Bio/PopGen/IO.pm

   1 # $Id$
   2 #
   3 # BioPerl module for Bio::PopGen::IO
   4 #
   5 # Please direct questions and support issues to <bioperl-l@bioperl.org>
   6 #
   7 # Cared for by Jason Stajich <jason-at-bioperl.org>
   8 #
   9 # Copyright Jason Stajich
  10 #
  11 # You may distribute this module under the same terms as perl itself
  12
  13 # POD documentation - main docs before the code
  14
  15 =head1 NAME
  16
  17 Bio::PopGen::IO - Input individual,marker,allele information
  18
  19 =head1 SYNOPSIS
  20
  21   use Bio::PopGen::IO;
  22   my $io = Bio::PopGen::IO->new(-format => 'csv',
  23                                -file   => 'data.csv');
  24
  25   # Some IO might support reading in a population at a time
  26
  27   my @population;
  28   while( my $ind = $io->next_individual ) {
  29       push @population, $ind;
  30   }
  31
  32
  33 =head1 DESCRIPTION
  34
  35 This is a generic interface to reading in population genetic data (of
  36 which there really isn't too many standard formats).  This implementation
  37 makes it easy to provide your own parser for the data.  You need to
  38 only implement one function next_individual.  You can also implement
  39 next_population if your data has explicit information about population
  40 memberhsip for the indidviduals.
  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
  48 the Bioperl mailing list.  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 of the bugs and their resolution. Bug reports can be submitted via
  68 the web:
  69
  70   http://bugzilla.open-bio.org/
  71
  72 =head1 AUTHOR - Jason Stajich
  73
  74 Email jason-at-bioperl.org
  75
  76 =head1 APPENDIX
  77
  78 The rest of the documentation details each of the object methods.
  79 Internal methods are usually preceded with a _
  80
  81 =cut
  82
  83
  84 # Let the code begin...
  85 #TODO
  86 # Set the Individual creation as a factory rather than
  87 # hardcoded
  88
  89 package Bio::PopGen::IO;
  90 use strict;
  91
  92 # Object preamble - inherits from Bio::Root::Root
  93
  94 use Bio::Root::Root;
  95
  96 use base qw(Bio::Root::IO);
  97
  98 =head2 new
  99
 100  Title   : new
 101  Usage   : my $obj = Bio::PopGen::IO->new();
 102  Function: Builds a new Bio::PopGen::IO object
 103  Returns : an instance of Bio::PopGen::IO
 104  Args    :
 105
 106
 107 =cut
 108
 109 sub new {
 110   my($class,@args) = @_;
 111
 112   if( $class =~ /Bio::PopGen::IO::(\S+)/ ) {
 113     my ($self) = $class->SUPER::new(@args);
 114     $self->_initialize(@args);
 115     return $self;
 116   } else {
 117     my %param = @args;
 118     @param{ map { lc $_ } keys %param } = values %param; # lowercase keys
 119     my $format = $param{'-format'} ||
 120       $class->_guess_format( $param{'-file'} || $ARGV[0] ) || 'csv';
 121
 122     # normalize capitalization to lower case
 123     $format = "\L$format";
 124
 125     return unless( $class->_load_format_module($format) );
 126     return "Bio::PopGen::IO::${format}"->new(@args);
 127   }
 128 }
 129
 130 # _initialize is chained for all PopGen::IO classes
 131
 132 sub _initialize {
 133     my($self, @args) = @_;
 134 #    my ($indfact, $popfact) = $self->_rearrange([qw(INDIVIDUAL_FACTORY
 135 #                                                   POPULATION_FACTORY)],
 136 #                                               @args);
 137 #    $indfact = Bio::PopGen::IndividualBuilder->new() unless $indfact;
 138 #    $indfact = Bio::PopGen::PopulationBuilder->new() unless $indfact;
 139
 140     # initialize the IO part
 141     $self->_initialize_io(@args);
 142     return 1;
 143 }
 144
 145 =head2 next_individual
 146
 147  Title   : next_individual
 148  Usage   : my $ind = $popgenio->next_individual;
 149  Function: Retrieve the next individual from a dataset
 150  Returns : L<Bio::PopGen::IndividualI> object
 151  Args    : none
 152
 153
 154 =cut
 155
 156 sub next_individual{
 157     my ($self) = @_;
 158     $self->throw_not_implemented();
 159 }
 160
 161
 162 =head2 next_population
 163
 164  Title   : next_population
 165  Usage   : my $pop = $popgenio->next_population;
