Package 'interplex'

Title: Coercion Methods for Simplicial Complex Data Structures
Description: Computational topology, which enables topological data analysis (TDA), makes pervasive use of abstract mathematical objects called simplicial complexes; see Edelsbrunner and Harer (2010) <doi:10.1090/mbk/069>. Several R packages and other software libraries used through an R interface construct and use data structures that represent simplicial complexes, including mathematical graphs viewed as 1-dimensional complexes. This package provides coercers (converters) between these data structures. Currently supported structures are complete lists of simplices as used by 'TDA'; the simplex trees of Boissonnat and Maria (2014) <doi:10.1007/s00453-014-9887-3> as implemented in 'simplextree' and in Python GUDHI (by way of 'reticulate'); and the graph classes of 'igraph' and 'network', by way of the 'intergraph' package.
Authors: Jason Cory Brunson [aut, cre] , Yara Skaf [ctb]
Maintainer: Jason Cory Brunson <[email protected]>
License: GPL (>= 3)
Version: 0.1.2
Built: 2025-02-15 05:10:53 UTC
Source: https://github.com/tdaverse/interplex

Help Index

Coerce objects to lists of simplices

Description

Coerce objects to lists of simplices, as used by the TDA package.

Usage

as_cmplx(x, ...)

## Default S3 method:
as_cmplx(x, ...)

## S3 method for class 'Rcpp_SimplexTree'
as_cmplx(x, ...)

## S3 method for class 'gudhi.simplex_tree.SimplexTree'
as_cmplx(x, ...)

## S3 method for class 'igraph'
as_cmplx(x, index = NULL, ...)

## S3 method for class 'network'
as_cmplx(x, index = NULL, ...)

Arguments

x

An R object to be coerced. See Details.
...

Additional arguments passed to methods.
index

Integer-valued vertex attribute to be used as 0-simplex indices. Ignored if NULL (the default).

Details

as_cmplx() is a generic function with specific methods for different simplicial complex S3 classes. It returns a list of integer vectors, each of which represents a simplex, and all simplices are included in the list. When a filtration is constructed using ⁠TDA::*Filtration()⁠, the first named element of the returned list, cmplx, is a list whose ith element contains the vertices of the ith simplex.

Value

A list of integer vectors, each encoding one simplex.

Examples

if (requireNamespace("TDA", quietly = TRUE)) {
  # pick the simplicial complex from a TDA filtration object
  t <- 2 * pi * c(0, 1, 3, 6) / 7
  rf <- TDA::ripsFiltration(
    cbind(x = cos(t), y = sin(t)),
    maxdimension = 2L, maxscale = 1.7
  )
  print(rf$cmplx)
  cp_rf <- as_cmplx(rf)
  print(cp_rf)
}

if (requireNamespace("simplextree", quietly = TRUE)) {
  # convert a simplextree object
  st <- simplextree::simplex_tree()
  st$insert(list(3:5, 5:6, 8))
  cp_st <- as_cmplx(st)
  print(cp_st)
}

## Not run: 

if (requireNamespace("reticulate", quietly = TRUE)) {
  # convert a Python GUDHI simplex tree
  gd <- reticulate::import("gudhi")
  gd_st <- gd$SimplexTree()
  for (s in list(3:5, 5:6, 8)) gd_st$insert(as.list(s))
  cp_gd <- as_cmplx(gd_st)
  print(cp_gd)
}

## End(Not run)

if (requireNamespace("igraph", quietly = TRUE)) {
  # convert an igraph object
  ig <- igraph::graph(c(1,2, 2,3, 1,3, 3,4))
  set.seed(0L)
  ig <- igraph::set_vertex_attr(ig, "id", value = sample(igraph::vcount(ig)))
  print(ig)
  cp_ig <- as_cmplx(ig, index = "id")
  print(cp_ig)
}

if (requireNamespace("network", quietly = TRUE)) {
  # convert a network object
  el <- data.frame(tails = c(1, 2, 1, 3), heads = c(2, 3, 3, 4))
  nw <- network::network.edgelist(el, network::network.initialize(4))
  print(nw)
  cp_nw <- as_cmplx(nw)
  print(cp_nw)
}

Coerce objects to class 'igraph'

Description

Coerce objects to 'igraph' objects, as implemented in the igraph package.

