faster_poisson

This is a library that implements various methods of Poisson disk sampling.

Each method returns a set of points where no two are less than a fixed distance from each other.

Usage

The easiest way to use this library is with the [Poisson2D], [Poisson3D], and [PoissonND] types. You can generate points all at once with the run method, or lazily with the iter method. To change parameters like the grid dimensions or minimum distance, [Poisson] uses a fluent interface pattern.

use faster_poisson::{Poisson2D, Poisson3D, PoissonND};

// Sample points from a 6 x 4 grid with minimum distance 0.5.
let poisson_2d = Poisson2D::new().dims([6.0, 4.0]).radius(0.5);
let samples_2d = poisson_2d.run();

// The default side length of the grid is 1.0 and the default radius is 0.1.
let poisson_3d = Poisson3D::new();
let samples_3d = poisson_3d.run();

// For dimensions higher than 3, use PoissonND.
let poisson_4d = PoissonND::<4>::new();
// Points are generated lazily, so this is fast.
let samples_4d_100: Vec<[f64; 4]> = poisson_4d.iter().take(100).collect();

For smaller 2D grids (< 1,000,000 points) it is probably faster to use [PoissonDart2D].

Features

• gpu

  Use the wgpu crate to perform Poisson disk sampling on the GPU with [pixelpie::PoissonPixelPie].

• plotly

  Use the plotly crate to plot distributions with [plot::plot_2d] and [plot::plot_3d].

• fourier

  Use the rustfft crate to generate 2D periodograms of distributions with [fourier::fourier].

  

• stipple

  Stipple images with [stipple::stipple_image_luma] and [stipple::stipple_image_rgb].