1 package Text::Tradition::Collation;
4 use Encode qw( decode_utf8 );
8 use IPC::Run qw( run binary );
10 use Text::Tradition::Collation::Data;
11 use Text::Tradition::Collation::Reading;
12 use Text::Tradition::Collation::RelationshipStore;
13 use Text::Tradition::Error;
14 use XML::Easy::Syntax qw( $xml10_namestartchar_rx $xml10_namechar_rx );
16 use XML::LibXML::XPathContext;
20 isa => 'Text::Tradition::Collation::Data',
58 isa => 'Text::Tradition',
59 writer => '_set_tradition',
65 Text::Tradition::Collation - a software model for a text collation
70 my $t = Text::Tradition->new(
71 'name' => 'this is a text',
73 'file' => '/path/to/tei_parallel_seg_file.xml' );
75 my $c = $t->collation;
76 my @readings = $c->readings;
77 my @paths = $c->paths;
78 my @relationships = $c->relationships;
80 my $svg_variant_graph = $t->collation->as_svg();
84 Text::Tradition is a library for representation and analysis of collated
85 texts, particularly medieval ones. The Collation is the central feature of
86 a Tradition, where the text, its sequence of readings, and its relationships
87 between readings are actually kept.
93 The constructor. Takes a hash or hashref of the following arguments:
97 =item * tradition - The Text::Tradition object to which the collation
100 =item * linear - Whether the collation should be linear; that is, whether
101 transposed readings should be treated as two linked readings rather than one,
102 and therefore whether the collation graph is acyclic. Defaults to true.
104 =item * baselabel - The default label for the path taken by a base text
105 (if any). Defaults to 'base text'.
107 =item * wit_list_separator - The string to join a list of witnesses for
108 purposes of making labels in display graphs. Defaults to ', '.
110 =item * ac_label - The extra label to tack onto a witness sigil when
111 representing another layer of path for the given witness - that is, when
112 a text has more than one possible reading due to scribal corrections or
113 the like. Defaults to ' (a.c.)'.
115 =item * wordsep - The string used to separate words in the original text.
126 =head2 wit_list_separator
134 Simple accessors for collation attributes.
138 The meta-reading at the start of every witness path.
142 The meta-reading at the end of every witness path.
146 Returns all Reading objects in the graph.
148 =head2 reading( $id )
150 Returns the Reading object corresponding to the given ID.
152 =head2 add_reading( $reading_args )
154 Adds a new reading object to the collation.
155 See L<Text::Tradition::Collation::Reading> for the available arguments.
157 =head2 del_reading( $object_or_id )
159 Removes the given reading from the collation, implicitly removing its
160 paths and relationships.
162 =head2 has_reading( $id )
164 Predicate to see whether a given reading ID is in the graph.
166 =head2 reading_witnesses( $object_or_id )
168 Returns a list of sigils whose witnesses contain the reading.
172 Returns all reading paths within the document - that is, all edges in the
173 collation graph. Each path is an arrayref of [ $source, $target ] reading IDs.
175 =head2 add_path( $source, $target, $sigil )
177 Links the given readings in the collation in sequence, under the given witness
178 sigil. The readings may be specified by object or ID.
180 =head2 del_path( $source, $target, $sigil )
182 Links the given readings in the collation in sequence, under the given witness
183 sigil. The readings may be specified by object or ID.
185 =head2 has_path( $source, $target );
187 Returns true if the two readings are linked in sequence in any witness.
188 The readings may be specified by object or ID.
192 Returns all Relationship objects in the collation.
194 =head2 add_relationship( $reading, $other_reading, $options )
196 Adds a new relationship of the type given in $options between the two readings,
197 which may be specified by object or ID. Returns a value of ( $status, @vectors)
198 where $status is true on success, and @vectors is a list of relationship edges
199 that were ultimately added.