Usage

as_igraph(x, ...)

## Default S3 method:
as_igraph(x, index = NULL, ...)

## S3 method for class 'Rcpp_SimplexTree'
as_igraph(x, index = NULL, ...)

## S3 method for class 'gudhi.simplex_tree.SimplexTree'
as_igraph(x, index = NULL, ...)

## S3 method for class 'igraph'
as_igraph(x, ...)

## S3 method for class 'network'
as_igraph(x, ...)

Arguments

x

An R object to be coerced. See Details.
...

Additional arguments passed to methods.
index

Character string to be added as a vertex attribute containing 0-simplex indices. Ignored if NULL (the default).

Details

as_igraph() is a generic function with specific methods for different simplicial complex S3 classes. It returns an igraph object.

Value

An object of class 'igraph'.

Examples

if (requireNamespace("TDA", quietly = TRUE)) {
  # convert a TDA filtration object
  t <- 2 * pi * c(0, 1, 3, 6) / 7
  rf <- TDA::ripsFiltration(
    cbind(x = cos(t), y = sin(t)),
    maxdimension = 2L, maxscale = 1.7
  )
  print(rf$cmplx)
  ig_rf <- as_igraph(rf)
  print(ig_rf)
  ig_rf2 <- as_igraph(rf$cmplx)
  print(ig_rf2)
}

if (requireNamespace("simplextree", quietly = TRUE)) {
  # convert a simplextree object
  st <- simplextree::simplex_tree()
  st$insert(list(3:5, 5:6, 8))
  ig_st <- as_igraph(st)
  print(ig_st)
}

## Not run: 

if (requireNamespace("reticulate", quietly = TRUE)) {
  # convert a Python GUDHI simplex tree
  gd <- reticulate::import("gudhi")
  gd_st <- gd$SimplexTree()
  for (s in list(3:5, 5:6, 8)) gd_st$insert(as.list(s))
  ig_gd <- as_igraph(gd_st, index = "id")
  print(ig_gd)
}

## End(Not run)

if (requireNamespace("network", quietly = TRUE)) {
  # convert a network object
  el <- data.frame(tails = c(1, 2, 1, 3), heads = c(2, 3, 3, 4))
  nw <- network::network.edgelist(el, network::network.initialize(4))
  print(nw)
  ig_nw <- as_igraph(nw)
  print(ig_nw)
}

Coerce objects to class 'network'

Description

Coerce objects to 'network' objects, as implemented in the network package.

Usage

as_network(x, ...)

## Default S3 method:
as_network(x, index = NULL, ...)

## S3 method for class 'Rcpp_SimplexTree'
as_network(x, index = NULL, ...)

## S3 method for class 'gudhi.simplex_tree.SimplexTree'
as_network(x, index = NULL, ...)

## S3 method for class 'igraph'
as_network(x, ...)

## S3 method for class 'network'
as_network(x, ...)

Arguments

x

An R object to be coerced. See Details.
...

Additional arguments passed to methods.
index

Character string to be added as a vertex attribute containing 0-simplex indices. Ignored if NULL (the default).

Details

as_network() is a generic function with specific methods for different simplicial complex S3 classes. It returns a network object.

Value

An object of class 'network'.

