Convert netgen mesh to numpy arrays?

rick42
Topic Author
Offline
New Member

3 years 7 months ago #3618 by rick42

Convert netgen mesh to numpy arrays? was created by rick42

Is there an efficient way to extract or convert a netgen mesh to numpy arrays in Python. I'm currently using the following

Code:

p = mesh.Points()
vertices = np.empty(shape=(len(p), 3), dtype='float32')
for i, p_i in enumerate(p):
    vertices[i] = list(p_i)
    pass

faceels = mesh.Elements2D()
faces = np.empty(shape=(len(faceels), 3), dtype='uint32')
for i, el in enumerate(faceels):
    faces[i] = [p.nr-1 for p in el.points]
    pass

which isn't very efficient. Is there a better way?

rick42
Topic Author
Offline
New Member

3 years 7 months ago #3633 by rick42

Replied by rick42 on topic Convert netgen mesh to numpy arrays?

Is there anyone who can comment on this (is it possible at all?), or point out something in the documentation (as I couldn't find anything about this)?

christopher
Offline
Administrator

3 years 7 months ago - 3 years 7 months ago #3637 by christopher

Replied by christopher on topic Convert netgen mesh to numpy arrays?

Hm depends what you want to get from python with regards to efficiency, but you could write it more compact and more efficient like this:

Code:

from netgen.geom2d import unit_square
mesh = unit_square.GenerateMesh(maxh=0.2)

import numpy as np
# verts = np.array([v.point for v in mesh.vertices])
verts = np.array([list(p) for p in mesh.Points()])
print(verts)
els = np.array([el.vertices for el in mesh.Elements2D()])

print(els)

Last edit: 3 years 7 months ago by christopher.

Time to create page: 0.109 seconds