R/get_edge_betweenness.R

In mselensky/bngal: Biological Network Graphing and Learning

#' Return edge between cluster data
#'
#' This is a wrapper around [`igraph::cluster_edge_betweenness()`]. This can
#' be a memory- and time-intensive operation, especially for lower taxonomic
#' levels. In this wrapper, edge betweenness is processed in parallel for each
#' subcommunity by default in an attempt to ease runtime requirements. For larger
#' matrices of a single community, consider using [`parallel::mcmapply()`] around
#' `igraph::cluster_edge_betweenness()` to increase computational efficiency
#' (support coming soon).
#'
#' Before running this function, consider
#'
#' @param igraph.list Output from [`bngal::get_igraph()`]
#'
#' @return
#' @export
#'
#' @examples
#' get_igraph_output <- bngal::get_igraph(prepro_data)
#' get_edge_betweenness(get_igraph_output)
get_edge_betweenness <- function(igraph.list) {

  NCORES <- bngal::check_cores()

  # better multicore performance currently being worked on, for now this placeholder:
  cluster_eb <- function (igraph.list) {
    igraph::cluster_edge_betweenness(igraph.list)
  }

  input.data.class = c("tbl_df", "tbl", "data.frame", "igraph")
  if (class(igraph.list) == "list") {
    edge_betweenness <- mclapply(X = igraph.list,
                                 FUN = cluster_eb,
                                 mc.cores = NCORES)
  } else if (class(igraph.list) %in% input.data.class) {
    edge_betweenness <- cluster_eb(igraph.list)
    list(all=edge_betweenness)
  } else {
    stop("\n | [", Sys.time(), "] Unexpected input data.\n",
    " |   ->Requires output from bngal::get_igraph()")
  }

}