Examples

if (requireNamespace("TDA", quietly = TRUE)) {
  # convert a TDA filtration object
  t <- 2 * pi * c(0, 1, 3, 6) / 7
  rf <- TDA::ripsFiltration(
    cbind(x = cos(t), y = sin(t)),
    maxdimension = 2L, maxscale = 1.7
  )
  print(rf$cmplx)
  nw_rf <- as_network(rf)
  print(nw_rf)
  nw_rf2 <- as_network(rf$cmplx)
  print(nw_rf2)
}

if (requireNamespace("simplextree", quietly = TRUE)) {
  # convert a simplextree object
  st <- simplextree::simplex_tree()
  st$insert(list(3:5, 5:6, 8))
  nw_st <- as_network(st)
  print(nw_st)
}

## Not run: 

if (requireNamespace("reticulate", quietly = TRUE)) {
  # convert a Python GUDHI simplex tree
  gd <- reticulate::import("gudhi")
  gd_st <- gd$SimplexTree()
  for (s in list(3:5, 5:6, 8)) gd_st$insert(as.list(s))
  nw_gd <- as_network(gd_st, index = "id")
  print(nw_gd)
}

## End(Not run)

if (requireNamespace("igraph", quietly = TRUE)) {
  # convert an igraph object
  ig <- igraph::graph(c(1,2, 2,3, 1,3, 3,4))
  print(ig)
  nw_ig <- as_network(ig)
  print(nw_ig)
}

Coerce objects to Python GUDHI simplex trees

Description

Coerce objects to 'SimplexTree' objects in Python GUDHI, accessed via the reticulate package.

Usage

as_py_gudhi_simplextree(x, ...)

## Default S3 method:
as_py_gudhi_simplextree(x, ...)

## S3 method for class 'Rcpp_SimplexTree'
as_py_gudhi_simplextree(x, ...)

## S3 method for class 'igraph'
as_py_gudhi_simplextree(x, index = NULL, ...)

## S3 method for class 'network'
as_py_gudhi_simplextree(x, index = NULL, ...)

Arguments

x

An R object to be coerced. See Details.
...

Additional arguments passed to methods.
index

Integer-valued vertex attribute to be used as 0-simplex indices. Ignored if NULL (the default).

Details

as_py_gudhi_simplextree() is a generic function with specific methods for different simplicial complex S3 classes. It returns an object of class 'gudhi.simplex_tree.SimplexTree', which is a reticulate accessor to a Python object of class 'SimplexTree' implemented in GUDHI.

Value

A simplex tree instantiated in Python GUDHI accessed through reticulate.

Author(s)

Jason Cory Brunson

Yara Skaf

Examples

## Not run: 

if (requireNamespace("reticulate", quietly = TRUE)) {
  # print GUDHI simplices
  print_py_gudhi <- function(x) {
    reticulate::iterate(
      x$get_skeleton(x$dimension()),
      function(s) print(s[[1]]),
      simplify = FALSE
    )
  }
}

if (requireNamespace("TDA", quietly = TRUE)) {
  # convert a TDA filtration object
  t <- 2 * pi * c(0, 1, 3, 6) / 7
  rf <- TDA::ripsFiltration(
    cbind(x = cos(t), y = sin(t)),
    maxdimension = 2L, maxscale = 1.7
  )
  print(rf$cmplx)
  gd_rf <- as_py_gudhi_simplextree(rf)
  print_py_gudhi(gd_rf)
  gd_rf2 <- as_py_gudhi_simplextree(rf$cmplx)
  print_py_gudhi(gd_rf2)
}

if (requireNamespace("simplextree", quietly = TRUE)) {
  # convert a simplextree object
  st <- simplextree::simplex_tree()
  st$insert(list(3:5, 5:6, 8))
  gd_st <- as_py_gudhi_simplextree(st)
  print_py_gudhi(gd_st)
}

if (requireNamespace("igraph", quietly = TRUE)) {
  # convert an igraph object
  ig <- igraph::graph(c(1,2, 2,3, 1,3, 3,4))
  print(ig)
  gd_ig <- as_py_gudhi_simplextree(ig)
  print_py_gudhi(gd_ig)
  
  # specify 0-simplex indices
  set.seed(0L)
  ig <- igraph::set_vertex_attr(ig, "id", value = sample(igraph::vcount(ig)) + 1L)
  igraph::V(ig)$id
  igraph::as_edgelist(ig)
  gd_ig2 <- as_py_gudhi_simplextree(ig, index = "id")
  print_py_gudhi(gd_ig2)
}

if (requireNamespace("network", quietly = TRUE)) {
  # convert a network object
  el <- data.frame(tails = c(1, 2, 1, 3), heads = c(2, 3, 3, 4))
  nw <- network::network.edgelist(el, network::network.initialize(4))
  print(nw)
  gd_nw <- as_py_gudhi_simplextree(nw)
  print_py_gudhi(gd_nw)
}

## End(Not run)

Coerce objects to class 'Rcpp_SimplexTree'

Description

Coerce objects to 'Rcpp_SimplexTree' objects, as implemented in the simplextree package.

