1 unstable release
0.1.0 | Apr 26, 2023 |
---|
#130 in #times
20KB
342 lines
This crate implements the alpha and weak-alpha filtration of a set of points in the plane.
Both are filtrations of the Delaunay triangulation.
With the optional lophat
feature, the crate also provides a function for computing the boundary matrix of the filtration.
- The alpha filtration is constructed similarly to the Čech filtration. Grow balls of radius r around each point and intersect each ball with the corresponding Voronoi cell. This nerve of this collection of open sets is the alpha filtration at radius r.
- The weak alpha filtration is a sub-filtration of the Vietoris-Rips filtration. Namely, at each filtration value r, the weak alpha filtration is equal to the Vietoris-Rips filtration intersected with the Delaunay triangulation.
Filtrations of the Delauny triangulation are implemented as a DelaunayTriangulation
from the spade
crate.
The filtration value is stored in the data()
struct associated to each
vertex,
undirected edge and
inner face.
WARNING: To avoid unecessary square roots, the filtration times are squared from their theoretical value.
Dependencies
~5MB
~76K SLoC