R/load_emln.R

Defines functions load_emln

Documented in load_emln

#' Load networks include din the package
#'
#' @name load_emln
#'
#' @description Load a specific network by specifying its ID.
#'
#' @param network_id The network ID.
#'
#' @return A multilayer object (see ?multilayer).
#'
#' @details
#' Use the \code{search_emln} and \code{view_emln} functions to obtain network IDs.
#'
#' @examples
#' \donttest{
#' load_emln(network_id = 38)
#' }
#'
#' @seealso \code{multilayer}
#'
#' @export load_emln
#' @import dplyr
#' @import tidyr
#' @importFrom hablar retype
#' @importFrom stringr str_detect
#' @importFrom rlang .data

load_emln <- function(network_id) {
  #get the network name in order to load it -- depending on the network id entered
  #FOR DEBUG: network_name <- list.files('./data/',pattern = paste0('emln',as.character(network_id),'_'))
  pattern <- paste0("emln", network_id, "_")
  network_name <- descriptions %>% dplyr::filter(stringr::str_detect(.data$network_name, pattern)) %>% dplyr::slice(1) %>% dplyr::pull(.data$network_name)
  if (length(network_name) == 0) {
    return(paste0('Error: Network id',network_name, ' does not exist'))
  } else {
    #FOR DEBUG: network <- loadRData(paste0('./data/',network_name))
    network <- get(network_name)
  }

  #Pivot nodes to a wide format
  network$nodes %<>%
    tidyr::pivot_wider(names_from = 'attribute',values_from = 'value', values_fn = list) %>%
    tidyr::unnest(cols = everything())
    #fix the column types (so not all columns are character type)
    for (i in colnames(network$nodes)) {network$nodes[[i]] <- hablar::retype(network$nodes[[i]])}

  #if the same name id has different node id
  nodes_vector <- network$nodes[,c(1,2)] %>% dplyr::group_by(.data$node_id, .data$node_name) %>% dplyr::summarise()
  nodes_vector<- nodes_vector$node_name
  # nodes_vector <- network$nodes$node_name
  if(any(duplicated(nodes_vector))){
    message('WARNING! Some node names appear more than once:')
    message(nodes_vector[which(duplicated(nodes_vector))])
  }


  #Pivot layers to a wide format
  network$layers %<>%
    tidyr::pivot_wider(names_from = 'attribute', values_from = 'value', values_fn = list) %>%
    #if there are multiple types to the same layer, write them as individual rows and not lists
    tidyr::unnest(cols = everything()) %>%
    #remove type column
    dplyr::select(-"type") %>%
    dplyr::distinct() %>%
    # The layer_id and layer_name are necessary because they are used by functions like get_sam
    dplyr::rename(layer_id=layer) %>%
    dplyr::mutate(layer_name=paste('layer_',layer_id,sep=''))

  #fix the column types
  for (i in colnames(network$layers)) {
    if('date' %in% tolower(i)) {
      if(all(grepl(pattern = '/', x = network$layers$date))) { #if there are dates that are written with '/' (wrong format to fix)
        network$layers$date <- strptime(as.character(network$layers$date), "%d/%m/%Y") #fix the format
        format(network$layers$date, "%Y-%m-%d")
      }
    }
    network$layers[[i]] <- hablar::retype(network$layers[[i]])
  }

  if(any(duplicated(network$layers$layer_id))){
    message('WARNING! Some layer names appear more than once!')
  }

  #Pivot links to a wide format
  network$interactions %<>%
    tidyr::pivot_wider(names_from = 'attribute', values_from = 'value', values_fn = list) %>%
     # dplyr::select(interaction_id,layer_from,node_from,layer_to,node_to,weight) %>%
     tidyr::unnest(cols = everything()) %>%
     dplyr::select(-"interaction_id") %>%
    dplyr::select("layer_from", "node_from", "layer_to", "node_to", everything())

