#!/usr/bin/perl -w

# INPUT: .cdf file
# OUTPUT: .psc file
# creates table:    probesetid    x1   y1      (PM only)
#                       "         x2   y2
#                       "         x3   y3
#
# assumes each PM (at y) has corresponging MM (at y+1)

use strict;
use Getopt::Long;

my ($cdf_file,$psc_file);
my %probeset;

&GetOptions("cdf=s"    => \$cdf_file,
            "out=s"    => \$psc_file);
($cdf_file) && ($psc_file) || die "Usage: cdf2psc OPTIONS

where OPTIONS are:
                  -cdf   .CDF FILE
                  -out   output (.PSC FILE)
\n";

open (CDF, $cdf_file) || die "Can't open $cdf_file: $!\n";
foreach () {
    if (/^Cell/ && !/^CellHeader/) {
	my ($x,$y,$probe) = (split /\t/,$_)[0,1,4];
	($x) = (split /=/,$x)[1];
	if ($probe =~ /\D/ && /\S/ ) {
	    $probeset{$probe}{$x}{$y}++;
	}
    }
}
close (CDF) || die "Can't close $cdf_file: $!\n";
# filters out .cdf header and pairs each x,y coordinate pair with a 
probeset id
# probeset must contain a non digit, non whitespace character

#$cdf_file =~ s/\.cdf$//i;

open (PSC, ">$psc_file") || die "Can't create $psc_file: $!\n";

my ($probekey, $xkey, $ykey);
my $z = 0;

foreach $probekey (sort keys %probeset) {
    if ($probekey eq "") {
	next;
    }
    foreach $xkey (sort {$a <=> $b} keys %{$probeset{$probekey}}) {
	foreach $ykey (sort {$a <=> $b} keys 
%{$probeset{$probekey}{$xkey}}) {
	    if ($z == 0) {
		print PSC "$probekey\t$xkey\t$ykey\n"; #PM 
		$z = 1;
	    } else {
		$z = 0; #skip MM
	    }
	}
    }
}
close (PSC) || die "Can't close $psc_file: $!\n";
# sorts by probeset id, then x coordinate, then y coordinate
# prints to psc_file only if probeset id exists
# prints only PM cells