List of interpolation algorithms and boolean/topological operations

[banesullivan]

Description

I was asked:

is there a list of interpolation algorithms and boolean/topological operations (intersect, cut, merge, etc.) implemented in PyVista? The tutorials are quite focused on visualization but I understand that there is much more.

We should create an example in the docs gallery to showcase all of the boolean operations with PolyData

[banesullivan]

boolean/topological operations (intersect, cut, etc.) methods are implemented for PolyData mesh types only and are accessible directly from any PolyData mesh. Check out https://docs.pyvista.org/core/filters.html#polydata-filters and take a look at the following filters:

• boolean_add
• boolean_cut
• boolean_difference
• boolean_union

For merging, the + operator can be used between any two meshes in PyVista (see pyvista/pyvista#510) which simply calls the .merge() filter to combine any two meshes.

Interpolation algorithms - we are lacking here, but the .interpolate() and .sample() filters will be what you want to use here. All we have are simple Gaussian kernels for the interpolations. This example shows how to use these filters.

[Sunil7545]

Hi Bane,
These methods are also applicable surfaces with boundaries or they are working only for the watertight surfaces ?

Thanks
Sunil

[akaszynski]

@Sunil7545,

It looks like operations with non-manifold surfaces aren't always successful. For example, the following leads to a segfault:

import pyvista

sphere = pyvista.Sphere()
plane = pyvista.Plane(i_size=1.5, j_size=1.5).triangulate()
(sphere - plane).plot(color='w', show_edges=True)

However, the following works:

import pyvista

sphere = pyvista.Sphere()
sphere2 = pyvista.Sphere(center=(-0.25, -0.25, 0.25))
cut_sphere = sphere.boolean_difference(sphere2)

pl = pyvista.Plotter()
pl.add_mesh(sphere, style='wireframe')
pl.add_mesh(sphere2, style='wireframe')
pl.add_mesh(cut_sphere, color='r')
pl.show()

So, to answer your question, non-manifold mesh boolean operations may fail, though I've been able to successfully perform them in some circumstances, for example:

[Sunil7545]

Thank you very much.

[punyidea]

Wanted to add to this thread, that the .probe() filter is also helpful if one has a point set, and one is interested in evaluating the mesh at specific points.