200 See L<Text::Tradition::Collation::Relationship> for the available options.
205 my ( $class, @args ) = @_;
206 my %args = @args == 1 ? %{ $args[0] } : @args;
207 # TODO determine these from the Moose::Meta object
208 my @delegate_attrs = qw(sequence relations readings wit_list_separator baselabel
209 linear wordsep start end cached_table _graphcalc_done);
211 for my $attr (@delegate_attrs) {
212 $data_args{$attr} = delete $args{$attr} if exists $args{$attr};
214 $args{_data} = Text::Tradition::Collation::Data->new(%data_args);
220 $self->_set_relations( Text::Tradition::Collation::RelationshipStore->new( 'collation' => $self ) );
221 $self->_set_start( $self->add_reading(
222 { 'collation' => $self, 'is_start' => 1, 'init' => 1 } ) );
223 $self->_set_end( $self->add_reading(
224 { 'collation' => $self, 'is_end' => 1, 'init' => 1 } ) );
227 sub register_relationship_type {
229 my %args = @_ == 1 ? %{$_[0]} : @_;
230 if( $self->relations->has_type( $args{name} ) ) {
231 throw( 'Relationship type ' . $args{name} . ' already registered' );
233 $self->relations->add_type( %args );
236 ### Reading construct/destruct functions
239 my( $self, $reading ) = @_;
240 unless( ref( $reading ) eq 'Text::Tradition::Collation::Reading' ) {
241 my %args = %$reading;
242 if( $args{'init'} ) {
243 # If we are initializing an empty collation, don't assume that we
244 # have set a tradition.
245 delete $args{'init'};
246 } elsif( $self->tradition->can('language') && $self->tradition->has_language
247 && !exists $args{'language'} ) {
248 $args{'language'} = $self->tradition->language;
250 $reading = Text::Tradition::Collation::Reading->new(
251 'collation' => $self,
254 # First check to see if a reading with this ID exists.
255 if( $self->reading( $reading->id ) ) {
256 throw( "Collation already has a reading with id " . $reading->id );
258 $self->_graphcalc_done(0);
259 $self->_add_reading( $reading->id => $reading );
260 # Once the reading has been added, put it in both graphs.
261 $self->sequence->add_vertex( $reading->id );
262 $self->relations->add_reading( $reading->id );
266 around del_reading => sub {
271 if( ref( $arg ) eq 'Text::Tradition::Collation::Reading' ) {
274 # Remove the reading from the graphs.
275 $self->_graphcalc_done(0);
276 $self->_clear_cache; # Explicitly clear caches to GC the reading
277 $self->sequence->delete_vertex( $arg );
278 $self->relations->delete_reading( $arg );
281 $self->$orig( $arg );
284 =head2 merge_readings( $main, $second, $concatenate, $with_str )
286 Merges the $second reading into the $main one. If $concatenate is true, then
287 the merged node will carry the text of both readings, concatenated with either
288 $with_str (if specified) or a sensible default (the empty string if the
289 appropriate 'join_*' flag is set on either reading, or else $self->wordsep.)
291 The first two arguments may be either readings or reading IDs.
297 my $cxfile = 't/data/Collatex-16.xml';
298 my $t = Text::Tradition->new(
300 'input' => 'CollateX',
303 my $c = $t->collation;
305 my $rno = scalar $c->readings;
306 # Split n21 ('unto') for testing purposes
307 my $new_r = $c->add_reading( { 'id' => 'n21p0', 'text' => 'un', 'join_next' => 1 } );
308 my $old_r = $c->reading( 'n21' );
309 $old_r->alter_text( 'to' );
310 $c->del_path( 'n20', 'n21', 'A' );
311 $c->add_path( 'n20', 'n21p0', 'A' );
312 $c->add_path( 'n21p0', 'n21', 'A' );
313 $c->add_relationship( 'n21', 'n22', { type => 'collated', scope => 'local' } );
315 ok( $c->reading( 'n21p0' ), "New reading exists" );
316 is( scalar $c->readings, $rno, "Reading add offset by flatten_ranks" );
318 # Combine n3 and n4 ( with his )
319 $c->merge_readings( 'n3', 'n4', 1 );
320 ok( !$c->reading('n4'), "Reading n4 is gone" );
321 is( $c->reading('n3')->text, 'with his', "Reading n3 has both words" );
323 # Collapse n9 and n10 ( rood / root )
324 $c->merge_readings( 'n9', 'n10' );
325 ok( !$c->reading('n10'), "Reading n10 is gone" );
326 is( $c->reading('n9')->text, 'rood', "Reading n9 has an unchanged word" );
328 # Combine n21 and n21p0
329 my $remaining = $c->reading('n21');
330 $remaining ||= $c->reading('n22'); # one of these should still exist
331 $c->merge_readings( 'n21p0', $remaining, 1 );
332 ok( !$c->reading('n21'), "Reading $remaining is gone" );
333 is( $c->reading('n21p0')->text, 'unto', "Reading n21p0 merged correctly" );
343 my( $kept_obj, $del_obj, $combine, $combine_char ) = $self->_objectify_args( @_ );
344 my $mergemeta = $kept_obj->is_meta;
345 throw( "Cannot merge meta and non-meta reading" )
346 unless ( $mergemeta && $del_obj->is_meta )
347 || ( !$mergemeta && !$del_obj->is_meta );
349 throw( "Cannot merge with start or end node" )
350 if( $kept_obj eq $self->start || $kept_obj eq $self->end
351 || $del_obj eq $self->start || $del_obj eq $self->end );
352 throw( "Cannot combine text of meta readings" ) if $combine;
354 # We only need the IDs for adding paths to the graph, not the reading
355 # objects themselves.
