Define Delaunay triangulation for n = 0, n = 1 and n = 2. #19
Labels
enhancement
New feature or request
Comments
|
Hello Mike, The n=0 and n=1 cases are easy to work around (it's just the entire bounding area). But the n=2 requires cooperation from the underlying voronoi region algorithm. I'm not super familiar with the mathematics behind delaunay triangulation or voronoi regions or else I'd write up a pull request. If you could find some time to fix this issue, it would be much appreciated. |
|
Here's my tentative It's gross (try/catch, and add external points at near-infinity if the triangulation failed). But hey! |
Fil
added a commit
that referenced
this issue
Jun 19, 2019
Fix Delaunay and Voronoi for degenerate cases For 0, 1, 2 points, and for collinear points: - returns the correct topology (neighbors, find) - returns the correct Voronoi cells (bounding box, half plane…) - resorts to deterministic jittering only in the case of multiple collinear points Fixes: - #19 - #20 - #50 - #65
|
Fixed by 2ee9847 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Currently we throw an error from Delaunator:
There is a valid Voronoi diagram for n = 1 and n = 2, so we should allow an empty Delaunay triangulation with no triangles in these cases. And similarly for n = 0, we can have an empty Voronoi diagram with no cells.
The text was updated successfully, but these errors were encountered: