualising thin shells. When looking at the normals in the screen recording attached, the values seem incorrect, but the problem disappears when zooming in. Have you encountered this before? Is there a way to fix this?

Steps to reproduce the bug.

Try to visualize a mesh with a thin shell.

System Information

--------------------------------------------------------------------------------
  Date: Mon Oct 17 14:57:25 2022 UTC

                OS : Linux
            CPU(s) : 8
           Machine : x86_64
      Architecture : 64bit
               RAM : 54.9 GiB
       Environment : Jupyter
        GPU Vendor : Mesa/X.org
      GPU Renderer : llvmpipe (LLVM 12.0.0, 256 bits)
       GPU Version : 4.5 (Core Profile) Mesa 21.2.6

  Python 3.8.10 (default, Mar 15 2022, 12:22:08)  [GCC 9.4.0]

           pyvista : 0.33.2
               vtk : 9.0.1
             numpy : 1.19.5
           imageio : 2.9.0
           appdirs : 1.4.4
            scooby : 0.5.7
        matplotlib : 3.3.4
           IPython : 7.16.1
        ipyvtklink : 0.2.2
             scipy : 1.8.1
              tqdm : 4.60.0
            meshio : 4.4.1
--------------------------------------------------------------------------------

Screenshots

https://user-images.githubusercontent.com/37087741/196212363-5df2c0a6-f7ee-4a28-9976-81cee8e9e692.mov

[banesullivan]

Ooof, that does not look good. Can you provide more details?

• Is this rendering in a PyVista render window or with one of the notebook rendering back ends?
• Are there multiple geometries in the scene? I suspect there is and we’re seeing coincident topology not properly being handled.

Try to visualize a mesh with a thin shell.

I’m not too sure what this actually means… Can you share the data in question and the code to make this visualization? Or perhaps some similar data?

[banesullivan]

Do you have any estimated time of release for it?

Not exactly… I’m hoping before End of Year for a conservative timeline but hopefully sooner.

The work is in progress at this point. There is a very, very rough prototype in #3385