PyMesh is a code base developed by Qingnan Zhou for his PhD research at New York University. It is a rapid prototyping platform focused on geometry processing. PyMesh is written with both C++ and Python, where computational intensive functionalities are realized in C++, and Python is used for creating minimalistic and easy to use interfaces.

(Model source: Bust of Sappho)

Latest documentation

Perhaps the easiest way of trying out PyMesh is through docker:

docker run -it pymesh/pymesh
Python 3.6.4 (default, Feb 17 2018, 09:32:33)
[GCC 4.9.2] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import pymesh

For example, to run meshstat.py:

docker run -it --rm -v `pwd`:/models pymesh/pymesh meshstat.py -x /models/model.obj

This command mounts your current working directory to the /models directory in a docker container and executes the meshstat.py script on the mesh file model.obj in the current directory.

To retrieve the code:

git clone https://github.com/PyMesh/PyMesh.git
cd PyMesh
git submodule update --init

PyMesh has the following required dependencies:

• Python v2.7 and v3.x.
• NumPy v1.8 or higher
• SciPy v0.13 or higher
• nose v1.3.7 or higher

The following C++ libraries are required. They are included in $PYMESH_PATH/third_party directory.

• Eigen v3.2 or higher
• PyBind11

PyMesh also has a number of optional dependencies:

• Carve: A fast, robust constructive solid geometry library.
• CGAL: The Computational Geometry Algorithms Library.
• Clipper: An open source freeware library for clipping and offsetting lines and polygons.
• Cork: A 3D boolean/CSG library.
• Draco: An open-source library for compressing and decompressing 3D geometric meshes and point clouds
• Geogram: A programming library of geometric algorithms
• libigl: A simple C++ geometry processing library.
• MMG: Robust, open source & multidisciplinary software for remeshing.
• Qhull: Engine for convex hulls, Delaunay triangulations, Voronoi diagrams computations.
• Quartet: A tetrahedral mesh generator that does isosurface stuffing with an acute tetrahedral tile.
• TBB: Thread building blocks from Intel.
• Tetgen: Tetrahedral mesh generation engine.
• Triangle: A two-Dimensional quality mesh generator and Delaunay triangulator.

All of the optional libraries are included in $PYMESH_PATH/third_party directory.

All dependencies are included as git submodules in the third_party directory. It is recommended to build these dependencies directly (see Build section). However, it is possible to use an existing dependency installed on your system. If the dependent library is not installed in standard locations (/usr/local/, /opt/local), one needs to set environment variables that point to the correct directories. PyMesh check the following environment variables:

• Eigen: Set EIGEN_PATH to the directory containing the eigen3 directory.
• CGAL: Set CGAL_PATH to the directory containing UseCGAL.cmake file.
  • Boost: Set BOOST_INC to the directory containing boost.
  • GMP: Set GMP_INC and GMP_LIB to the directories containing GMP header and library.
• libigl: Set LIBIGL_PATH the include directory of libigl sources.
• Cork: Set CORK_PATH to the install directory of Cork.
• Tetgen: Set TETGEN_PATH to the install directory of Tetgen.
• Triangle: Set TRIANGLE_PATH to the install directory of Triangle.
• Qhull: Set QHULL_PATH to the install directory of Qhull.
• Clipper: Set CLIPPER_PATH to the install directory of Clipper.
• Carve: Set CARVE_PATH to the install directory of Carve.
• Geogram: Set GEOGRAM_PATH to the install directory of Geogram.
• Quartet: Set QUARTET_PATH to the install directory of Quartet.
• Draco: Set Draco_PATH to the install directory of Draco.
• MMG: Set MMG_PATH to the install directory of MMG.

Let $PYMESH_PATH be the root directory of the repository. The first step is to compile the optional third party dependencies:

cd $PYMESH_PATH/third_party
build.py all

Third party dependencies will be installed in $PYMESH_PATH/python/pymesh/third_party directory.

Now we can build the main project. It is recommended to build out of source:

cd $PYMESH_PATH
mkdir build
cd build
cmake ..

To build the PyMesh library:

make
make tests

Make sure all unit tests are passed before using the library.

The output of building PyMesh consists a set of C++ libraries and a python module. Installing the C++ library is currently not available. However, installing the python package can be done:

./setup.py build # This an alternative way of calling cmake/make
./setup.py install

To check PyMesh is installed correctly, run the following python unit tests::

python -c "import pymesh; pymesh.test()"

Once again, make sure all unit tests are passed, and report any unit test failures.

Thank you for using PyMesh! Please consider help PyMesh improve by leaving feedback!