R/internal_split_mult_node_refs.R

In grishagin/RIGplotbiopax: Plot BioPAX Pathways

internal_split_mult_node_refs<-
  function(nodes_edges_list
           ,ntype
           ,verbose=FALSE){
    
    #' @keywords internal

    allnodes<-
      nodes_edges_list[[1]]
    alledges<-
      nodes_edges_list[[2]]
    #find all "to" nodes of the desired type
    #make individual ids for them
    # alledges[type %in% ntype
    #          ,newto := paste0(paste(substr(ntype,1,2)
    #                                 ,collapse="")
    #                           ,1:length(to))]
    alledges$newto<-NA_character_
    alledges[type %in% ntype]$newto<-
      alledges[type %in% ntype
               ,.(newto=paste0(paste(substr(ntype,1,2)
                                     ,collapse="")
                               ,1:length(to)))]$newto
    #and a dictionary
    #also merge the dict by the old id
    #such that there are no repetitions
    replace_dict<-
      alledges[type %in% ntype
               ,.(to
                  ,newto)] %>% 
      merge_cols_shorten_df(colKey="to"
                            ,patternToMerge=";")
    
    #replace "to" ids and remove auxiliary column
    # alledges[!is.na(newto)
    #          ,to := newto]
    alledges[!is.na(newto)]$to<-
      alledges[!is.na(newto)]$newto
    
    alledges$newto<-NULL
    
    #replace the "from" nodes' ids
    alledges$from<-
      alledges$from %>% 
      mapvalues(from=replace_dict$to
                ,to=replace_dict$newto
                ,warn_missing = verbose)
    
    #now split up the newly replaced from ids
    alledges<-
      alledges %>% 
      split_cols_lengthen_df(colsToSplit = "from"
                             ,patternToSplit = ";"
                             ,at_once = FALSE) %>% 
      as.data.table
    
    
    #do the same thing to nodes df
    allnodes$id<-
      allnodes$id %>% 
      mapvalues(from=replace_dict$to
                ,to=replace_dict$newto
                ,warn_missing = verbose)
    
    allnodes<-
      allnodes %>% 
      split_cols_lengthen_df(colsToSplit = "id"
                             ,patternToSplit = ";") %>% 
      as.data.table
    
    return(list(allnodes=unique(allnodes)
                ,alledges=unique(alledges)))
  }