#!/usr/bin/env perl

#
# by  Xiaodong Bai
# for the extraction of sequences from GenBank based on user-defined information
#

use strict;
use warnings;
use Bio::DB::GenBank;
use Bio::SeqIO;
use vars qw(%hash);

die("Usage: $0 <info_file> <output_file>\n"), unless (@ARGV == 2);

my $infofile = shift;
open (IN,"$infofile");
while (<IN>) {
    chomp;
    next if /^\s/;
    next if /^Gene/;
    my @tmp = split /\t/;
    $hash{$tmp[0]} = 1;
}
close IN;

my $gb = new Bio::DB::GenBank;
my $outfile = shift;
my $out = Bio::SeqIO->new(-file => ">$outfile", -format => 'fasta');
foreach my $key (keys %hash) {
    my $query_str = $key . " AND \"complete cds\"";
    my $query = Bio::DB::Query::GenBank->new(-query => $query_str, -db => 'nucleotide');
    next unless (my $seqin = $gb->get_Stream_by_query($query));

    while (my $seqobj = $seqin->next_seq) {
	$out->write_seq($seqobj);
    }
}