PyRT (pronounced pirate) is a raytracer/image generator for Python 3.5 and higher. This project is mainly done with the following in mind:
- Ray Tracing in the Jupyter Notebook
- Teaching ray tracing
- Exploring ray tracing concepts for geo data using Python.
- Rendering geo data, including large point clouds.
- Implementing new algorithms for rendering large 3D city models.
- Creating 3D-Maps from OpenStreetMap data
- Server-side rendering / cloud based rendering
- ...
PyRT is work in progress.
Installation can be done with pip. Please note that pyrt is still under heavy development and not yet meant for production. (API breaks are possible!)
pip install pyrtCurrently it is recommended to pip directly from the repository:
pip install https://github.com/martinchristen/pyRT/archive/master.zip --upgradePyRT doesn't have any dependencies, however, pillow and numpy are highly recommended.
There are a lot of examples available how to use this module. This is the recommended way to get started with pyRT.
PyRT is not a 3D-modelling package. It is all about rendering from code.
In PyRT you create a scene first. Scenes consist of atleast one camera and geometry. Creation of scenes is done in an object oriented way:
from pyrt.math import *
from pyrt.geometry import Triangle, Vertex
from pyrt.material import PhongMaterial
from pyrt.camera import PerspectiveCamera
from pyrt.renderer import SimpleRT
from pyrt.scene import Scene
camera = PerspectiveCamera(640,480)
scene = Scene()
scene.add(Triangle(Vertex(position=(0, 0, 0)),
Vertex(position=(0, 5, 0)),
Vertex(position=(1, 5, 0)), material=PhongMaterial()))
scene.setCamera(camera)
engine = SimpleRT()
imgdata = engine.render(scene)PyRT has an open rendering concept, you can create your own renderer. In the example above "SimpleRT" was used, which is a minimalistic reference implementation.
No. Custom renderers can be written in C with Python bindings. This is planned in future, later versions will even support the GPU using OpenCL and/or other libraries, but at the moment the primary focus is to create a "pythonic" ray tracer.
PyRT is released under MIT. More information about this license can be found under: https://opensource.org/licenses/MIT
PyRT is created and maintained by Martin Christen. You can contact me by e-mail: martin.christen@fhnw.ch
Please note: This project is mostly developed in my spare time, there is no official support or anything.