  #layer from and to are always ids
  #if the layer from and to are not in the layer ids, then convert to ids in edge list
  if (all(!(unique(network$interactions$layer_from) %in%  unique(network$layers$layer_id)))) {
  #if (!all(grepl(x = network$interactions$layer_from,pattern = '^[0-9]*$'))) {
  for (i in 1:nrow(network$interactions)) {
     network$interactions$layer_from[i] <- network$layers[network$layers$name == network$interactions$layer_from[i],]$layer_id
     }
}
  if (all(!(unique(network$interactions$layer_to) %in%  unique(network$layers$layer_id)))) {
  #if (!all(grepl(x = network$interactions$layer_to,pattern = '^[0-9]*$'))) {
    for (i in 1:nrow(network$interactions)) {
      network$interactions$layer_to[i] <- network$layers[network$layers$name == network$interactions$layer_to[i],]$layer_id
    }
  }

  #fix the column types
  for (i in colnames(network$interactions)) {network$interactions[[i]] <- hablar::retype(network$interactions[[i]])}

  # If the layers are ids, change them to layer_name
  if (all(grepl(x = network$interactions$layer_from,pattern = '^[0-9]*$'))) {network$interactions$layer_from <- paste('layer_',network$interactions$layer_from,sep='')}
  if (all(grepl(x = network$interactions$layer_to,pattern = '^[0-9]*$'))) {network$interactions$layer_to <- paste('layer_',network$interactions$layer_to,sep='')}

  # Create a map of state nodes
  message('Creating state node map')
  state_nodes_map <-
  dplyr::bind_rows(network$interactions %>% dplyr::distinct(layer_name=.data$layer_from, node_name=.data$node_from),
            network$interactions %>% dplyr::distinct(layer_name=.data$layer_to, node_name=.data$node_to)) %>%
    dplyr::distinct() %>%
    dplyr::left_join(network$nodes, by='node_name') %>%
    dplyr::left_join(network$layers %>% dplyr::select("layer_id", "layer_name"), by='layer_name') %>%
    dplyr::select("layer_id", "node_id", "layer_name", "node_name", everything())

  #Pivot state nodes to a wide format if more data on state nodes exist
  if ('state_nodes' %in% names(network)) {
  network$state_nodes %<>%
      tidyr::pivot_wider(names_from = 'attribute',values_from = 'value', values_fn = list) %>%
      tidyr::unnest(cols = everything())
  message('Joining with state node information in the data set')
  state_nodes_map %<>% dplyr::left_join(network$state_nodes %>% dplyr::select(-"node_name"), by=c('layer_id', 'node_id'))
  }

  ###  conversion of the node names to node ids in the edge list  ###
  message('Creating extended link list with node IDs')
  extended_edge_list <- extended_edge_list_ids <- as.data.frame(network$interactions)

  i=1
  for (i in 1:nrow(extended_edge_list)) {
    extended_edge_list_ids[i,2] <- unique(network$nodes[network$nodes$node_name == extended_edge_list[i,2],]$node_id)
    extended_edge_list_ids[i,4] <- unique(network$nodes[network$nodes$node_name == extended_edge_list[i,4],]$node_id)
    # Change layer names for ids
    extended_edge_list_ids[i,1] <- network$layers[network$layers$layer_name == extended_edge_list[i,1],]$layer_id
    extended_edge_list_ids[i,3] <- network$layers[network$layers$layer_name == extended_edge_list[i,3],]$layer_id
  }

  #fix the column type in extended_edge_list_ids
  for (i in colnames(extended_edge_list_ids)) {extended_edge_list_ids[[i]] <- hablar::retype(extended_edge_list_ids[[i]])}


  out <- list(nodes=network$nodes,
              layers=network$layers,
              extended=extended_edge_list,
              extended_ids=extended_edge_list_ids,
              state_nodes=state_nodes_map,
              description=network$description,
              references=network$references)
  class(out) <- 'multilayer'

  return(out)
}


# #check if there are networks that return an error when loading the emln
# errors_ids <- NULL
# for (network_id in 1:78){result <- tryCatch(expr = load_emln(network_id), error = function(e) {errors_ids<<-c(errors_ids,network_id)})}

Try the emln package in your browser

Any scripts or data that you put into this service are public.

emln documentation built on May 6, 2026, 1:06 a.m.