#!/usr/bin/perl
# 2009-04-09
# Converts between the R output format and the XML format for clustering (for Ken).
# Bryan Bishop (kanzure@gmail.com) http://heybryan.org/ Automated Design Lab UT Austin

# Note: generate list.txt via: ls -1 *.dat > list.txt

open(FIL, "<list.txt"); # Open a list of files.
@myfiles = <FIL>;
close(FIL);

open(IND, "<index.txt"); # Generic index where the line number that a candidates000.dsm is listed represents the line number that its sum was given in the other text files that were used to generate the kmeans clustering data.
@indices = <IND>;
close(IND);
chop(@indices); # get rid of the newline at the end of each line (we have no need for its kind here).

print "the size of the indices array is: " . $#indices . "\n\n\n";

#$ARGV[0]; # Let's set this to be the initial part of the file name- all the way up to the "$i.dat"i 
if(!$ARGV[0]) { print "ERROR! ARGV0?!?!"; exit(); }

foreach $filename (@myfiles) {
	if(!($filename =~ /$ARGV[0]/)) { next; } # we're reading through "list.txt" and checking for the relevant filenames. cheap hack. ignore.
	open(TFIL, "<$filename");
	@lines = <TFIL>;
	close(TFIL);
	@overallvalues = "";
	@values = "";
	foreach my $line (@lines) {
		chop($line); # get rid of trailing characters.
		#print "the line is: $line\n";
		@values = split(/ /, $line);
		foreach my $val (@values) {
			print "val: $val ($#overallvalues)\n";
			push(@overallvalues, $val);
		}
	}
	#print "size of the array is: $#overallvalues\n";
	# Now correlate the index of a value in overallvalues with a name in 50-most-different/waterlifting-output-50-most-different.txt (or whatever) (same line number)
	# <cluster clusterid="1"><member>candidates000.dsm</member></cluster>
	# How many clusters exist in this file? Well, check the filename. 
	$numberOfClusters = $filename;
	$numberOfClusters =~ s/$ARGV[0]//;
	$numberOfClusters =~ s/\.dat//;
	chop($numberOfClusters);
	#print "numberOfClusters = $numberOfClusters\n";
	$filename2 = $filename;
	chop($filename2);
	open (OUTP, ">$filename2.cluster.xml");
	print "------ starting a new clusterset (w/ $numberOfClusters clusters)-----------\n";
	for($i = 1; $i <= $numberOfClusters; $i++) {
		$j = 0;
		#open (OUTP, ">$filename.clustering.xml");
		print OUTP "<cluster clusterid=\"$i\">\n";
		print "<cluster>\n";
		foreach my $subval (@overallvalues) {
			if($subval == $i) {
				# current cluster ID: $i
				# current place in the array: $j;
				# so, the name is then: $indices[j]
				#chomp($indices[$j]);
				print OUTP "<member>$indices[$j]</member>\n";
				print "value of \$j is $j\n";
			}
			$j++;
			#print "value of \$j is $j\n";
		}
		print "</cluster>\n";
		print OUTP "</cluster>\n";
	
	}
	close(OUTP);
	if($#overallvalues == 0) { print "filename: $filename\n"; }
}