Bio/SeqIO/raw.pm

   1 #-----------------------------------------------------------------------------
   2 # PACKAGE : Bio::SeqIO::raw
   3 # AUTHOR  : Ewan Birney <birney@ebi.ac.uk>
   4 # CREATED : Feb 16 1999
   5 # REVISION: $Id$
   6 #
   7 # Copyright (c) 1997-9 bioperl, Ewan Birney. All Rights Reserved.
   8 #           This module is free software; you can redistribute it and/or
   9 #           modify it under the same terms as Perl itself.
  10 #
  11 # _History_
  12 #
  13 # Ewan Birney <birney@ebi.ac.uk> developed the SeqIO
  14 # schema and the first prototype modules.
  15 #
  16 # This code is based on his Bio::SeqIO::Fasta module with
  17 # the necessary minor tweaks necessary to get it to read
  18 # and write raw formatted sequences made by
  19 # chris dagdigian <dag@sonsorol.org>
  20 #
  21 # October 18, 1999  Largely rewritten by Lincoln Stein
  22 #
  23 # Copyright Ewan Birney
  24 #
  25 # You may distribute this module under the same terms as perl itself
  26
  27 # POD documentation - main docs before the code
  28
  29 =head1 NAME
  30
  31 Bio::SeqIO::raw - raw sequence file input/output stream
  32
  33 =head1 SYNOPSIS
  34
  35 Do not use this module directly.  Use it via the L<Bio::SeqIO> class.
  36
  37 =head1 DESCRIPTION
  38
  39 This object can transform Bio::Seq objects to and from raw flat
  40 file databases.
  41
  42 =head1 FEEDBACK
  43
  44 =head2 Mailing Lists
  45
  46 User feedback is an integral part of the evolution of this and other
  47 Bioperl modules. Send your comments and suggestions preferably to one
  48 of the Bioperl mailing lists.  Your participation is much appreciated.
  49
  50   bioperl-l@bioperl.org                  - General discussion
  51   http://bioperl.org/wiki/Mailing_lists  - About the mailing lists
  52
  53 =head2 Support
  54
  55 Please direct usage questions or support issues to the mailing list:
  56
  57 L<bioperl-l@bioperl.org>
  58
  59 rather than to the module maintainer directly. Many experienced and
  60 reponsive experts will be able look at the problem and quickly
  61 address it. Please include a thorough description of the problem
  62 with code and data examples if at all possible.
  63
  64 =head2 Reporting Bugs
  65
  66 Report bugs to the Bioperl bug tracking system to help us keep track
  67 the bugs and their resolution.
  68 Bug reports can be submitted via the web:
  69
  70   http://bugzilla.open-bio.org/
  71
  72 =head1 AUTHORS
  73
  74   Ewan Birney   E<lt>birney@ebi.ac.ukE<gt>
  75   Lincoln Stein E<lt>lstein@cshl.orgE<gt>
  76
  77 =head1 CONTRIBUTORS
  78
  79   Jason Stajich E<lt>jason@bioperl.org<gt>
  80
  81 =head1 APPENDIX
  82
  83 The rest of the documentation details each of the object methods.
  84 Internal methods are usually preceded with a _
  85
  86 =cut
  87
  88
  89 # Let the code begin...
  90
  91 package Bio::SeqIO::raw;
  92 use strict;
  93
  94 use Bio::Seq::SeqFactory;
  95
  96 use base qw(Bio::SeqIO);
  97
  98 sub _initialize {
  99     my($self,@args) = @_;
 100     $self->SUPER::_initialize(@args);
 101     my ($variant) = $self->_rearrange([qw(VARIANT)], @args);
 102     $variant ||= 'multiple';
 103     $self->variant($variant);
 104     if( ! defined $self->sequence_factory ) {
 105         $self->sequence_factory(Bio::Seq::SeqFactory->new
 106                     (-verbose => $self->verbose(),
 107                      -type => 'Bio::Seq'));
 108     }
 109     $self->variant;
 110 }
 111
 112 =head2 next_seq
 113
 114  Title   : next_seq
 115  Usage   : $seq = $stream->next_seq()
 116  Function: returns the next sequence in the stream
 117  Returns : Bio::Seq object
 118  Args    :
 119
 120
 121 =cut
 122
 123 sub next_seq{
 124     my ($self,@args) = @_;
 125     ## When its 1 sequence per line with no formatting at all,
 126     ## grabbing it should be easy :)
 127
 128     ## adding an option to assume the file is one sequence
 129     local $/ = $self->{record_separator};
 130     my $nextline = $self->_readline();
 131     return unless defined $nextline;
 132
 133     my $sequence = uc($nextline);
 134     $sequence =~ s/\W//g;
 135     return unless $sequence;
 136     return $self->sequence_factory->create(-seq => $sequence) if $sequence;
 137 }
 138
 139 =head2 write_seq
 140
 141  Title   : write_seq
 142  Usage   : $stream->write_seq($seq)
 143  Function: writes the $seq object into the stream
 144  Returns : 1 for success and 0 for error
 145  Args    : Array of Bio::PrimarySeqI objects
 146
 147
 148 =cut
 149
 150 sub write_seq {
 151     my ($self,@seq) = @_;
 152     foreach my $seq (@seq) {
 153         $self->throw("Must provide a valid Bio::PrimarySeqI object")
 154             unless defined $seq && ref($seq) && $seq->isa('Bio::PrimarySeqI');
 155         $self->_print($seq->seq, "\n") or return;
 156     }
 157     $self->flush if $self->_flush_on_write && defined $self->_fh;
 158     return 1;
 159 }
 160
 161 =head2 write_qual
 162
 163  Title   : write_qual
 164  Usage   : $stream->write_qual($seq)
 165  Function: writes the $seq object into the stream
 166  Returns : 1 for success and 0 for error
 167  Args    : Bio::Seq::Quality object
 168
 169
 170 =cut
 171
 172 sub write_qual {
 173     my ($self,@seq) = @_;
 174     my @qual = ();
 175     foreach (@seq) {
 176         unless ($_->isa("Bio::Seq::Quality")){
 177            warn("You cannot write raw qualities without supplying a Bio::Seq::Quality object! You passed a ", ref($_), "\n");
 178            next;
 179         }
 180         @qual = @{$_->qual};
 181         if(scalar(@qual) == 0) {
 182            $qual[0] = "\n";
 183         }
 184         $self->_print (join " ", @qual,"\n") or return;
 185     }
 186     return 1;
 187 }
 188
 189 =head2 variant
 190
 191  Title   : variant
 192  Usage   : $format  = $obj->variant();
 193  Function: Get and set method for the sequence variant.  For raw sequence, this
 194            indicates whether to treat the input as multiple sequences (the
 195            default) or as a single sequence.
 196
 197            Current values accepted are:
 198
 199             'single'    single sequence
 200             'multiple'  multiple sequences (default)
 201  Returns : string
 202  Args    : new value, string
 203
 204 =cut
 205
 206 sub variant {
 207     my ($self, $var) = @_;
 208     if (defined $var || !defined $self->{variant}) {
 209         $var ||= 'multiple';
 210         $var = lc $var;
 211         $self->throw("Accepted raw format variants: 'single', 'multiple'") unless
 212             $var eq 'single' || $var eq 'multiple';
 213         $self->{record_separator} = $var eq 'single' ? undef : $/;
 214         $self->{variant} = $var;
 215     }
 216     return $self->{variant};
 217 }
 218
 219 # private method for testing record separator
 220
 221 sub _separator {
 222     shift->{record_separator};
 223 }
 224
 225 1;