CoordSystem.pm 9.75 KB
Newer Older
1
2
=head1 LICENSE

Andy Yates's avatar
Andy Yates committed
3
  Copyright (c) 1999-2012 The European Bioinformatics Institute and
4
5
6
7
8
9
10
11
12
13
  Genome Research Limited.  All rights reserved.

  This software is distributed under a modified Apache license.
  For license details, please see

    http://www.ensembl.org/info/about/code_licence.html

=head1 CONTACT

  Please email comments or questions to the public Ensembl
14
  developers list at <dev@ensembl.org>.
15
16
17
18
19

  Questions may also be sent to the Ensembl help desk at
  <helpdesk@ensembl.org>.

=cut
20
21
22
23
24
25
26
27
28
29
30
31
32
33

=head1 NAME

Bio::EnsEMBL::CoordSystem

=head1 SYNOPSIS

  my $db = Bio::EnsEMBL::DBSQL::DBAdaptor->new(...);

  my $csa = $db->get_CoordSystemAdaptor();

  #
  # Get all coord systems in the database:
  #
34
35
  foreach my $cs ( @{ $csa->fetch_all() } ) {
    my $str = join ':', $cs->name(), $cs->version(), $cs->dbID();
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
    print "$str\n";
  }

=head1 DESCRIPTION

This is a simple object which contains a few coordinate system attributes:
name, internal identifier, version.  A coordinate system is uniquely defined
by its name and version.  A version of a coordinate system applies to all
sequences within a coordinate system.  This should not be confused with
individual sequence versions.

Take for example the Human assembly.  The version 'NCBI33' applies to
to all chromosomes in the NCBI33 assembly (that is the entire 'chromosome'
coordinate system).  The 'clone' coordinate system in the same database would
have no version however.  Although the clone sequences have their own sequence
versions, there is no version which applies to the entire set of clones.

Coordinate system objects are immutable. Their name and version, and other
attributes may not be altered after they are created.

=head1 METHODS

=cut


use strict;
use warnings;

package Bio::EnsEMBL::CoordSystem;

use Bio::EnsEMBL::Storable;

68
69
use Bio::EnsEMBL::Utils::Argument  qw(rearrange);
use Bio::EnsEMBL::Utils::Exception qw(throw);
70
71
72
73
74
75
76
77
78
79

use vars qw(@ISA);

@ISA = qw(Bio::EnsEMBL::Storable);


=head2 new

  Arg [..]   : List of named arguments:
               -NAME      - The name of the coordinate system
80
81
82
83
               -VERSION   - (optional) The version of the coordinate system.
                            Note that if the version passed in is undefined,
                            it will be set to the empty string in the
                            resulting CoordSystem object.
84
85
86
87
88
89
               -RANK      - The rank of the coordinate system. The highest
                            level coordinate system should have rank 1, the
                            second highest rank 2 and so on.  An example of
                            a high level coordinate system is 'chromosome' an
                            example of a lower level coordinate system is
                            'clone'.
90
               -TOP_LEVEL - (optional) Sets whether this is a top-level coord
91
92
93
                            system. Default = 0. This should only be set to
                            true if you are creating an artificial toplevel
                            coordsystem by the name of 'toplevel'
94
95
               -SEQUENCE_LEVEL - (optional) Sets whether this is a sequence
                            level coordinate system. Default = 0
96
               -DEFAULT   - (optional)
97
98
99
100
101
102
                            Whether this is the default version of the 
                            coordinate systems of this name. Default = 0
               -DBID      - (optional) The internal identifier of this
                             coordinate system
               -ADAPTOR   - (optional) The adaptor which provides database
                            interaction for this object
103
  Example    : $cs = Bio::EnsEMBL::CoordSystem->new(-NAME    => 'chromosome',
104
                                                    -VERSION => 'NCBI33',
105
                                                    -RANK    => 1,
106
107
108
                                                    -DBID    => 1,
                                                    -ADAPTOR => adaptor,
                                                    -DEFAULT => 1,
109
                                                    -SEQUENCE_LEVEL => 0);
110
111
112
113
114
  Description: Creates a new CoordSystem object representing a coordinate
               system.
  Returntype : Bio::EnsEMBL::CoordSystem
  Exceptions : none
  Caller     : general
115
  Status     : Stable
116
117
118
119
120
121
122
123
124

=cut

sub new {
  my $caller = shift;
  my $class = ref($caller) || $caller;

  my $self = $class->SUPER::new(@_);

125
126
127
128
129
  my ( $name, $version, $top_level, $sequence_level, $default, $rank ) =
    rearrange( [ 'NAME',      'VERSION',
                 'TOP_LEVEL', 'SEQUENCE_LEVEL',
                 'DEFAULT',   'RANK' ],
               @_ );
130

131
  $version = '' if ( !defined($version) );
132

133
134
135
  $top_level      = ($top_level)      ? 1 : 0;
  $sequence_level = ($sequence_level) ? 1 : 0;
  $default        = ($default)        ? 1 : 0;
136
137
  $rank ||= 0;

138
  if ( $top_level == 1 ) {
139
    if ( $rank != 0 ) {
140
141
142
      throw('RANK argument must be 0 if TOP_LEVEL is 1');
    }

143
144
145
    if ( defined($name) ) {
      if ( $name ne 'toplevel' ) {
        throw('The NAME argument must be "toplevel" if TOP_LEVEL is 1');
146
147
148
149
150
      }
    } else {
      $name = 'toplevel';
    }

151
    if ( $sequence_level == 1 ) {
152
153
154
155
156
157
      throw("SEQUENCE_LEVEL argument must be 0 if TOP_LEVEL is 1");
    }

