Introduction and Scope

This repository contains CPython bindings for libnest2d, a library to pack 2D polygons into a small space. Libnest2d implements the 2D bin packing problem.

The objective of this repository is to allow libnest2d to be called from Python using Numpy. There is a competing solution to provide Python bindings to this end. However this solution is licensed under AGPL. Since Cura uses an LGPL license, it could not use that solution. This implementation is also too limited for Cura's purposes, so a new bindings project was set up that allows more configuration of the nesting algorithm.

Usage

This is an example of how you can use these Python bindings to arrange multiple shapes in a volume.

>>> from pynest2d import *
>>> bin = Box(1000, 1000)  # A bounding volume in which the items must be arranged, a 1000x1000 square centered around 0.
>>> i1 = Item([Point(0, 0), Point(100, 100), Point(50, 100)])                # Long thin triangle.
>>> i2 = Item([Point(0, 0), Point(100, 0), Point(100, 100), Point(0, 100)])  # Square.
>>> i3 = Item([Point(0, 0), Point(100, 0), Point(50, 100)])                  # Equilateral triangle.
>>> num_bins = nest([i1, i2, i3], bin)  # The actual arranging!
>>> num_bins  # How many bins are required to add all objects.
1
>>> transformed_i1 = i1.transformedShape()  # The original item is unchanged, but the transformed shape is.
>>> print(transformed_i1.toString())
Contour {
    18 96
    117 46
    117 -4
    18 96
}
>>> transformed_i.vertex(0).x()
18
>>> transformed_i.vertex(0).y()
96
>>> i1.rotation()
4.71238898038469

For full documentation, see libnest2d. These bindings stay close to the original function signatures.

Building

This library has a couple of dependencies that need to be installed prior to building:

• libnest2d, the library for which this library offers CPython bindings, and its dependencies:
  • Clipper, a polygon clipping library.
  • NLopt, a library to solve non-linear optimization problems.
  • Boost, the headers only.
• Sip, an application to generate Python bindings more easily.