 166  Function: Retrieve the next population from a dataset
 167  Returns : L<Bio::PopGen::PopulationI> object
 168  Args    : none
 169  Note    : Many implementation will not implement this
 170
 171 =cut
 172
 173 sub next_population{
 174     my ($self) = @_;
 175     $self->throw_not_implemented();
 176 }
 177
 178 =head2 write_individual
 179
 180  Title   : write_individual
 181  Usage   : $popgenio->write_individual($ind);
 182  Function: Write an individual out in the implementation format
 183  Returns : none
 184  Args    : L<Bio::PopGen::PopulationI> object(s)
 185
 186 =cut
 187
 188 sub write_individual{
 189     my ($self) = @_;
 190     $self->throw_not_implemented();
 191 }
 192
 193
 194
 195 =head2 write_population
 196
 197  Title   : write_population
 198  Usage   : $popgenio->write_population($pop);
 199  Function: Write a population out in the implementation format
 200  Returns : none
 201  Args    : L<Bio::PopGen::PopulationI> object(s)
 202  Note    : Many implementation will not implement this
 203
 204 =cut
 205
 206 sub write_population{
 207     my ($self) = @_;
 208     $self->throw_not_implemented();
 209 }
 210
 211
 212 =head2 newFh
 213
 214  Title   : newFh
 215  Usage   : $fh = Bio::SeqIO->newFh(-file=>$filename,-format=>'Format')
 216  Function: does a new() followed by an fh()
 217  Example : $fh = Bio::SeqIO->newFh(-file=>$filename,-format=>'Format')
 218            $sequence = <$fh>;   # read a sequence object
 219            print $fh $sequence; # write a sequence object
 220  Returns : filehandle tied to the Bio::SeqIO::Fh class
 221  Args    :
 222
 223 See L<Bio::SeqIO::Fh>
 224
 225 =cut
 226
 227 sub newFh {
 228   my $class = shift;
 229   return unless my $self = $class->new(@_);
 230   return $self->fh;
 231 }
 232
 233 =head2 fh
 234
 235  Title   : fh
 236  Usage   : $obj->fh
 237  Function:
 238  Example : $fh = $obj->fh;      # make a tied filehandle
 239            $sequence = <$fh>;   # read a sequence object
 240            print $fh $sequence; # write a sequence object
 241  Returns : filehandle tied to Bio::SeqIO class
 242  Args    : none
 243
 244 =cut
 245
 246
 247 sub fh {
 248   my $self = shift;
 249   my $class = ref($self) || $self;
 250   my $s = Symbol::gensym;
 251   tie $$s,$class,$self;
 252   return $s;
 253 }
 254
 255 =head2 _load_format_module
 256
 257  Title   : _load_format_module
 258  Usage   : *INTERNAL Bio::PopGen::IO stuff*
 259  Function: Loads up (like use) a module at run time on demand
 260  Example :
 261  Returns :
 262  Args    :
 263
 264 =cut
 265
 266 sub _load_format_module {
 267   my ($self,$format) = @_;
 268   my $module = "Bio::PopGen::IO::" . $format;
 269   my $ok;
 270
 271   eval {
 272       $ok = $self->_load_module($module);
 273   };
 274   if ( $@ ) {
 275       print STDERR <<END;
 276 $self: $format cannot be found
 277 Exception $@
 278 For more information about the Bio::PopGen::IO system please see the
 279 Bio::PopGen::IO docs.  This includes ways of checking for formats at
 280 compile time, not run time
 281 END
 282   ;
 283   }
 284   return $ok;
 285 }
 286
 287
 288 =head2 _guess_format
 289
 290  Title   : _guess_format
 291  Usage   : $obj->_guess_format($filename)
 292  Function:
 293  Example :
 294  Returns : guessed format of filename (lower case)
 295  Args    :
 296
 297 =cut
 298
 299
 300 sub _guess_format {
 301    my $class = shift;
 302    return unless $_ = shift;
 303    return 'csv'   if (/csv/i or /\.dat\w$/i);
 304 }
 305
 306 sub close {
 307     my $self = shift;
 308     $self->SUPER::close(@_);
 309 }
 310
 311 sub DESTROY {
 312     my $self = shift;
 313     $self->close();
 314 }
 315
 316 sub TIEHANDLE {
 317   my $class = shift;
 318   return bless {processor => shift}, $class;
 319 }
 320
 321 sub READLINE {
 322   my $self = shift;
 323   return $self->{'processor'}->next_result() unless wantarray;
 324   my (@list, $obj);
 325   push @list, $obj while $obj = $self->{'processor'}->next_result();
 326   return @list;
 327 }
 328
 329 sub PRINT {
 330   my $self = shift;
 331   $self->{'processor'}->write_result(@_);
 332 }
 333
 334 1;