Well, actually, that's a lie. It only works in 2 or 3 dimensions. But! The code
is written such that it would be very easy to get it to work in any number of
dimensions! We just havn't had the need. (To do so, just generalize the
For examples on how to use the code, please look in the
This code isn't especially efficient; in particular, it runs in quadratic time. There are very clever divide-and-conquer algorithms for Delaunay triangulation (even in arbitrary dimensions!) that are O(n log n). That said, it should be fast enough for most purposes.
THIS CODE IS DEPRECATED. It was once used in the Forecast
API, but these days we use the similar (and
much faster and simpler)
npm install delaunay-fast). This code is kept around for historical
To the extend possible by law, The Dark Sky Company, LLC has waived all copyright and related or neighboring rights to this library.