Open-source python package for non-manifold subdivison surfaces algorithm
The PySubdiv library is a python-based open-source package that can apply non-manifold subdivision surfaces algorithm (Loop) on triangluar meshes with considering unique semi-sharp creases value for each edge. The implementation of semi-sharp creases in the algorithm allows to model of different varieties of structures with sharp edges. Fitting of smooth subdivision surfaces to the input dense meshes for reconstruction is implemented in this python package.
Currently the software can work well on Linux (without conda enviroment). Based on the user reports you may face some errors if you are using Mac OS or Windows, mostly becasue of the problems of PyQt with these framworks.
Tested on Ubuntu 22.04.1
- Clone the repository from source (We are currently working with a pip installation)
git clone https://github.com/SimBe-hub/PySubdiv.git- Install the packages and dependencies
cd PySubdiv
python3 setup.py installfor developers
python3 setup.py develop- Important：
install
easyguifrom apt, NOT pip
apt-get install python3-easyguiTest if the installation is successful
python3 examples/Introduction/CreateMesh.pyusign pip install PySubdiv
This method may not work, depend on your system.
import PySubdiv as psd
from PySubdiv.data import files
import os
absolute_path = os.path.dirname(__file__)
# Create a simple Mesh by passing vertices and faces
# Create empty object or pass vertices and faces directly
mesh = psd.Mesh()
# vertices do have the shape (n,3), we can pass a list or numpy array
# where n is the number of vertices
# each vertex is defined by it's x,y and z coordinate in space
vertices = [[-3., 3., -7.], [1., 2., 1.], [1., 3., -7.], [-3., -1., 1.], [1., -1., 1.], [-8., -1., -2.],
[-12., -1., -8.], [-8., -1., -10.], [1., -1., -7.], [-3., -1., -7.], [-8., 2., -2.], [-12., 3., -8.],
[-12., 2., 0.], [-3., 2., 1.], [-12., -1., 0.], [-8., 3., -10.]]
# set the vertices property
mesh.vertices = vertices
# Faces are defined by the indices of the vertices and have the shape (n,3) for triangular meshes.
# n is the number if faces. Each face is defined by three vertices
faces = [[0, 1, 2], [1, 3, 4], [5, 6, 7], [8, 3, 9], [8, 1, 4], [0, 8, 9],
[10, 11, 12], [0, 10, 13], [10, 14, 5], [7, 11, 15], [6, 12, 11],
[9, 5, 7], [13, 5, 3], [9, 15, 0], [0, 13, 1], [1, 13, 3], [5, 14, 6],
[8, 4, 3], [8, 2, 1], [0, 2, 8], [10, 15, 11], [0, 15, 10], [10, 12, 14],
[7, 6, 11], [6, 14, 12], [9, 3, 5], [13, 10, 5], [9, 7, 15]]
# set the faces property
mesh.faces = faces
# visualize the mesh
mesh.visualize_mesh()
# As it is quite tedious to set vertices and faces manually we simply can load and also save meshes from and to obj file
# load mesh by passing the file path as string
relative_path = "Meshes/CoarseChannel.obj"
full_path = os.path.join(absolute_path, relative_path)
mesh_loaded_from_obj = files.read(full_path)
# lets have a look:
mesh_loaded_from_obj.visualize_mesh()
# Yes indeed it is the same mesh
# saving the mesh to obj file by passing a permitted file path, here I'm just overwriting.
# The file ending will be added automatically if forgotten.
mesh.save_mesh(full_path)If you find our work useful in your research, please cite our work:
-
Moulaeifard M, Wellmann F, Bernard S, de la Varga M, Bommes D. Subdivide and Conquer: Adapting Non-Manifold Subdivision Surfaces to Surface-Based Representation and Reconstruction of Complex Geological Structures. Mathematical Geosciences. 2023 Jan;55(1):81-111.
-
Moulaeifard, M., Bernard, S., and Wellmann, F.: PySubdiv: open-source geological modelling and reconstruction by non-manifold subdivision surfaces, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2022-685, 2022.
The PySubdiv is under GNU license. https://zenodo.org/record/6878051#.YuGVynZByUk
Full stack developer : Simon Bernard (SimBe-hub)
Project manager and backend developer : Mohammad Moulaeifard (MohammadCGRE)
Project supervisor from RWTH Aachen university: Prof. Florian Wellmann