use strict;
use Data::Dumper;

my $usage = "usage: gather_pg_sets RepeatSets PGsets Precomputed1 Precomputed2...";
my($repeat_clustersF,$pg_setsF);
(
 ($repeat_clustersF = shift @ARGV) &&
 ($pg_setsF         = shift @ARGV)
)
    || die $usage;

my @dirs = @ARGV; # Something like ('PubSEED.Output.1','PubSEED.Output.2','BrucellaPGs');
(@dirs > 0) || die "You need to give the precomputed correspondence sets (see build_PGs_for_OTUs)";


open(RP,"| cluster_objects | tabs2rel > $repeat_clustersF") || die "FAILED";
open(PG,"| cluster_objects | tabs2rel > $pg_setsF")         || die "FAILED";

my %rep_peg;
foreach my $d (@dirs)
{
    opendir(SUB,$d) || die $d;
    my @sub = grep { $_ =~/^\d/ } readdir(SUB);
    closedir(SUB);
    foreach my $subD (@sub)
    {
	if (-s "$d/$subD/PG/repeat.clusters")
	{
	    open(CON,"rel2tabs < $d/$subD/PG/repeat.clusters |") || die "$d/$subD";
	    while (defined($_ = <CON>))
	    {
		chop;
		my @pegs = split(/\t/,$_);
		foreach $_ (@pegs) { $rep_peg{$_} = 1 }   #### mark pegs in repeats
		for (my $i=1; ($i < @pegs); $i++)
		{
		    print RP join("\t",($pegs[0],$pegs[$i])),"\n";
		}
	    }
	    close(CON);
	}
    }
}

foreach my $d (@dirs)
{
    opendir(SUB,$d) || die $d;
    my @sub = grep { $_ =~/^\d/ } readdir(SUB);
    closedir(SUB);
    foreach my $subD (@sub)
    {
	if (-s "$d/$subD/PG/pg.sets")
	{
	    open(CON,"rel2tabs < $d/$subD/PG/pg.sets |") || die "$d/$subD";
	    while (defined($_ = <CON>))
	    {
		chop;
		my @pegs = grep { ! $rep_peg{$_} } split(/\t/,$_);
		for (my $i=1; ($i < @pegs); $i++)
		{
		    print PG join("\t",($pegs[0],$pegs[$i])),"\n";
		}
	    }
	    close(CON);
	}
    }
}