I am seriously interested in a solution.
If there is no workaround to get some kind of categorical labeling, please point in the right direction, so I can create a feature request in the correct repository (vtk.js / trame-vtk).
I am pretty sure I am not the only one with a use case, and this will come up more often in the future.

The first step would be to add support in vtk.js. Probably in the ScalarBarActor and its implementations WebGL/WebGPU.
Once working on the vtk.js side, we may need to adjust the scene serializer, but it should be pretty straightforward if we have "parity" between the C server and the JS client.

OK, so I will go ahead and create an FR in the vtk.js repository. However, I am afraid I won't be of mutch help. JS and c are not on my strong side.

As mentioned in the first post, legends are also not working yet. But as these are implemented as widgets and widgets are also generally not supported in vtk.js, I do not expect this to be happening any time soon.

However, I want to thank you for the awesome framework you guys have created. It is a blast to work with!

With the current releases, the example from the initial post, exported to html, results in a white screen without anything being rendered.

This is not caused by the categorial labelling. The Problem occurs as soon as the scalars="xyz" kwarg is added to p.add_mesh.
The exported png is fine. It does not depend on the data type stored in the plotted array.

--------------------------------------------------------------------------------
  Date: Wed Sep 11 18:19:14 2024 CEST

                  OS : Linux (Rocky Linux 8.9)
              CPU(s) : 128
             Machine : x86_64
        Architecture : 64bit
                 RAM : 251.7 GiB
         Environment : IPython
         File system : xfs
          GPU Vendor : Mesa
        GPU Renderer : llvmpipe (LLVM 18.1.8, 256 bits)
         GPU Version : 4.5 (Core Profile) Mesa 24.1.5
    MathText Support : True

  Python 3.12.5 | packaged by conda-forge | (main, Aug  8 2024, 18:36:51) [GCC
  12.4.0]

             pyvista : 0.44.1
                 vtk : 9.3.1
               numpy : 2.1.0
          matplotlib : 3.9.2
              scooby : 0.10.0
               pooch : 1.8.2
              pillow : 10.4.0
             imageio : 2.35.1
           pyvistaqt : 0.11.1
               PyQt5 : 5.15.11
             IPython : 8.26.0
          ipywidgets : 8.1.5
               scipy : 1.14.1
                tqdm : 4.66.5
              meshio : 5.3.5
               trame : 3.6.3
        trame_client : 3.2.5
        trame_server : 3.1.0
           trame_vtk : 2.8.10
       trame_vuetify : 2.6.2
jupyter_server_proxy : 4.3.0
        nest_asyncio : 1.6.0
--------------------------------------------------------------------------------

@jourdain

trame-vtklocal and vtk 9.4 provides a solution but that also mean that PyVista and VTK 9.4 needs to play well together. And we are still working on that.

trame-vtklocal and vtk 9.4 provides a solution but that also mean that PyVista and VTK 9.4 needs to play well together. And we are still working on that.

Really looking forward this 👍 Is there any estimated time for this ❤️

I think PyVista needs 9.4.1 which is supposed to happen in begining on January. After that, I'm not sure about the PyVista side to be fully ok with 9.4.1. Also, someone will have to take on the trame-vtklocal into PyVista unless you do it on your side/app.

vtk.js does not support text labels hence the issue.