File Coverage

blib/lib/BioX/SeqHash.pm

Criterion	Covered	Total	%
statement	9	38	23.6
branch	0	16	0.0
condition			n/a
subroutine	3	10	30.0
pod	4	6	66.6
total	16	70	22.8

package BioX::SeqHash;
use strict;
use warnings;
use Exporter 'import';
our @EXPORT_OK = qw(fa2hs);

our $VERSION = '0.1.1'; # VERSION:
# ABSTRACT: get one or more sequences from a FASTA file quickly.


sub new {
  my $class = shift;
  my $self = {
    @_
  };
  $self->{file} || die `pod2text $0`;
  bless $self, $class;
}


sub fa2hs {
  my $fa = ref $_[0] ? $_[0]->{file} : $_[0];
  open my $file, "<", "$fa" or  die "Can not open $fa $!";
  my (%hs, $name);
  while (my $line = <$file>) {
    chomp($line);
    if ($line =~/^>(.+?)\s/) {
      $name = $1;
    } else {
      $hs{$name} .= $line;
    }
  }
  close($file);
  return \%hs;
}

sub _initialize {
  my $self = shift;
  $self->{id2seq} = $self->fa2hs();
}


sub get_id_seq {
  my ($self, $id) = @_;
  $self->_initialize() unless (exists $self->{id2seq});
  return ">$id\n" . $self->{id2seq}->{$id} . "\n";
}


sub get_seq {
  my ($self, $id) = @_;
  $self->_initialize() unless (exists $self->{id2seq});
  return $self->{id2seq}{$id};
}


sub get_seqs_batch {
  my ($self, $id_file, $outfile) = @_;
  open my $IN, "<",  $id_file or die "Can not open $id_file $!";
  my @ids = <$IN>;
  chomp @ids;
  open my $OUT, ">", $outfile or die "Can not open $outfile $!";
  for my $id (@ids) {
    print $OUT $self->get_id_seq($id);
  }
}

# may be use later
sub rename_seqs {

}

1;

__END__