356 my $kept = $kept_obj->id;
357 my $deleted = $del_obj->id;
358 $self->_graphcalc_done(0);
360 # The kept reading should inherit the paths and the relationships
361 # of the deleted reading.
362 foreach my $path ( $self->sequence->edges_at( $deleted ) ) {
363 my @vector = ( $kept );
364 push( @vector, $path->[1] ) if $path->[0] eq $deleted;
365 unshift( @vector, $path->[0] ) if $path->[1] eq $deleted;
366 next if $vector[0] eq $vector[1]; # Don't add a self loop
367 my %wits = %{$self->sequence->get_edge_attributes( @$path )};
368 $self->sequence->add_edge( @vector );
369 my $fwits = $self->sequence->get_edge_attributes( @vector );
370 @wits{keys %$fwits} = values %$fwits;
371 $self->sequence->set_edge_attributes( @vector, \%wits );
373 $self->relations->merge_readings( $kept, $deleted, $combine );
375 # Do the deletion deed.
377 # Combine the text of the readings
378 my $joinstr = $combine_char;
379 unless( defined $joinstr ) {
380 $joinstr = '' if $kept_obj->join_next || $del_obj->join_prior;
381 $joinstr = $self->wordsep unless defined $joinstr;
383 $kept_obj->_combine( $del_obj, $joinstr );
385 $self->del_reading( $deleted );
388 =head2 merge_related( @relationship_types )
390 Merge all readings linked with the relationship types given. If any of the selected type(s) is not a colocation, the graph will no longer be linear. The majority/plurality reading in each case will be the one kept.
392 WARNING: This operation cannot be undone.
401 my $t = Text::Tradition->new(
404 'file' => 't/data/legendfrag.xml',
406 my $c = $t->collation;
409 map { $rdg_ids{$_} = 1 } $c->readings;
410 $c->merge_related( 'orthographic' );
411 is( scalar( $c->readings ), keys( %rdg_ids ) - 8,
412 "Successfully collapsed orthographic variation" );
413 map { $rdg_ids{$_} = undef } qw/ r13.3 r11.4 r8.5 r8.2 r7.7 r7.5 r7.4 r7.1 /;
414 foreach my $rid ( keys %rdg_ids ) {
415 my $exp = $rdg_ids{$rid};
416 is( !$c->reading( $rid ), !$exp, "Reading $rid correctly " .
417 ( $exp ? "retained" : "removed" ) );
419 ok( $c->linear, "Graph is still linear" );
421 $c->calculate_ranks; # This should succeed
422 ok( 1, "Can still calculate ranks on the new graph" );
424 ok( 0, "Rank calculation on merged graph failed: $@" );
427 # Now add some transpositions
428 $c->add_relationship( 'r8.4', 'r10.4', { type => 'transposition' } );
429 $c->merge_related( 'transposition' );
430 is( scalar( $c->readings ), keys( %rdg_ids ) - 9,
431 "Transposed relationship is merged away" );
432 ok( !$c->reading('r8.4'), "Correct transposed reading removed" );
433 ok( !$c->linear, "Graph is no longer linear" );
435 $c->calculate_ranks; # This should fail
436 ok( 0, "Rank calculation happened on nonlinear graph?!" );
437 } catch ( Text::Tradition::Error $e ) {
438 is( $e->message, 'Cannot calculate ranks on a non-linear graph',
439 "Rank calculation on merged graph threw an error" );
448 # TODO: there should be a way to display merged without affecting the underlying data!
