user27182

> If I run `pytest --doctest-modules pyvista/core/filters/data_set.py::pyvista.core.filters.data_set.DataSetFilters.cell_data_to_point_data pyvista/core/filters/data_set.py::pyvista.core.filters.data_set.DataSetFilters.pack_labels -v` it _always_ fails. This seems suspicious. `cell_data_to_point_data` is called internally by the `connectivity` filter. And the docstring for `download_frog_tissue` uses both...

> I can reproduce this still on a codespaces with our devcontainer, but I cannot reproduce it on a local machine with seemingly all the same python packages. It is...

> I have an even more subtle report but it can be reproduced without doctest and it does **not** rely on `cell_data_to_point_data`. It gives a different error. Great! I think...

I've tracked this bug down to version 0.42.2 specifically, i.e. 0.42.1 (and prior, I assume) have the 'old' orientation, and version 0.42.2 and later have the 'new' orientation. So, if...

Also, I believe this change was the likely cause for breaking `add_floor`, see PR #5140

I’m going to attempt to make the argument that the Plane’s orientation was incorrect prior to #4853, and that it is now correct. Put another way, I think bug #4850...

After further testing, I can confirm that bug #4850 is also applicable to `Plane`, and this bug was fixed in #4853. If we run a similar example to above, but...

I wrote some tests which suggest that perhaps 22 objects have different orientations in 0.43.1 compared to 0.42.1. See https://github.com/pyvista/pyvista/pull/5435#issuecomment-1890377282 for details.

Here's a counter-example for why the `v0.43` version orientation is incorrect. Have a look at the +Y direction: With `Text3D`, the old `v0.42.1` orientation yields:  With the new...

The TLDR of #5516 is that `pandas` and `polars` have array-like data structures which are implicitly supported anywhere `np.asarray` is used. Not only that, but calling `np.asarray` likely does not...