List titles in subnetworks

Running Pajek from R, List titles

Running in Pajek the sequence:

• read network
• read partition
• extract selected subnetwork with a given class number
• save subnetwork to a file

we get a Pajek log file like this

NETBEGIN 1
CLUBEGIN 1
PERBEGIN 1
CLSBEGIN 1
HIEBEGIN 1
VECBEGIN 1

% Reading Network   ---    C:\Users\batagelj\work\Python\WoS\BM\results\jaccard\Jislands.net
N 1 RDN "C:\Users\batagelj\work\Python\WoS\BM\results\jaccard\Jislands.net" (477)
% Reading Partition   ---    C:\Users\batagelj\work\Python\WoS\BM\results\jaccard\Jislands.clu
C 1 RDC "C:\Users\batagelj\work\Python\WoS\BM\results\jaccard\Jislands.clu" (477)
% Extracting Subnetwork according to Partition
N 2 EXT 1 1 [5] 1 (15)
% Saving network to file   ---    C:\Users\batagelj\work\Python\WoS\BM\results\jaccard\Jislands\J5.net
N 2 WN "C:\Users\batagelj\work\Python\WoS\BM\results\jaccard\Jislands\J5.net" 0 (15)

We transform it into the following “template”

NETBEGIN 1
CLUBEGIN 1
PERBEGIN 1
CLSBEGIN 1
HIEBEGIN 1
VECBEGIN 1

N 1 RDN "network" (477)
C 1 RDC "partition" (477)
N 2 EXT 1 1 [cluster] 1 (15)
N 2 WN "subnetwork" 0 (15)
EXIT

Important: The sequence of commands ends with the EXIT command that exits the Pajek program.

The R function listSubNets for each (non-null) class

• substitutes template “parameters” with the class' values and save it on Pajek.log
• apply Pajek on the sequence of commands in Pajek.log - it extracts a current subnetwork
• apply R function description on the current subnetwork producing a file List<i>.csv

# WoS2Pajek - Extracting subnetworks and listing their titles
# V. Batagelj, May 1st, 2017
# ------------------------------------------------------
 
listSubNets <- function(net,clu,subdir,T){
  cat("WoS2Pajek - Extracting subnetworks and listing their titles\n",
    "V. Batagelj, May 2017\n\n",date(),"\n",sep="")
  t <- table(read.csv(clu,head=FALSE,skip=1)$V1)
  if(names(t)[1]=="0") t <- t[2:length(t)]
  cluA <- paste(getwd(),"/",clu,sep=""); netA <- paste(getwd(),"/",net,sep="")
#  L <- readLines("getSubnet.txt",n=-1)
  L <- c("NETBEGIN 1", "CLUBEGIN 1", "PERBEGIN 1", "CLSBEGIN 1", 
    "HIEBEGIN 1", "VECBEGIN 1", "", "N 1 RDN \"network\" (477)", 
    "C 1 RDC \"partition\" (477)", "N 2 EXT 1 1 [cluster] 1 (15)", 
    "N 2 WN \"subnet\" 0 (15)", "EXIT")                        
  L <- gsub("partition",cluA,gsub("network",netA,L))
  for(i in 1:length(t)) {
    cl <- names(t)[i]; cat("Subnetwork:",cl,t[i],"\n"); flush.console()
    sub <- paste("subNet",cl,".net",sep="")
    subA <- paste(getwd(),subdir,sub,sep="")
    N <- gsub("subnet",subA,gsub("cluster",cl,L))
    writeLines(N,con="Pajek.log")
    system("C:/programi/pajek/pajek.exe",invisible=TRUE,wait=TRUE)
    if(!file.exists(subA)) {
      cat(subA,"*** problems\n"); flush.console(); next
    } else {
      csv <- paste(".",subdir,"List",cl,".csv",sep="")
      d <- description(subA,csv,T)
    }
  }
}

github: description and listSubNets

We first prepare the right environment: load R functions description and listSubNets, set the working directory and read in a data frame T the information about works from the file titles.csv produced by WoS2Pajek:

> source("C:/Users/batagelj/work/Python/WoS/peere1/description.R")
> source("C:/Users/batagelj/work/Python/WoS/BM/listSubNets.R")
> setwd("C:/Users/batagelj/work/Python/WoS/BM/results/jaccard")
> T <- read.csv('../../titles.csv',sep=";",colClasses="character"); T$code <- 1

Now we can run the function listSubNets

> listSubNets("Jislands.net","Jislands.clu","/Jislands/",T)
WoS2Pajek - Extracting subnetworks and listing their titles
V. Batagelj, May 2017
 
Mon May 01 15:44:26 2017
Subnetwork: 1 18 
Subnetwork: 2 32 
Subnetwork: 3 21 
Subnetwork: 4 74 
Subnetwork: 5 15 
Subnetwork: 6 15 
Subnetwork: 7 54 
Subnetwork: 8 20 
Subnetwork: 9 22 
Subnetwork: 10 20 
Subnetwork: 11 22 
Subnetwork: 12 23 
Subnetwork: 13 45 
Subnetwork: 14 26 
Subnetwork: 15 70 
> 

We get the corresponding CSV files List<i>.csv that we can inspect using Excel.