#!/usr/bin/perl

use Getopt::Long;
use Pod::Usage;
use strict;

my $out_file = "";
my @tmp_dirs = ();
my $help = 0;

GetOptions("o=s" => \$out_file,
	   "T|temporary-directory=s" => \@tmp_dirs,
	   "h|help" => \$help) or pd2usage(-exitval => 2, -verbose => 2);

	   
pod2usage(-verbose => 2) if ($help == 1);
pod2usage(-msg => "Invalid number of arguments!", -exitval => 2, -verbose => 2) if (scalar(@ARGV) != 1);

my $command;

# grep header section
$command = "grep ^@ $ARGV[0]";
if ($out_file ne "") { $command .= " > $out_file"; }
&runCommand($command);

# sort alignment section
$command = "grep ^[^@] $ARGV[0] | sort -k 1,1 -s";
if (scalar(@tmp_dirs) > 0) { $" = " -T "; $command .= " -T @tmp_dirs"; }
if ($out_file ne "") { $command .= " >> $out_file"; }
&runCommand($command);

# finish
print STDERR "Conversion is completed successfully!\n";

# command, {err_msg}
sub runCommand {
    print STDERR $_[0]."\n";
    my $status = system($_[0]);
    if ($status != 0) { 
	my $errmsg;
	if (scalar(@_) > 1) { $errmsg = $_[1]; }
	else { $errmsg = "\"$command\" failed! Plase check if you provide correct parameters/options for the script!"; }
	print STDERR $errmsg."\n";
	exit(-1);
    }
    print STDERR "\n";
}

__END__

=head1 NAME

convert-sam-for-rsem

=head1 SYNOPSIS

=over

 convert-sam-for-rsem [options] input_sam

=back

=head1 ARGUMENTS

=over

=item B<input_sam>

The SAM file (*.sam) generated by user's aligner. If the aligner produces a BAM file, please use samtools to convert it to a SAM file (with header information). 

=back

=head1 OPTIONS

=over

=item B<-o> <file>

Output the converted SAM file into <file>. (Default: STDOUT)

=item B<-T/--temporary-directory> <directory>

'convert-sam-for-rsem' will call 'sort' command and this is the '-T/--temporary-directory' option of 'sort' command. The following is the description from 'sort' : "use DIR for temporaries, not $TMPDIR or /tmp; multiple options specify multiple directories". 

=item B<-h/--help>

Show help information.

=back

=head1 DESCRIPTION

This program converts the SAM file generated by user's aligner into a SAM file which RSEM can process. However, users should make sure their aligners use 'reference_name.idx.fa' generated by 'rsem-prepare-reference' as their references. In addition, their aligners should output header information and make two mates of the same alignment adjacent to each other for paired-end data. This program will output the converted file into standard output by default for the purpose of piping. By setting '-o' option, users can make the converted file written into disk.

Note: You do not need to run this script if Bowtie (not Bowtie 2) is used, or the alignment lines of a same read group together and the mates of the same alignment are adjacent each other for paired-end reads.

Note: This program can only recognize SAM files. See ARGUMENTS section.

=head1 EXAMPLES

Suppose input_sam is set to 'input.sam'. 

1) Output to standard output and gzip the output to 'input_for_rsem.sam.gz':

 convert-sam-for-rsem input.sam | gzip > input_for_rsem.sam.gz

2) Output to 'input_for_rsem.sam' directly:

 convert-sam-for-rsem input.sam -o input_for_rsem.sam

=cut