PyGeM¶
Tutorial 5: Deformation of an object stored into file¶
In the present tutorial, we are going to show the fundamental steps to compute in order to deform a given object stored in a given file.
To achieve this, we basically need a parser for extracting from the file all geometrical information (typically the nodes coordinates and the topology), then deforming such nodes with one of the PyGeM deformation techniques.
Here we show a simple FFD applied to a .vtp and to an .stl files. To deal with such files, we employ the Smithers package, a utilities toolbox that allows for an easy manipulation of several file formats.
As usually, at the beginning we import all the modules we need.
%matplotlib inline
import numpy as np
from pygem import FFD
from smithers import io
For visualization purpose, we also implement a small function that shows the object parsed with Smithers.
def plot(data, color=None):
from mpl_toolkits.mplot3d import Axes3D
from mpl_toolkits.mplot3d.art3d import Poly3DCollection
import matplotlib.pyplot as plt
if color is None:
color = (0, 0, 1, 0.1)
fig = plt.figure(figsize=(16,10))
verts = [data['points'][cell] for cell in data['cells']]
ax = fig.add_subplot(111, projection='3d')
faces = Poly3DCollection(verts, linewidths=1, edgecolors='k')
faces.set_facecolor(color)
ax.add_collection3d(faces)
ax.set_xlim3d(-.8, .8)
ax.set_ylim3d(-.8, .8)
ax.set_zlim3d(-.8, .8)
ax.set_aspect('equal','box')
plt.show()
Deformation of the VTP file¶
First of all, we need a .vtp file: we download a simple cube and we open it.
!rm -rf cube.vtp
!wget https://raw.githubusercontent.com/mathLab/Smithers/master/tests/test_datasets/cube.vtp
vtp_filename = "cube.vtp"
vtp_content = io.VTPHandler.read(vtp_filename)
plot(vtp_content)
We can now instantiate a new FFD object, setting position and lenght of the lattice of points as well as their displacements. We present here a very simple deformation, moving just one control points, but of course it is possible to change the FFD settings according to needed. We also remark that of course any other deformation (by PyGeM) can be used!
ffd = FFD()
ffd.origin_box = np.array([-.6, -.6, -.6])
ffd.box_length = np.array([1.2, 1.2, 1.2])
ffd.array_mu_x[1, 1, 1] += 1.5
Now the actual deformation: the points of our object are morphed through the ffd and the new coordinates are replaced to the older one.
vtp_content['points'] = ffd(vtp_content['points'])
Here a visual test to see the final outcome!
plot(vtp_content)
Of course, the deformed object is still contained within the vtp_content dictionary. To save it into a new file, we just need to use again the VTPHandler by Smithers.
io.VTPHandler.write('deform_cube.vtp', vtp_content)
Pretty easy, isn't?
Deformation of an STL file¶
Here we basically replicate (in a more compress way) the previous steps in order to deform an object stored in an STL file. As before, we deal with a simple cube.
!rm -rf cube.stl
!wget https://raw.githubusercontent.com/mathLab/Smithers/master/tests/test_datasets/cube.stl
stl_filename = "cube.stl"
stl_content = io.STLHandler.read(stl_filename)
stl_content['points'] = ffd(stl_content['points'])
io.STLHandler.write('deform_cube.stl', stl_content)
We can now plot the content of the original file and the deformed one, showing the altered geometry.
plot(io.STLHandler.read('cube.stl'))
plot(io.STLHandler.read('deform_cube.stl'), (0, 1, 0, .1))
Other file formats?¶
vtp and stl are common formats in the scientific community, but of course there are many more available. What if you have to deform a different file?
As first step, we suggest to check the Smithers documentation looking for the handler for the file format in hand.
If the latter is not implemented yet, you can write by your own the parser (and hopefully make a Pull-Request to Smithers to make your improvement also available for other users) or just use some third-party libraries to extract the node coordinates!
Happy deformations!