File Coverage

blib/lib/Microarray/GEO/SOFT/GPL.pm

Criterion	Covered	Total	%
statement	12	12	100.0
branch			n/a
condition			n/a
subroutine	4	4	100.0
pod			n/a
total	16	16	100.0

package Microarray::GEO::SOFT::GPL;

# parse the GPL part in GSE file
# or the GPL file itself

use List::Vectorize qw(!table);
use Carp;
use strict;

use base "Microarray::GEO::SOFT";

1;

sub new {

	my $invocant = shift;
	my $class = ref($invocant) || $invocant;
	my $self = { "file" => "",
	             "verbose" => 1,
	             @_ };
	bless($self, $class);

	return $self;

}

sub parse {

	my $self = shift;

	my $fh;
	if(! List::Vectorize::is_glob_ref($self->{file})) {

		open F, $self->{file} or croak "cannot open $self->{file}.\n";
		$fh = \*F;
	}
	else {
		$fh = $self->{file};
	}

	$self->_parse_platform($fh);

	return 1;
}

sub _parse_platform {

	my $self = shift;

	my $fh = shift;

	Microarray::GEO::SOFT::_set_fh($self->{verbose});

	my $accession;
	my $title;
	my $table_colnames = [];
	my $table_rownames = [];
	my $table_matrix = [];

	while(my $line = <$fh>) {

		chomp $line;
		if($line =~/^!Platform_geo_accession = (GPL\d+)$/
		   or $line =~/^!Annotation_platform = (GPL\d+)/) {
			$accession = $1;
		}

		if($line =~/^!Platform_title = (.*?)$/
		   or $line =~/^!Annotation_platform_title = (.*?)$/) {
			$title = $1;
		}

		if($line =~/^!platform_table_begin$/) {

			$line = <$fh>;
			chomp $line;

			@$table_colnames = split "\t", $line, -1;
			shift(@$table_colnames);

			while($line = <$fh>) {

				if($line =~/^!platform_table_end$/) {
					last;
				}

				chomp $line;
				my @tmp = split "\t", $line, -1;

				my $uid = shift(@tmp);

				push(@$table_rownames, $uid);
				push(@$table_matrix, [@tmp]);

			}


		}
		if($line =~/^!platform_table_end$/) {
			last;
		}

	}

	my $n_row = len($table_rownames);
	my $n_col = len($table_colnames);

	my $platform = $accession;

	print "Platform info:\n";
	print "  Accession: $accession\n";
	print "  Platform: $platform\n";
	print "  Title: $title\n";
	print "  Rows: $n_row\n";
	print "  Columns: $n_col\n";
	print "\n";

	$self->set_meta( accession => $accession,
	                 title     => $title,
					 platform  => $platform );
	$self->set_table( rownames => $table_rownames,
	                  colnames => $table_colnames,
					  matrix   => $table_matrix );

	Microarray::GEO::SOFT::_set_to_std_fh();

	return $self;
}

# map new ID from the order of the first column
sub _mapping {

	my $self = shift;
	my $to_id = shift;
	my $from_list = shift;

	my $mapping;

	my $to_index;
	my $colnames = $self->colnames;
	for(my $i = 0; $i < len($colnames); $i ++) {
		if($colnames->[$i] eq $to_id) {
			$to_index = $i;
			last;
		}
	}

	if(! defined($to_index)) {
		croak "ERROR: Cannot find ID ($to_id) in ".$self->platform."\n";
	}

	my $mat = $self->matrix;
	my $hash;
	my $rownames = $self->rownames;
	for(my $i = 0; $i < len($mat); $i ++) {
		if($mat->[$i]->[$to_index] =~/^(.*?)\/\/\//) {
			$hash->{$rownames->[$i]} = $1;
		}
		else {
			$hash->{$rownames->[$i]} = $mat->[$i]->[$to_index];
		}
	}

	for (@$from_list) {
		push(@$mapping, $hash->{$_});
	}

	return $mapping;

}


__END__