MiniQhull

MiniQhull (Qhull-based Delaunay triangulation).

Documentation

The MiniQhull julia package provides a single function delaunay overloaded with 3 methods:

delaunay(dim::Integer, numpoints::Integer, points::AbstractVector) -> Matrix{Int32}

Compute the Delaunay triangulation of a cloud of points in an arbitrary dimension dim. The length of vector points must be dim*numpoints. Vector points holds data in "component major order", i.e., components are consequitive within the vector. The returned matrix has shape (dim+1, nsimplices), where nsimplices is the number of simplices in the computed Delaunay triangulation.

delaunay(points::AbstractMatrix) -> Matrix{Int32}

In this variant, the cloud of points is specified by a matrix with size(matrix) == (dim, numpoints).

delaunay(points::AbstractVector) -> Matrix{Int32}

In this variant, the cloud of points is specified with a vector of dim-element vectors or tuples. It is also possible to use a vector of other tuple-like types, like SVector from StaticArrays.jl.

Adjusting Qhull flags

You can override the default set of flags that Qhull uses by passing an additional flags argument:

delaunay(dim::Integer, numpoints::Integer, points::AbstractVector, flags::AbstractString) -> Matrix{Int32}
delaunay(points::AbstractMatrix, flags::AbstractString) -> Matrix{Int32}
delaunay(points::AbstractVector, flags::AbstractString) -> Matrix{Int32}

The default set of flags is qhull d Qt Qbb Qc Qz for up to 3 dimensions, and qhull d Qt Qbb Qc Qx for higher dimensions. The flags you pass override the default flags, i.e. you have to pass all the flags that Qhull should use.

Examples

using MiniQhull
dim          = 2
numpoints    = 4
coordinates  = [0,0,0,1,1,0,1,1]
connectivity = delaunay(dim, numpoints, coordinates)
# output
3×2 Array{Int32,2}:
 4  4
 2  3
 1  1

using MiniQhull
coordinates  = [0 0 1 1; 0 1 0 1]
connectivity = delaunay(coordinates)
# output
3×2 Array{Int32,2}:
 4  4
 2  3
 1  1

using MiniQhull, StaticArrays
dim = 5
npts = 500
pts = [SVector{dim, Float64}(rand(dim)) for i = 1:npts];
flags = "qhull d Qbb Qc QJ Pp" # custom flags
connectivity = delaunay(pts, flags)

Installation

MiniQhull is a registered Julia package. If your system fulfills all the requirements (see below), MiniQhull can be installed using the command:

pkg> add MiniQhull

If, for any reason, you need to manually build the project, write down the following commands in the Julia REPL:

pkg> add MiniQhull
pkg> build MiniQhull

Requirements

The MiniQhull package requires the Qhull reentrant library installed in your system.

By default, if julia >= 1.3, it will use Qhull_jll library and you don't need to manually install anything.
If you need to perform a custom Qhull installation. It can be installed in any path on your local machine as long as you export the environment variable QHULL_ROOT_DIR pointing to the installation directory.
If none of the previous choices are used, MiniQhull will try to find the library in the usual linux user library directory (/usr/lib).

MiniQhull also requires any C compiler installed on the system.

Qhull installation

From Sources

Custom installation of Qhull can be performed as described in the official Qhull installation instructions. You can find the latest source code in the oficial Qhull download section.

Debian-based installation from package manager

The reentrant Qhull library can be installed with apt in recent Debian-based linux distributions as follows:

$ sudo apt-get install update
$ sudo apt-get install libqhull-r7 libqhull-dev

If you need to install a C compiler, it can be also obtained by means of apt tool:

$ sudo apt-get update
$ sudo apt-get gcc

Continuous integration

In order to use MiniQhull in continuous integration jobs, you must ensure that its installation requirements are fullfilled in the CI environment.

For julia < 1.3 jobs, if your CI process is based on Travis-CI you can add the following block at the beginning of your .travis.yml file:

os:
  - linux
dist:
  - bionic
addons:
  apt:
    update: true
    packages:
    - gcc
    - libqhull-r7
    - libqhull-dev

Gnimuc/MiniQhull.jl