GHPython mesh deconstruction


I’m using ghpython to extract the faces of a mesh, where each face becomes its own mesh. I found a component that did this in c#, but wanted to see if I could write the same thing in python. here is the resulting script, which works, but seems like it could be done much more efficiently. Any pointers much appreciated. I’ve done a lot of coding, but I’m new to coding within rhino/grasshopper.

input M : single Mesh
output FaceMeshes : list of Meshes representing the faces of the input mesh

import rhinoscriptsyntax as rs

def dePoint( p ):
  xyz = []
  return xyz

def meshExplode( m ):
  exploded = []
  for mf in m.Faces:
    v = []
    fv = []
    if mf.IsQuad: 
      fv.append( (0,1,2,3) )
      fv.append( (0,1,2) )
    exploded.append( rs.AddMesh( v, fv ) )
  return exploded

FaceMeshes = meshExplode( M )

in particular it seems odd that I had to write a function to extract the XYZ coords of the original mesh vertices, but if I dont and just try:


When I then call rs.AddMesh, I get the error:
Runtime error (TypeErrorException): ‘Point3f’ object is not iterable

Also, the python script as I’ve written it seems to take about 10 times as long to execute as the C# script it was based on, which didnt seem have issues just passing the vertexes from the existing mesh to the new ones. The original C# script was written by Andrew Heumann. I’ve posted it below for reference:

List<Mesh> meshesOut = new List<Mesh>();
foreach(MeshFace mf in M.Faces){
  Mesh meshToAdd = new Mesh();
    meshToAdd.Faces.AddFace(new MeshFace(0, 1, 2, 3));
  } else {
    meshToAdd.Faces.AddFace(new MeshFace(0, 1, 2));
A = meshesOut;

While you certainly could iterate the faces and construct new meshes from these, I imagine a simpler/more computationally efficient solution here might be to first unweld the mesh and then explode it at the unwelded edges.

Edit: Like so:

Your suggestion is a really good one, and works really well. Almost as well as the native MeshExplode component (which suggests to me that your method is exactly what the native component is doing).

Though I still don’t know why in my code I needed to manually convert the vertices from 3d point objects to coordinate lists. seems unnecessarily clunky.