pyvista/pyvista: VTK 9 Support and PyQt5 Refactor

Release 0.25.1

It's been three months since the last release and there are quite a few bug fixes and features that are ready to be tested. Please see the release candidate wheels here: https://github.com/pyvista/pyvista/releases/tag/0.25.0RC

@pyvista/developers, the next release is ready to be tested out. Please install the wheel and post any issues you encounter here.

Ideally, we can test the RC and if there are no problems we can have the release out by the end of this week. There are a few projects that need Python 3.8 support, and which means supporting VTKv9.

Here are the release notes. I've drafted it here so we can just copy and paste it into the release.

• Moved BackgroundPlotter and QtInteractor to pyvistaqt. See #719 for the reasoning, but as a summary this boils down to licencing, unit testing, and improved development. See pyvistaqt to test out the new module. Wheels will be released for that as soon as we've released wheels for pyvista==0.25.0 as it's a dependancy.
• Added better support for VTKv9. There were several changes under the hood to VTK's UnstructuredGrids that need to be implemented here, as well as many changes to our unit testing and CI to support Python 3.8 and VTKv9. See #732, #772, and #693.

• Removing cells or faces (#653)
• Progress bar using tqdm (#608)
• Plotting floors (#424)
• Additional geometry features (#449)
• Subplot groups (#696)
• Custom orientation widget (#702)

• Checking vectors when setting vectors #763
• Check overwriting dataset to avoid segfault #738
• Fix faulthandler (#694 and #645)
• Implicit point setters (#769)
• Ellipsoid wrapping (#705)

• Unit testing is between 25-40% faster. MacOS testing no supports multiple testing processes using pytest-xdist. See #739 and #744
• Improved code coverage #745 and #709
• Removed all example downloading from unit testing. Greatly improves test stability (#731)
• Various speedups (#661)

• Removed untested pyansys quality method.
• Using underscore numpy types (#762)
• Optimized imports (#743)
• Line break consistent (#737)
• Use new style super() calls (#741)
• Generic load/save (#736)
• Refactored active scalars/vectors (#746)
• Better error types (#759)
• And many many more...

Thanks to all those who contributed to this release! In no particular order:

• @JevinJ
• @mli0603
• @banesullivan
• @GuillaumeFavelier
• @dcbr
• @tomsal
• @tkoyama010
• @ralovich
• @florian-wagner
• @imsodin
• @larsoner
• @adeak
• @akaszynski