#!/usr/bin/perl
#
#  Usage: processanda infile [outfile a b]
#
#  infile has a N x K (K >= 2) matrix of (other than the first two columns) numbers 
#   The a-th and b-th columns of infile have nodenames. (a and b are assumed to be 0 and 1 by default ... indexing starts at 0)
#
#   It creates the following files:
#     outfile.label, a file whose n-th line (n>=1) has the label of 
#       the n-th node found in infile
#     outfile.ncol  is a N x 2 matrix representing an undirected graph
#       with no repeated edges or loops.
#       It's suitable to be used as an .ncol file for 
#       submission to http://orion.icmb.utexas.edu/cgi-bin/lgl/web_lgl.pl
#
# Dinoj Surendran (dinojATATATATATcs.uchicago.edu)

my %nodenames;
my %names;

$infile = $ARGV[0];
$outfile = $infile; $outfile =~ s/\./_/g;
if ($#ARGV >= 1)
{ $outfile = $ARGV[1]; }

$col1 = 0; 
$col2 = 1;
if ($#ARGV >= 2)
{ $col1 = $ARGV[2]; }
if ($#ARGV >= 3)
{ $col2 = $ARGV[3]; }

open (IN,$infile);
open (LAB,">".$outfile.".label");

$count = 0;
while ($line = <IN>)
{
    @tmp = split /[\n\t\r\ ]+/, $line;         # important to remove \r too!
    $n1 = $tmp[$col1];
    $n2 = $tmp[$col2];
    if (! exists($nodenames{$n1}))
    {
	$count++;
	$nodenames{$n1} = $count;
	print LAB "$n1 \n";
    }
    if (! exists($nodenames{$n2}))
    {
	$count++;
	$nodenames{$n2} = $count;
	print LAB "$n2 \n";
    }
}

close (IN);
open (OUT,">".$outfile.".ncol");

open (IN,$infile);
my %existingedges;


$count = 0;
while ($line = <IN>) 
{ 
    @tmp = split /[\n\t\r\ ]+/, $line; 
    $n1 = $nodenames{$tmp[$col1]}; 
    $n2 = $nodenames{$tmp[$col2]};
    
    $edgename = $n1."_".$n2;
    if ($n1 > $n2) { $edgename = $n2."_".$n1; }

    if (! exists ($existingedges{$edgename}))
    {
 	print OUT "$n1 $n2 \n"; 
	$existingedges{$edgename} = 1;
    }
}