nan_color ignored from VTK panel==1.4.1 after exporting to HTML

[edsaac]

Description of expected behavior and the observed behavior

Cells with NaN values should be colored according to the provided nan_color kwarg. However, they always default to a redish color. I believe this is a follow up from #6627, where the color mapping was fixed, except for handling NaN colors.

Screenshots or screencasts of the bug in action

No bug	Bug

Software version info

No bug	Bug
panel==1.3.8	panel==1.4.1
bokeh==3.3.4	bokeh==3.4.1
pyvista==0.43.5	pyvista==0.43.5

Complete, minimal, self-contained example code that reproduces the issue

import pyvista as pv
import panel as pn
import numpy as np
pn.extension("vtk")

## Initialize a plotter object
plotter = pv.Plotter(window_size=[400, 400])

## Create a mesh
mesh = pv.Sphere(radius=1.0)

## Associate a scalar field to the mesh
z = mesh.cell_centers().points.T[2]
mesh["myscalar"] = np.where(z < 0.7, z, np.nan)

## Add mesh to the plotter
plotter.add_mesh(
    mesh, scalars="myscalar", cmap="viridis", show_edges=True, nan_color="magenta"
)

## Some final touches
plotter.background_color = "white"
plotter.view_isometric()
plotter.add_title(f"Using panel=={pn.__version__}")

## Pass plotter to panel
sphere_pan = pn.panel(plotter.ren_win, width=400, orientation_widget=True)

## Export as HTML
with open(f"sphere_pan_{pn.__version__}.html", 'w') as f:
    sphere_pan.save(f)

Stack traceback and/or browser JavaScript console output

With panel==1.4.1:

[bokeh 3.4.1] [bokeh-3.4.1.min.js:166:1469](https://cdn.bokeh.org/bokeh/release/bokeh-3.4.1.min.js) WebGL warning: getParameter: Framebuffer not complete. (status: 0x8cd7)

[edsaac]

This is perhaps related: the above_color and below_color values are also ignored.

No bug	Bug

Code:

import pyvista as pv
import panel as pn

pn.extension("vtk")

## Initialize a plotter object
plotter = pv.Plotter(window_size=[400, 400])

## Create a mesh
mesh = pv.Sphere(radius=1.0)

## Associate a scalar field to the mesh
z = mesh.cell_centers().points.T[2]
mesh["myscalar"] = z 

## Add mesh to the plotter
plotter.add_mesh(
    mesh,
    scalars="myscalar",
    cmap="viridis",
    show_edges=True,
    clim=[-0.5, 0.5],
    above_color="magenta",
    below_color="orange",
)

## Some final touches
plotter.background_color = "white"
plotter.view_isometric()
plotter.add_title(f"Using panel=={pn.__version__}")

## Pass a plotter to panel
sphere_pan = pn.panel(plotter.ren_win, width=400, orientation_widget=True)

## Export as HTML
with open(f"sphere_pan_{pn.__version__}.html", 'w') as f:
    sphere_pan.save(f)