    $default = 0;

  } else {
158
159
160

    if ( $rank == 0 ) {
      throw("RANK argument must be non-zero unless TOP_LEVEL is 1");
161
    }
162
163
164
165
166
167

    if ( !defined($name) ) {
      throw('The NAME argument is required');
    } elsif ( $name eq 'toplevel' ) {
      throw(   "Cannot name coord system 'toplevel' "
             . "unless TOP_LEVEL is 1" );
168
    }
169

170
171
  }

172
  if ( $rank !~ /^\d+$/ ) {
173
174
    throw('The RANK argument must be a positive integer');
  }
175

176
177
178
  $self->{'version'}        = $version;
  $self->{'name'}           = $name;
  $self->{'top_level'}      = $top_level;
179
  $self->{'sequence_level'} = $sequence_level;
180
181
  $self->{'default'}        = $default;
  $self->{'rank'}           = $rank;
182
183

  return $self;
184
} ## end sub new
185
186
187
188
189
190
191
192
193
194


=head2 name

  Arg [1]    : (optional) string $name
  Example    : print $coord_system->name();
  Description: Getter for the name of this coordinate system
  Returntype : string
  Exceptions : none
  Caller     : general
195
  Status     : Stable
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215

=cut

sub name {
  my $self = shift;
  return $self->{'name'};
}



=head2 version

  Arg [1]    : none
  Example    : print $coord->version();
  Description: Getter for the version of this coordinate system.  This
               will return an empty string if no version is defined for this
               coordinate system.
  Returntype : string
  Exceptions : none
  Caller     : general
216
  Status     : Stable
217
218
219
220
221
222
223
224
225
226

=cut

sub version {
  my $self = shift;
  return $self->{'version'};
}



Andrew Jenkinson's avatar
Andrew Jenkinson committed
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
=head2 species

  Arg [1]    : none
  Example    : print $coord->species();
  Description: Shortcut method to get the species this CoordSystem refers to.
  Returntype : string
  Exceptions : none
  Caller     : general
  Status     : Stable

=cut

sub species {
  my $self = shift;
  return $self->adaptor->db->species;
}


245
246
247
248
249
250
251
252
253
254
255
256

=head2 equals

  Arg [1]    : Bio::EnsEMBL::CoordSystem $cs
               The coord system to compare to for equality.
  Example    : if($coord_sys->equals($other_coord_sys)) { ... }
  Description: Compares 2 coordinate systems and returns true if they are
               equivalent.  The definition of equivalent is sharing the same
               name and version.
  Returntype : string
  Exceptions : none
  Caller     : general
257
  Status     : Stable
258
259
260
261
262
263
264
265

=cut

sub equals {
  my $self = shift;
  my $cs = shift;

  if(!$cs || !ref($cs) || !$cs->isa('Bio::EnsEMBL::CoordSystem')) {
Andrew Jenkinson's avatar
Andrew Jenkinson committed
266
267
268
269
    if ($cs->isa('Bio::EnsEMBL::ExternalData::DAS::CoordSystem')) {
      return $cs->equals($self);
    }
    throw('Argument must be a CoordSystem');
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
  }

  if($self->{'version'} eq $cs->version() && $self->{'name'} eq $cs->name()) {
    return 1;
  }

  return 0;
}




=head2 is_top_level

  Arg [1]    : none
  Example    : if($coord_sys->is_top_level()) { ... }
286
287
288
289
290
  Description: Returns true if this is the toplevel pseudo coordinate system.
               The toplevel coordinate system is not a real coordinate system
               which is stored in the database, but it is a placeholder that
               can be used to request transformations or retrievals to/from
               the highest defined coordinate system in a given region.
Kieron Taylor's avatar
Kieron Taylor committed
291
  Returntype : 1 or 0
292
293
  Exceptions : none
  Caller     : general
294
  Status     : Stable
295
296
297
298
299
300
301
302
303
304
305
306
307
308

=cut

sub is_top_level {
  my $self = shift;
  return $self->{'top_level'};
}


=head2 is_sequence_level

  Arg [1]    : none
  Example    : if($coord_sys->is_sequence_level()) { ... }
  Description: Returns true if this is a sequence level coordinate system
Kieron Taylor's avatar
Kieron Taylor committed
309
  Returntype : 1 or 0
310
311
  Exceptions : none
  Caller     : general
312
  Status     : Stable
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327

=cut

sub is_sequence_level {
  my $self = shift;
  return $self->{'sequence_level'};
}


=head2 is_default

  Arg [1]    : none
  Example    : if($coord_sys->is_default()) { ... }
  Description: Returns true if this coordinate system is the default
               version of the coordinate system of this name.
Kieron Taylor's avatar
Kieron Taylor committed
328
  Returntype : 1 or 0
329
330
  Exceptions : none
  Caller     : general
331
  Status     : Stable
332
333
334
335
336
337
338
339
340

=cut

sub is_default {
  my $self = shift;
  return $self->{'default'};
}


341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356


=head2 rank

  Arg [1]    : none
  Example    : if($cs1->rank() < $cs2->rank()) {
                 print $cs1->name(), " is a higher level coord system than",
                       $cs2->name(), "\n";
               }
  Description: Returns the rank of this coordinate system.  A lower number
               is a higher coordinate system.  The highest level coordinate
               system has a rank of 1 (e.g. 'chromosome').  The toplevel
               pseudo coordinate system has a rank of 0.
  Returntype : int
  Exceptions : none
  Caller     : general
357
  Status     : Stable
358
359
360
361
362
363
364
365

=cut

sub rank {
  my $self = shift;
  return $self->{'rank'};
}

366
1;