fsplit

#!/usr/bin/perl

use warnings;
use strict;
use fralib;
use POSIX qw(ceil floor);
use File::Basename;
use Getopt::Long;
use Pod::Usage;

=head1 NAME

fsplit

=head1 SYNOPSIS

fsplit [options] <filename>
  
    -v  verbose
    -d  debug
    -n  number of rows in each split file (inclusive of header) >1 
    -o  out file name to append split to.
    
    usage: fsplit -n 500 pscalare.txt
    
    Splits a genotype/table file.
  
=head1 DESCRIPTION

=cut

#option variables
my $help;
my $n = 0;
my $outFile;

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

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

my $file = $ARGV[0];

if (!defined($outFile))
{
    my ($name, $dir, $ext) = fileparse($file, '\..*');
	$outFile = "$name$ext";
}

open(IN, $file) || die "Cannot open $file\n";

my $header = "";
my $header_line_count = 0;

while (<IN>)
{	
	
	if (/^#/)
	{
	    $header .= $_;
    	++$header_line_count;
	}
	else
	{
    	if (($.-$header_line_count)%$n==1)
    	{
        	my $tempOutFile = "$outFile" . "_" . ceil(($.-1)/$n) . ".sites.vcf";
    		#open new file		
    		open(OUT, ">$tempOutFile") || die "Cannot open $tempOutFile\n";
    		print OUT $header;
    	}
    	
    	print OUT $_;
    }
}

#checks if file is splitted properly
my $splitFileNo = ceil(($.-$header_line_count)/$n);

close(OUT);
close(IN);

#checks if file is splitted properly
my $totalSize = 0;
for my $i (1..$splitFileNo)
{
	$totalSize += -s "$outFile" . "_$i" . ".sites.vcf";
}

if (-s $file != $totalSize-(($splitFileNo-1)*length($header)))
{
	warn "Splitting of $file appears to be corrupted";
}