Create an irregular polygon shaped cylinder and use it to clip a mesh

[craigmillernz]

I'd like to be able to clip a mesh by an irregular shaped polygon cylinder. This polygon could be defined by the perimeter of a set of datapoints (ie like observations in a geophysical survey), or it could be defined manually using coordinates of the top and bottom of the cylinder.
I would then use this polygon cylinder surface as input into the dataset.clip_surface() filter. The intention is to clip a geophysical model (which is a rectangular cube) to the actual data measurement point distribution, which is a random shape.

[banesullivan]

So just to clarify and follow up on some points from Slack - the clip_surface filter requires a closed ("watertight") mesh to work - we'd need to create a PolyData mesh of the defined surface that encloses all the data in the mesh that you'd like to extract.

Here is an example of how we might create a cyclider and clip a gridded dataset to extract a non-uniform ROI.

FYI - I'm experiencing a significant issue/bug with the clip_surface filter when using custom, irregular surfaces that will need to be mitigated, so I don't have an immediate solution for you when it comes to clipping with custom surfaces - I'll report back

Example A

Used PyVista' cylinder helper method to create a cylinder PolyData object that is closed and clip a UniformGrid with that surface:

import pyvista as pv
from pyvista import examples

# The gridded dataset to clip
mesh = examples.load_uniform()

# The surface that bounds the ROI
surf = pv.Cylinder(center=mesh.center, direction=(1,2,1),
                   height=15, radius=3.3)

p = pv.Plotter(notebook=0) 
p.add_mesh(surf, color='red', opacity=0.5,)
p.add_mesh(mesh, show_edges=True)
p.show()

# Perform the clip
clipped = mesh.clip_surface(surf, compute_distance=True)

p = pv.Plotter(notebook=0) 
p.add_mesh(mesh.outline(), color=True)
p.add_mesh(clipped, show_edges=True)
p.show()

[banesullivan]

Also, the clip_box filter now allows a PolyData box to be passed which should address this.

ref pyvista/pyvista#433 which landed on v0.23.0