453 map { $reltypehash{$_} = 1 } @_;
455 # Set up the filter for finding related readings
457 exists $reltypehash{$_[0]->type};
461 # Go through all readings looking for related ones
462 foreach my $r ( $self->readings ) {
463 next unless $self->reading( "$r" ); # might have been deleted meanwhile
464 my @related = $self->related_readings( $r, $filter );
466 push( @related, $r );
468 scalar $b->witnesses <=> scalar $a->witnesses
470 my $keep = shift @related;
471 foreach my $delr ( @related ) {
473 unless( $self->get_relationship( $keep, $delr )->colocated );
474 $self->merge_readings( $keep, $delr );
478 $self->linear( $linear );
481 =head2 compress_readings
483 Where possible in the graph, compresses plain sequences of readings into a
484 single reading. The sequences must consist of readings with no
485 relationships to other readings, with only a single witness path between
486 them and no other witness paths from either that would skip the other. The
487 readings must also not be marked as nonsense or bad grammar.
489 WARNING: This operation cannot be undone.
493 sub compress_readings {
495 # Anywhere in the graph that there is a reading that joins only to a single
496 # successor, and neither of these have any relationships, just join the two
498 foreach my $rdg ( sort { $a->rank <=> $b->rank } $self->readings ) {
499 # Now look for readings that can be joined to their successors.
500 next unless $rdg->is_combinable;
502 while( $self->sequence->successors( $rdg ) == 1 ) {
503 my( $next ) = $self->reading( $self->sequence->successors( $rdg ) );
504 throw( "Infinite loop" ) if $seen{$next->id};
505 $seen{$next->id} = 1;
506 last if $self->sequence->predecessors( $next ) > 1;
507 last unless $next->is_combinable;
508 say "Joining readings $rdg and $next";
509 $self->merge_readings( $rdg, $next, 1 );
512 # Make sure we haven't screwed anything up
513 foreach my $wit ( $self->tradition->witnesses ) {
514 my $pathtext = $self->path_text( $wit->sigil );
515 my $origtext = join( ' ', @{$wit->text} );
516 throw( "Text differs for witness " . $wit->sigil )
517 unless $pathtext eq $origtext;
518 if( $wit->is_layered ) {
519 $pathtext = $self->path_text( $wit->sigil.$self->ac_label );
520 $origtext = join( ' ', @{$wit->layertext} );
521 throw( "Ante-corr text differs for witness " . $wit->sigil )
522 unless $pathtext eq $origtext;
526 $self->relations->rebuild_equivalence();
527 $self->calculate_ranks();
530 # Helper function for manipulating the graph.
531 sub _stringify_args {
532 my( $self, $first, $second, @args ) = @_;
534 if ref( $first ) eq 'Text::Tradition::Collation::Reading';
535 $second = $second->id
536 if ref( $second ) eq 'Text::Tradition::Collation::Reading';
537 return( $first, $second, @args );
540 # Helper function for manipulating the graph.
541 sub _objectify_args {
542 my( $self, $first, $second, $arg ) = @_;
543 $first = $self->reading( $first )
544 unless ref( $first ) eq 'Text::Tradition::Collation::Reading';
545 $second = $self->reading( $second )
546 unless ref( $second ) eq 'Text::Tradition::Collation::Reading';
547 return( $first, $second, $arg );
550 =head2 duplicate_reading( $reading, @witlist )
552 Split the given reading into two, so that the new reading is in the path for
553 the witnesses given in @witlist. If the result is that certain non-colocated
554 relationships (e.g. transpositions) are no longer valid, these will be removed.
