fconcatvcf

#!/usr/bin/perl

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

=head1 NAME

fconcatvcf

=head1 SYNOPSIS

 fconcatvcf [options] <vcf_file> 

  -h             help
  -o             output file (required)
  genotype-file  gt or tg file
    
  example: fconcatvcf -o ra-317459.txt *.txt
            
  Concatenates VCF files. Assumes the columns are all ordered.
       
=head1 DESCRIPTION

=cut

#option variables
my $help;
my $verbose;
my $outFile = "out.vcf.gz";

#variables
my $check; # function pointer to isGt or isTg
my $colNo;
my %ELEMENT;
my @sortedElements;
my $headerProcessed;
my $missingData;

#initialize options
Getopt::Long::Configure ('bundling');

if(!GetOptions ('h'=>\$help, 'v'=>\$verbose, 'o=s'=>\$outFile) 
   || $help || scalar(@ARGV)==0)
{
    if ($help)
    {
        pod2usage(-verbose => 2);
    }
    else
    {
        pod2usage(1);
    }
}

my @orderedFiles;

@orderedFiles = sort {$a<=>$b} @ARGV;

#prepare output file
open(OUT, "|gzip>$outFile") || die "Cannot open $outFile\n";

my $file1Header;

#iterates through each file and concatenates it to OUT
foreach my $file (@orderedFiles) 
{   
	print "processing $file\n";
	
	if(`gzip -t $file` eq "")
	{
		open(IN, "gunzip -c $file |") || die "can't open pipe to $file";
	}
	else
	{
		open(IN, $file) || die "can't open $file";
	}
    
    $headerProcessed = 0;
    
    while (<IN>)
    {        
        if (!$headerProcessed)
        {
			if (/^#CHROM/)
			{
	            if($file eq $orderedFiles[0])
	    		{
	    			$file1Header = $_;
	    			print OUT $_;
	    		}
	    		
	            if ($file1Header ne $_)
	            {
	            	die "Header for $file not the same as the original file";
	            }
	            
	            $headerProcessed = 1;
        	}
        }
        else
        {           
            print OUT $_;
        }
    }
    
    close(IN);
}

close(OUT);