Online Demo

Thanks to the magic of binder, you can try out petrify online. We have demos for several core petrify features:

• Construction of many basic solids.

• Methods for defining advanced solids.

• Combining multiple solids into a single complex solid.

• Setting visual properties on solids to create the above logo.

Have patience; notebooks on binder can take up to a minute to start. Due to a (possible) bug in pythreejs, you’ll need to click “restart and run all” to see the relevant interactive visualizations.

Design Principles

• Novice focus. This library should make it as easy as possible to build and manipulate solids, shapes, and other geometry.

• Immutable operation. Immutable math is easier for most audiences to reason about. This library explicitly does not target strong real-time applications like games that demand mutability for performance.

• Pluggable engines. CSG is complicated. Different engines have different tradeoffs:

  • The pymesh engine is the default where the pymesh2 library is installed. It uses the CGAL CSG engine, which is mature and quite fast. However, building pymesh with CGAL support can be difficult for novices and in certain environments (like Windows and OSX).

  • cython-csg is relatively fast, but still requires a cython build toolchain.

  • pycsg is a pure-python implementation. It is obviously quite slow, but works everywhere python does. For example, pure python environments like pyiodide can utilize this engine easily.

Quickstart with Docker

We recommend using petrify inside our prepackaged docker image. This image has all the necessary libraries for running a Jupyter notebook server, and the pymesh engine already installed:

docker run -it -p 8888:8888 \
  -v $(pwd)/notebooks:/home/jovyan/work \
  anirishduck/petrify

The above command will expose a notebook on port 8888 with the necessary volume mounts for persisting the work directory. The server will output a ?token=<xxxxx> query parameter you will need for authentication.

Installation

petrify is published online via pip:

pip install petrify

We strongly recommend using its visualization capabilities in combination with Jupyter:

pip install notebook pythreejs

While petrify is functional from this point, you almost certainly want to install a more powerful engine than the default pycsg one. See our csg example for more detail on why; read on to learn how.

sudo apt-get install libcgal-dev

pip install pymesh2

pip install cython

pip install cython-csg

Contributors

This library is a fusion of:

• pycsg

• pyeuclid v3 fork (https://github.com/ezag/pyeuclid/)

• the reverbat STL PR on pycsg (https://github.com/timknip/pycsg/pull/9)