#!/usr/bin/perl
#
#########1#########2#########3#########4#########5#########6#########7########
#
#   qsub_megablast_validation.pl
#
#   Program generates a bash script to run with Sun Grid Engine.
#
#   Runs 'validate_megablast_output.pl' against all output files 
#   from mega_blast in your current directory having names similar to :
#
#     chr##.xml.#.mega_out
#
#   Use of Sun Grid Engine allows multiple jobs to be generated
#   by this program and SGE queues them and executes each as
#   machine resources are available.
#
#   Command-line Parameters :
#
#       none
#
#  Before running this program you must 'cd' into directory where
#  your mega_blast output files are located  (see README for Build 2).
#
#  syntax : [path]/qsub_megablast.pl
#
#########1#########2#########3#########4#########5#########6#########7########
#
#
use Env qw(UPROBE_BIN_DIR);
if (! (-x "$UPROBE_BIN_DIR/validate_megablast_output.pl") )
{
  my $ErrMsg = "Error : Did not find 'validate_megablast_output.pl'\n ";
  my $ErrMsg = $ErrMsg."  Did you forget to define UPROBE_BIN_DIR ? \n";
  die $ErrMsg."\n";
}
#
open (IN,"ls chr*.mega_out |") || die "Hey - You dont have any MegaBlast output files for me !";
while (<IN>) {
        chop;
        ($chromosome, $fractional) = ($_ =~ m/chr(.+).xml.(.+)/);
        $file = $_;
        ($outfile) = ($file =~ m/(.+)\.mega_out/);  
        $outfile = "$outfile.valid";
        $base_name = "validate_".$chromosome."_".$fractional;
        $stdout = "$base_name.out";
        my $jobname = "Val-$chromosome-$fractional";
        my $outname= "$base_name.bash";
        print "Creating $outname\n";
        open(OUT,"> $outname");

        print OUT "#!/bin/bash \n";
        print OUT "#\$ -N $jobname \n";
        print OUT "#\$ -S /bin/bash \n";
        print OUT "#\$ -o $stdout \n";
        print OUT "#\$ -cwd \n";
        #print OUT "#\$ -m e \n";
      # print OUT "#\$ -M yourEmailId\@your.Email.Domain \n";
        print OUT "# \n";


        print OUT "cat $file | /usr/bin/sort > $file.sorted\n";
        print OUT "$UPROBE_BIN_DIR/validate_megablast_output.pl  -f $file.sorted > $outfile\n";
#        print OUT "rm $file.sorted\n";
        
        close OUT;
        system("qsub $outname");
}


##############################################################################
# Start Description
# For each input file in current directory, sort the file then run validate_megablast_output.pl
# End Description
##############################################################################