WRL and VRML to Matplotlib/Numpy

Question

I would like to import (or parse ??) a WRL/VRML file format in python in order to work the data with Matplotlib. Is there any way of converting this file format into plottable numpy arrays ?

I attached a photo a WRL file visualized with the FreeWRL viewer

Answer 1

{Assuming you still need the code} Let A1.wrl is your wrl file.

import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D
import numpy as np
import re
holder = []
with open("A1.wrl", "rb") as vrml:
    for lines in vrml:
        a = re.findall("[-0-9]{1,3}.[0-9]{6}", lines)
        if len(a) == 3:
            holder.append(map(float, a))

holder_array = np.array(holder) #if you want numpy array

#3D Plotting
x,y,z = zip(*holder)
fig = plt.figure()
ax = Axes3D(fig)
ax.plot(x,y,z)
plt.show()

My A.wrl: http://pastebin.com/PeVXbR5C

File generate:

Answer 2

For the people who end up here. I typically use trimesh to handle and visualize meshes and since trimesh does not support WRL/VRML...

import trimesh

def read_wrl(file_path):

    with open(file_path, 'r') as f:
        s = f.read()

    s2 = re.sub(r'\n|\r|,', ' ', s)
    s3 = re.sub('\s+', ' ', s2)
    
    l1 = re.search(r'point \[(.*?)\]', s3).group(1).strip().split()
    points = np.float32(l1).reshape((-1,3))

    l2 = re.search(r'coordIndex \[(.*?)\]', s3).group(1).strip().split()
    coordIndexs = np.int32(l2).reshape((-1,4))

    mesh = trimesh.Trimesh(vertices=points, faces=coordIndexs[:,:3])

    return mesh