#!/usr/bin/perl -w
use strict;

use Bio::SearchIO;

my $dir = $ARGV[0] || '.';
opendir(DIR, $dir) || die "cannot open dir '$dir': $!";
my %result; # where we put the results of BLAST parsing
my %gene2sp;
for my $file ( readdir(DIR) ) {
    if( $file =~ /(\S+)\.(BLASTP)$/ ) {
	my ($stem,$ext) = ($1,$2);
	warn "stem is $stem for $file\n";
	my ($query, $database);
	if( $stem =~ /(\S+)\.pep-vs-(\S+)\.pep/ ) {
	    ($query,$database) = ($1,$2);
	} else {
	    warn("Filename $file does not have a stem that I know how to match\n");
	    next;    # go to the next $file in the loop of the readdir
	}
	warn("query= $query and database= $database\n");
	my $input = Bio::SearchIO->new(-format => 'blast',
				       -file   => "$dir/$file");
	while( my $r = $input->next_result ) {
	    my $qname = $r->query_name;
	    my $qlen  = $r->query_length;
	    $gene2sp{$qname} = $query;
	    while( my $h = $r->next_hit ) {
		my $hname = $h->name;
		my $hlen  = $h->length;
		next if $qname eq $hname;
		my $evalue = $h->significance;
		#while( my $hsp = $h->next_hsp ) {
		#} 
		$result{$qname}->{$hname} = [$evalue];
	    } 
	} 
    }
}
my @clusters = &make_linkages(\%result);

my $i = 0;
for my $cluster (@clusters ) {
    for my $gene ( @$cluster ) {
	print join("\t", $i, $gene2sp{$gene}, $gene), "\n";
    }
    $i++;
}
sub make_linkages {
    my ($data) = @_;
    my @clusters;
    my %seen;
    for my $g1 ( keys %$data ) {
	for my $g2 ( keys %{$data->{$g1}} ) {
	    if( $data->{$g2}->{$g1} ) {	# reciprocal, yah!
		my $cluster_id1 = $seen{$g1};
		my $cluster_id2 = $seen{$g2};
		if( defined $cluster_id1 && defined $cluster_id2 ) {
		    if( $cluster_id1 == $cluster_id2 ) {
			#we've already seen and classified these genes into the same cluster
			next;
		    } else { # we need to merge two separate clusters
			# we've already seen and classified these genes, 
                        # but they are in separate clusters
			#we need to join these two clusters together
			# so we need move all the genes into a single cluster
			for my $gene ( @{$clusters[$cluster_id2]} ) {
			    push @{$clusters[$cluster_id1]}, $gene;
			    # change the stored cluster ID for the gene
			    $seen{$gene} = $cluster_id1;
			}
			undef $clusters[$cluster_id2];
		    }
		} elsif( defined $cluster_id1 ) { # gene1 is already assigned to cluster
		    push @{$clusters[$cluster_id1]}, $g2;
		    $seen{$g2} = $cluster_id1;
		} elsif( defined $cluster_id2 ) { # gene2 is already assigned to cluster
		    push @{$clusters[$cluster_id2]}, $g1;
		    $seen{$g1} = $cluster_id2;
		} else { # neither gene has been assigned to a cluster
		    # need to make a new cluster
		    # add it to the end of the list of clusters
		    $cluster_id1 =  scalar @clusters;
                    $cluster_id2 = scalar @clusters;
		    push @clusters, [$g1,$g2];
		    $seen{$g1} = $cluster_id1;
		    $seen{$g2} = $cluster_id2;
		}
	    } else {
		# search is not reciprocal 
	    }
	}
    }
    return grep { defined } @clusters;
}

sub unique {
    my %unique;
    return grep { ! $unique{$_}++ } @_;
}