Switch to using sphinx keep-going flag

[akaszynski]

This PR changes our documentation build to use --keep-going rather than our custom error catching script. See sphinx-build options.

Within the PyAnsys project, we've discovered that it's a brittle script and it's better to use a tried and true method rather than potentially letting warnings slip by. Plus, it's one less script to maintain.

[codecov]

Codecov Report

Merging #2719 (33662fe) into main (feb3664) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##             main    #2719   +/-   ##
=======================================
  Coverage   93.81%   93.81%           
=======================================
  Files          76       76           
  Lines       16195    16195           
=======================================
  Hits        15194    15194           
  Misses       1001     1001           

[tkoyama010]

Good catch. LGTM.
Reference: https://www.sphinx-doc.org/en/master/man/sphinx-build.html#cmdoption-sphinx-build-keep-going

[tkoyama010]

Recommending merge on 1 June 2022 unless further comments are given or issues are raised.

[akaszynski]

it's a brittle script

 File "<string>", line 5, in <module>
  File "/home/runner/work/pyvista/pyvista/pyvista/plotting/helpers.py", line 241, in plot
    plotter.add_mesh(var_item, **kwargs)
  File "/home/runner/work/pyvista/pyvista/pyvista/plotting/plotting.py", line 2224, in add_mesh
    mesh, scalars = prepare_smooth_shading(
  File "/home/runner/work/pyvista/pyvista/pyvista/plotting/_plotting.py", line 88, in prepare_smooth_shading
    mesh.point_data[indices_array] = np.arange(arr_sz, dtype=np.int32)
  File "/home/runner/work/pyvista/pyvista/pyvista/core/datasetattributes.py", line 227, in __setitem__
    self.set_array(value, name=key)
  File "/home/runner/work/pyvista/pyvista/pyvista/core/datasetattributes.py", line 613, in set_array
    vtk_arr = self._prepare_array(data, name, deep_copy)
  File "/home/runner/work/pyvista/pyvista/pyvista/core/datasetattributes.py", line 769, in _prepare_array
    raise ValueError(f'data length of ({data.shape[0]}) != required length ({array_len})')
ValueError: data length of (12) != required length (20)

Turns out we're already not capturing these warnings.

non-captured warnings

[akaszynski]

Main issue here is somehow the split_sharp_edges setting is persisting despite restoring the cache. I've fixed that by simply adding to the split_sharp_edges example.

Since the underlying issue will be fixed anyway in #2695, I think this is sufficient for now to clean up these errors.

[tkoyama010]

Approving after checking the CI pass.

[tkoyama010]

Suggest merging on June 5th.