pulsatrix 1.1.0
Loading...
Searching...
No Matches
rips_persistence.hpp File Reference

Persistent homology of the Vietoris-Rips filtration in every dimension up to a maximum: components (H0), loops (H1), voids (H2) and beyond, with each feature's critical edges and a representative cocycle. More...

#include <array>
#include <cstdint>
#include <vector>
#include "pulsatrix/tda/distance_matrix.hpp"
#include "pulsatrix/tda/persistence_diagram.hpp"
Include dependency graph for rips_persistence.hpp:
This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Classes

struct  pulsatrix::RipsOptions
 ComputeRipsPersistence()'s settings. More...
 
struct  pulsatrix::CocycleEntry
 One simplex of a cocycle with its coefficient. More...
 
struct  pulsatrix::RipsFeature
 One feature of the diagram, with where it comes from. More...
 
struct  pulsatrix::RipsResult
 ComputeRipsPersistence()'s result. More...
 

Namespaces

namespace  pulsatrix
 

Functions

RipsResult pulsatrix::ComputeRipsPersistence (const DistanceMatrix &distances, const RipsOptions &options={})
 Persistent homology of the Vietoris-Rips filtration on the given distances, in dimensions 0 to options.max_dimension.
 

Detailed Description

Persistent homology of the Vietoris-Rips filtration in every dimension up to a maximum: components (H0), loops (H1), voids (H2) and beyond, with each feature's critical edges and a representative cocycle.

The Vietoris-Rips complex at scale t holds every set of points whose pairwise distances are all at most t; a simplex appears at its diameter, the largest distance among its vertices. As t grows, loops and voids form and fill in. Each one is a pair (birth, death) of a persistence diagram.

The computation follows Bauer's Ripser (J. Appl. Comput. Topol. 2021): persistent cohomology, which pairs the same intervals as homology; simplices indexed by the combinatorial number system and their coboundaries enumerated on demand, never stored; columns reduced from the largest scale down, with clearing (a simplex that kills a class one dimension down can't start one); and columns already reduced when first seen, the common case, settled without a reduction. It is written for pulsatrix rather than vendored, over any prime field Z/p.

Cost grows quickly with the dimension. H1 is practical up to a few thousand points and H2 up to about a thousand; beyond that, choose landmarks first. Memory holds every simplex of the highest dimension asked for below the threshold that a lower one doesn't already account for.