BUG: pyvista backend on headless machine

I’m trying to setup the Pyvista backend for plotting on a headless machine with Jupyter port redirection.

After setting up a new server environment,

and (painfully*) installing pyvista with pyqt5 via

conda install -c conda-forge pyqt vtk pyvista 

(* pip install pyqt5 or pyqt as well as conda install install pyqt and sudo apt-get install python-pyqt5 all lead to pyvista to fail, or miss a python or C dependency one way or another),

Pyvista asked to install pyvistaqt, which I did.

However, I now face an AttributeError:

import mne
mne.viz.set_3d_backend("notebook")

data_dir = mne.datasets.sample.data_path()
subjects_dir = data_dir + '/subjects'
stc_path = data_dir + '/MEG/sample/sample_audvis-meg-eeg'
stc = mne.read_source_estimate(stc_path, 'sample')
stc.plot(hemi='lh', 
         subjects_dir=subjects_dir, 
         initial_time=0.07,
         size=(800, 600), 
         backend='pyvista')

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-1-39692bcd540a> in <module>
      6 stc_path = data_dir + '/MEG/sample/sample_audvis-meg-eeg'
      7 stc = mne.read_source_estimate(stc_path, 'sample')
----> 8 stc.plot(hemi='lh', 
      9          subjects_dir=subjects_dir,
     10          initial_time=0.07,

~/.conda/envs/bm/lib/python3.8/site-packages/mne/source_estimate.py in plot(self, subject, surface, hemi, colormap, time_label, smoothing_steps, transparent, alpha, time_viewer, subjects_dir, figure, views, colorbar, clim, cortex, size, background, foreground, initial_time, time_unit, backend, spacing, title, show_traces, src, volume_options, view_layout, add_data_kwargs, brain_kwargs, verbose)
    649              src=None, volume_options=1., view_layout='vertical',
    650              add_data_kwargs=None, brain_kwargs=None, verbose=None):
--> 651         brain = plot_source_estimates(
    652             self, subject, surface=surface, hemi=hemi, colormap=colormap,
    653             time_label=time_label, smoothing_steps=smoothing_steps,

<decorator-gen-153> in plot_source_estimates(stc, subject, surface, hemi, colormap, time_label, smoothing_steps, transparent, alpha, time_viewer, subjects_dir, figure, views, colorbar, clim, cortex, size, background, foreground, initial_time, time_unit, backend, spacing, title, show_traces, src, volume_options, view_layout, add_data_kwargs, brain_kwargs, verbose)

~/.conda/envs/bm/lib/python3.8/site-packages/mne/viz/_3d.py in plot_source_estimates(stc, subject, surface, hemi, colormap, time_label, smoothing_steps, transparent, alpha, time_viewer, subjects_dir, figure, views, colorbar, clim, cortex, size, background, foreground, initial_time, time_unit, backend, spacing, title, show_traces, src, volume_options, view_layout, add_data_kwargs, brain_kwargs, verbose)
   1778     else:
   1779         with use_3d_backend(backend):
-> 1780             return _plot_stc(
   1781                 stc, overlay_alpha=alpha, brain_alpha=alpha,
   1782                 vector_alpha=alpha, cortex=cortex, foreground=foreground,

~/.conda/envs/bm/lib/python3.8/site-packages/mne/viz/_3d.py in _plot_stc(stc, subject, surface, hemi, colormap, time_label, smoothing_steps, subjects_dir, views, clim, figure, initial_time, time_unit, background, time_viewer, colorbar, transparent, brain_alpha, overlay_alpha, vector_alpha, cortex, foreground, size, scale_factor, show_traces, src, volume_options, view_layout, add_data_kwargs, brain_kwargs)
   1956             TimeViewer(brain)
   1957         else:  # PyVista
-> 1958             brain.setup_time_viewer(time_viewer=time_viewer,
   1959                                     show_traces=show_traces)
   1960     else:

~/.conda/envs/bm/lib/python3.8/site-packages/mne/viz/_brain/_brain.py in setup_time_viewer(***failed resolving arguments***)
    655         self._configure_scalar_bar()
    656         self._configure_shortcuts()
--> 657         self._configure_picking()
    658         self._configure_tool_bar()
    659         self._configure_dock()

~/.conda/envs/bm/lib/python3.8/site-packages/mne/viz/_brain/_brain.py in _configure_picking(self)
   1219                 self.act_data_smooth[hemi] = (act_data, smooth_mat)
   1220 
-> 1221         self._renderer._update_picking_callback(
   1222             self._on_mouse_move,
   1223             self._on_button_press,

~/.conda/envs/bm/lib/python3.8/site-packages/mne/viz/backends/_pyvista.py in _update_picking_callback(self, on_mouse_move, on_button_press, on_button_release, on_pick)
    669                                  on_button_release,
    670                                  on_pick):
--> 671         self.plotter.iren.AddObserver(
    672             vtk.vtkCommand.RenderEvent,
    673             on_mouse_move

AttributeError: 'NoneType' object has no attribute 'AddObserver'

Here is my mne.sys_info()

Platform:      Linux-5.4.0-52-generic-x86_64-with-glibc2.10
Python:        3.8.8 | packaged by conda-forge | (default, Feb 20 2021, 16:22:27)  [GCC 9.3.0]
Executable:    /private/home/jeanremi/.conda/envs/bm/bin/python
CPU:           x86_64: 80 cores
Memory:        503.8 GB

mne:           0.23.dev0
numpy:         1.20.2 {blas=NO_ATLAS_INFO, lapack=lapack}
scipy:         1.6.2
matplotlib:    3.4.1 {backend=module://ipykernel.pylab.backend_inline}

sklearn:       0.24.1
numba:         0.53.1
nibabel:       3.2.1
nilearn:       0.7.1
dipy:          Not found
cupy:          Not found
pandas:        1.2.4
mayavi:        Not found
pyvista:       0.29.0 {pyvistaqt=0.3.0, OpenGL 3.3 (Core Profile) Mesa 18.3.1 via llvmpipe (LLVM 7.0, 256 bits)}
vtk:           9.0.0
PyQt5:         5.12.3

May relate to https://github.com/mne-tools/mne-python/issues/8659 cc @GuillaumeFavelier

[larsoner]

If you do backend='notebook' does it work?

[larsoner]

That seems like a bug, if you hack in 'notebook' does it work?