Usage

as_rcpp_simplextree(x, ...)

## Default S3 method:
as_rcpp_simplextree(x, ...)

## S3 method for class 'Rcpp_SimplexTree'
as_rcpp_simplextree(x, ...)

## S3 method for class 'gudhi.simplex_tree.SimplexTree'
as_rcpp_simplextree(x, ...)

## S3 method for class 'igraph'
as_rcpp_simplextree(x, index = NULL, ...)

## S3 method for class 'network'
as_rcpp_simplextree(x, index = NULL, ...)

Arguments

x

An R object to be coerced. See Details.
...

Additional arguments passed to methods.
index

Integer-valued vertex attribute to be used as 0-simplex indices. Ignored if NULL (the default).

Details

as_rcpp_simplextree() is a generic function with specific methods for different simplicial complex S3 classes. It returns an object of class 'Rcpp_SimplexTree', which is an Rcpp Module that exposes an instance of a C++ instance of a simplex tree.

Value

An instance of a simplex tree, exposed as an Rcpp Module with class 'Rcpp_SimplexTree'.

Examples

if (requireNamespace("TDA", quietly = TRUE)) {
  # convert a TDA filtration object
  t <- 2 * pi * c(0, 1, 3, 6) / 7
  rf <- TDA::ripsFiltration(
    cbind(x = cos(t), y = sin(t)),
    maxdimension = 2L, maxscale = 1.7
  )
  print(rf$cmplx)
  st_rf <- as_rcpp_simplextree(rf)
  print(st_rf)
  st_rf2 <- as_rcpp_simplextree(rf$cmplx)
  print(st_rf2)
}

## Not run: 

if (requireNamespace("reticulate", quietly = TRUE)) {
  # convert a Python GUDHI simplex tree
  gd <- reticulate::import("gudhi")
  gd_st <- gd$SimplexTree()
  for (s in list(3:5, 5:6, 8)) gd_st$insert(as.list(s))
  st_gd <- as_rcpp_simplextree(gd_st)
  st_gd$as_list()
}

## End(Not run)

if (requireNamespace("igraph", quietly = TRUE)) {
  # convert an igraph object
  ig <- igraph::graph(c(1,2, 2,3, 1,3, 3,4))
  print(ig)
  st_ig <- as_rcpp_simplextree(ig)
  print(st_ig)
  
  # specify 0-simplex indices
  set.seed(0L)
  ig <- igraph::set_vertex_attr(ig, "id", value = sample(igraph::vcount(ig)) + 1L)
  igraph::V(ig)$id
  igraph::as_edgelist(ig)
  st_ig <- as_rcpp_simplextree(ig, index = "id")
  st_ig$vertices
  st_ig$edges
}

if (requireNamespace("network", quietly = TRUE)) {
  # convert a network object
  el <- data.frame(tails = c(1, 2, 1, 3), heads = c(2, 3, 3, 4))
  nw <- network::network.edgelist(el, network::network.initialize(4))
  print(nw)
  st_nw <- as_rcpp_simplextree(nw)
  print(st_nw)
}

interplex package

Description

This is a helper package to coerce simplicial complexes between different data structures.

Details

This package helps interface between different topological data analytic packages and workflows by coercing simplicial complex objects stored using different data structures. Each coercion is designed to retain as much annotation as possible, whether of simplices or of the complex.

The package supports coercions between simplicial complexes stored using the following data structures:

Author(s)

Maintainer: Jason Cory Brunson [email protected] (ORCID)

Other contributors:

  • Yara Skaf [contributor]

See Also

Useful links: