File Coverage

blib/lib/Microarray/GEO/SOFT/GSM.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::GSM;

# parse SOFT file
# get the GSM part, only the first record of GSM from the current position to read

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 $self = { "file" => "",
	             "verbose" => 1,
				 "sample_value_column" => "VALUE",
	             @_ };
	bless($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};
	}

	$self->_parse_sample($fh);

	return 1;
}

sub _parse_sample {

	my $self = shift;

	my $fh = shift;

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

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

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

		chomp $line;
		if($line =~/^!Sample_geo_accession = (GSM\d+)$/) {
			$accession = $1;
		}

		elsif($line =~/^!Sample_title = (.*?)$/) {
			$title = $1;
		}

		elsif($line =~/^!Sample_platform_id = (GPL\d+)$/) {
			$platform = $1;
		}

		elsif($line =~/^!sample_table_begin$/) {

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

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

			my $value_index = -1;
			for(my $i = 0; $i < len($table_colnames); $i ++) {
				if($table_colnames->[$i] eq $self->{sample_value_column}) {
					$value_index = $i;
					last;
				}
			}

			if($value_index == -1) {
				croak "ERROR: Cannot find sample value column ($self->{sample_value_column}).";
			}

			while($line = <$fh>) {

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

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

				my $uid = shift(@tmp);

				push(@$table_rownames, $uid);
				# one column matrix
				push(@$table_matrix, [$tmp[$value_index]]);
			}


		}

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

	}

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

	print "Sample 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";

	#my $table_rownames_sorted = sort_array($table_rownames, sub {$_[0] cmp $_[1]});
	#my $table_rownames_sorted_index = order($table_rownames, sub {$_[0] cmp $_[1]});
	#my $table_matrix_sorted = subset($table_matrix, $table_rownames_sorted_index);

	$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;

}


__END__