555 Returns the newly-created reading.
559 use Test::More::UTF8;
563 my $st = Text::Tradition->new( 'input' => 'Self', 'file' => 't/data/collatecorr.xml' );
564 is( ref( $st ), 'Text::Tradition', "Got a tradition from test file" );
565 ok( $st->has_witness('Ba96'), "Tradition has the affected witness" );
567 my $sc = $st->collation;
569 ok( $sc->reading('n131'), "Tradition has the affected reading" );
570 is( scalar( $sc->readings ), $numr, "There are $numr readings in the graph" );
571 is( $sc->end->rank, 14, "There are fourteen ranks in the graph" );
573 # Detach the erroneously collated reading
574 my( $newr, @del_rdgs ) = $sc->duplicate_reading( 'n131', 'Ba96' );
575 ok( $newr, "New reading was created" );
576 ok( $sc->reading('n131_0'), "Detached the bad collation with a new reading" );
577 is( scalar( $sc->readings ), $numr + 1, "A reading was added to the graph" );
578 is( $sc->end->rank, 10, "There are now only ten ranks in the graph" );
579 my $csucc = $sc->common_successor( 'n131', 'n131_0' );
580 is( $csucc->id, 'n136', "Found correct common successor to duped reading" );
582 # Check that the bad transposition is gone
583 is( scalar @del_rdgs, 1, "Deleted reading was returned by API call" );
584 is( $sc->get_relationship( 'n130', 'n135' ), undef, "Bad transposition relationship is gone" );
586 # The collation should not be fixed
587 my @pairs = $sc->identical_readings();
588 is( scalar @pairs, 0, "Not re-collated yet" );
590 ok( $sc->merge_readings( 'n124', 'n131_0' ), "Collated the readings correctly" );
591 @pairs = $sc->identical_readings( start => 'n124', end => $csucc->id );
592 is( scalar @pairs, 3, "Found three more identical readings" );
593 is( $sc->end->rank, 11, "The ranks shifted appropriately" );
594 $sc->flatten_ranks();
595 is( scalar( $sc->readings ), $numr - 3, "Now we are collated correctly" );
597 # Check that we can't "duplicate" a reading with no wits or with all wits
599 my( $badr, @del_rdgs ) = $sc->duplicate_reading( 'n124' );
600 ok( 0, "Reading duplication without witnesses throws an error" );
601 } catch( Text::Tradition::Error $e ) {
602 like( $e->message, qr/Must specify one or more witnesses/,
603 "Reading duplication without witnesses throws the expected error" );
605 ok( 0, "Reading duplication without witnesses threw the wrong error" );
609 my( $badr, @del_rdgs ) = $sc->duplicate_reading( 'n124', 'Ba96', 'Mü11475' );
610 ok( 0, "Reading duplication with all witnesses throws an error" );
611 } catch( Text::Tradition::Error $e ) {
612 like( $e->message, qr/Cannot join all witnesses/,
613 "Reading duplication with all witnesses throws the expected error" );
615 ok( 0, "Reading duplication with all witnesses threw the wrong error" );
622 sub duplicate_reading {
623 my( $self, $r, @wits ) = @_;
624 # Check that we are not doing anything unwise.
625 throw( "Must specify one or more witnesses for the duplicated reading" )
627 unless( ref( $r ) eq 'Text::Tradition::Collation::Reading' ) {
628 $r = $self->reading( $r );
630 throw( "Cannot duplicate a meta-reading" )
632 my $ordered_req_wits = join( ',', sort @wits );
633 my $ordered_rdg_wits = join( ',', $r->witnesses );
634 throw( "Cannot join all witnesses to the new reading" )
635 if $ordered_req_wits eq $ordered_rdg_wits;
637 # Get all the reading attributes and duplicate them.
638 my $rmeta = Text::Tradition::Collation::Reading->meta;
640 foreach my $attr( $rmeta->get_all_attributes ) {
641 next if $attr->name =~ /^_/;
642 my $acc = $attr->get_read_method;
643 if( !$acc && $attr->has_applied_traits ) {
644 my $tr = $attr->applied_traits;
645 if( $tr->[0] =~ /::(Array|Hash)$/ ) {
647 my %methods = reverse %{$attr->handles};
648 $acc = $methods{elements};
649 $args{$attr->name} = $which eq 'Array'
650 ? [ $r->$acc ] : { $r->$acc };
653 $args{$attr->name} = $r->$acc if $acc;
656 # By definition the new reading will no longer be common.
