#!/usr/perl/perl580/bin/perl

use strict;
use Algorithm::Cluster qw/kcluster/;

#if ($ARGV[0] || $ARGV[1] || $ARGV[2] || $ARGV[3] eq "") { 
#	die ("This program has 4 ARGV values that must be given:\n
#	ARGV[0] -> file path;\n
#	ARGV[1] -> first collum to be taken;\n
#	ARGV[2] -> last collum to be taken;\n
#	ARGV[3] -> nuber of iteractions of k-means\n");
#}
my $file = "$ARGV[0]";
my $i = 0;
my (@orfname,@orfdata,@weight,@mask);

open(DATA,"<$file") or die "Can't open file $file: $!";


#------------------
# Read in the data file, and save the data to @orfdata
# We know that the file is intact and has no holes, 
# so just set the mask to 1 for every item.
# We don't check for errors in this case, because the file
# is short and we can spot errors by eye. 

my $diferenca = (($ARGV[2] - $ARGV[1])+1); #this will generate how much masks and weights will be generated

my $firstline = <DATA>;  # Skip the title line
while(<DATA>) {

	chomp(my $line = $_);
	my @field = split /\t/, $line;
	$orfname[$i] = $field[0];
#	print $orfname[$i];
	$orfdata[$i] = [@field[$ARGV[1]..$ARGV[2]]]; ##selecting what columms will be taken by ARGV;
#	print $orfdata[$i];
#	$a = <STDIN>;
	$mask[$i] = (1.0) x $diferenca;
	$i++;
}
close(DATA);

#------------------
# Make a reverse-lookup index of the @orfnames hash:
#
my %orfname_by_rowid;
$i=0;
$orfname_by_rowid{$i++} = $_, foreach(@orfname);

@weight = (1.0) x $diferenca;


#------------------
# Define the params we want to pass to kcluster
my %params = (
	nclusters =>         2,
	transpose =>         1,
	npass     =>       $ARGV[3], #how much interactions will happen
	method    =>       'a',
	dist      =>       'e',
	data      =>    \@orfdata,
	mask      =>       \@mask,
	weight    =>     \@weight,
);


#------------------
# Here is where we invoke the library function!
#
my ($clusters, $centroids, $error, $found) = kcluster(%params);
#
#------------------



#------------------
# Create a reverse index of the ORF names,  by cluster ID
#
my %orfname_by_cluster;
$i=0;
foreach(@{$clusters}) {
	push @{$orfname_by_cluster{$_}}, $orfname_by_rowid{$i++};
}


#------------------
# Print out a list of the ORFs, grouped by cluster ID,
# as returned by the kcluster() function.
#
$i=0;
foreach(@{$centroids}) {

	print "------------------\n";
	printf("Cluster %d:  %d ORFs,  centroids: ( %s )\n\n", 
		$i, scalar(@{$orfname_by_cluster{$i} }),
		join ', ', map { sprintf "%.3f",$_ } @{$_}
	);

	print "\t$_\n", foreach( sort { $a cmp $b } @{$orfname_by_cluster{$i} } );
	print "\n";
	++$i;
}