2021 Edition.
This is a library for computing the Voronoi diagram of a set of two-dimensional points.
This is a port of d3-delaunay. It is in a very early development phase.
It is part of a collection d3 modules ported into RUST
- d3_geo_rs
- d3_delaunay_rs
- d3_geo_voronoi_rs
Currently the code coverage, as reported by Cargo tarpaulin is 80%.
Three example web pages are provided in the git repository associated with crate
This demo renders the meshes associated with a set of 500 points. The points are created at random.
to run the application
cd examples/500_points
cargo run
This produces a file "mesh.svg"
-
The delaunay mesh is in blue.
-
The voronoi mesh is in green.
Using this library - I have ported this example into RUST
https://observablehq.com/@mbostock/voronoi-stippling
to run the example
cd examples/stippling
npm install
npm run build
npm run serve
Currently the RUST port of this example runs in javascrtipt's main event loop. This needs to be refactored so that the main computation can be run in parallel, ( runs in a web worker ).
This is a confidence building exercise. With only 5 points in a symmetric pattern the meshes can be predicted.
API finalization. There maybe optimization in the area of generics.
We need a profile taget based on the stippling example. To profile and identify bottlenecks.
Functions that use the generator crate are now availble only when the "generator" feature is enabled.
The following functions are under going rapid development.
The following generators functions are missing.
delaunay | voronoi | |
---|---|---|
neighbors() | cellPolygons() |
d3-geo-delaunay has a dependency on this npm package delauantor the function update() allow for rapid retriangulation - in a memory efficient mannor.
This module has a parallel depenecy on delaunator-rs unfortunatly this is missing from the rust port.
There is an open issue to add a update function mourner/delaunator-rs#30