#!/usr/bin/perl -w
#
#
#
use strict;
no strict "subs";
use XML::Smart;
use IO::File;
use Data::Dumper;
use Digest::MD5 qw(md5_hex);

# simple postprocess NCBI BLAST -m 7
# for diff based signature comparisons
# 
# Copyright (c) 2005 Scalable Informatics LLC
# Dual licensed:
#  For GPL and open source projects, this code carries a GPL v2.0 License
#  For commercially redistributed and closed source projects, you will need 
#  to obtain an appropriate closed source license from Scalable Informatics LLC
#
# Support for this code is available for a fee from Scalable Informatics LLC
# http://www.scalableinformatics.com
# landman@scalableinformatics.com
# 
my ($tag,$d,$in,$xml,@hits,@lines,$file,$all,@sub_documents);
my (@results,$query,$x1);

$in  = shift;
$file = IO::File->new;
if ($file->open("< ".$in))
   {
     @lines = $file->getlines;
     $all=join("",@lines);
   }
@sub_documents  = split(/\<\?xml version=\"1.0\"\?>/,$all);
shift @sub_documents;

foreach (@sub_documents)   
  {
    undef $xml;
    $xml  = XML::Smart->new($_);
    $query= md5_hex($xml->{BlastOutput}{'BlastOutput_query-def'});
    @hits=$xml->{'BlastOutput'}{'BlastOutput_iterations'}{'Iteration'}{'Iteration_hits'}{'Hit'}('@');
    foreach $x1 ( sort { $a->{'Hit_id'} <=> $b->{'Hit_id'}  } @hits )
      {
        push @results,sprintf("%s:%s:%s",$query,md5_hex($x1->{'Hit_id'}),$x1->{'Hit_hsps'}{'Hsp'}{'Hsp_evalue'});
      }
  }

foreach (@results)
  {
    printf "%s\n",$_;
  }