657 $args{is_common} = 0;
658 # The new reading also needs its own ID.
659 $args{id} = $self->_generate_dup_id( $r->id );
661 # Try to make the new reading.
662 my $newr = $self->add_reading( \%args );
663 # The old reading is also no longer common.
666 # For each of the witnesses, dissociate from the old reading and
667 # associate with the new.
668 foreach my $wit ( @wits ) {
669 my $prior = $self->prior_reading( $r, $wit );
670 my $next = $self->next_reading( $r, $wit );
671 $self->del_path( $prior, $r, $wit );
672 $self->add_path( $prior, $newr, $wit );
673 $self->del_path( $r, $next, $wit );
674 $self->add_path( $newr, $next, $wit );
677 # If the graph is ranked, we need to look for relationships that are now
678 # invalid (i.e. 'non-colocation' types that might now be colocated) and
679 # remove them. If not, we can skip it.
682 my @deleted_relations;
683 if( $self->end->has_rank ) {
684 # Find the point where we can stop checking
685 $succ = $self->common_successor( $r, $newr );
687 # Hash the existing ranks
688 foreach my $rdg ( $self->readings ) {
689 $rrk{$rdg->id} = $rdg->rank;
691 # Calculate the new ranks
692 $self->calculate_ranks();
694 # Check for invalid non-colocated relationships among changed-rank readings
695 # from where the ranks start changing up to $succ
696 my $lastrank = $succ->rank;
697 foreach my $rdg ( $self->readings ) {
698 next if $rdg->rank > $lastrank;
699 next if $rdg->rank == $rrk{$rdg->id};
700 my @noncolo = $rdg->related_readings( sub { !$_[0]->colocated } );
701 next unless @noncolo;
702 foreach my $nc ( @noncolo ) {
703 unless( $self->relations->verify_or_delete( $rdg, $nc ) ) {
704 push( @deleted_relations, [ $rdg->id, $nc->id ] );
709 return ( $newr, @deleted_relations );
712 sub _generate_dup_id {
713 my( $self, $rid ) = @_;
718 if( $self->has_reading( $newid ) ) {
731 # We only need the IDs for adding paths to the graph, not the reading
732 # objects themselves.
733 my( $source, $target, $wit ) = $self->_stringify_args( @_ );
735 $self->_graphcalc_done(0);
736 # Connect the readings
737 unless( $self->sequence->has_edge( $source, $target ) ) {
738 $self->sequence->add_edge( $source, $target );
739 $self->relations->add_equivalence_edge( $source, $target );
741 # Note the witness in question
742 $self->sequence->set_edge_attribute( $source, $target, $wit, 1 );
748 if( ref( $_[0] ) eq 'ARRAY' ) {
755 # We only need the IDs for removing paths from the graph, not the reading
756 # objects themselves.
757 my( $source, $target, $wit ) = $self->_stringify_args( @args );
759 $self->_graphcalc_done(0);
760 if( $self->sequence->has_edge_attribute( $source, $target, $wit ) ) {
761 $self->sequence->delete_edge_attribute( $source, $target, $wit );
763 unless( $self->sequence->has_edge_attributes( $source, $target ) ) {
764 $self->sequence->delete_edge( $source, $target );
765 $self->relations->delete_equivalence_edge( $source, $target );
770 # Extra graph-alike utility
773 my( $source, $target, $wit ) = $self->_stringify_args( @_ );
774 return undef unless $self->sequence->has_edge( $source, $target );
775 return $self->sequence->has_edge_attribute( $source, $target, $wit );
778 =head2 clear_witness( @sigil_list )
780 Clear the given witnesses out of the collation entirely, removing references
781 to them in paths, and removing readings that belong only to them. Should only
782 be called via $tradition->del_witness.
787 my( $self, @sigils ) = @_;
789 $self->_graphcalc_done(0);
790 # Clear the witness(es) out of the paths
791 foreach my $e ( $self->paths ) {
792 foreach my $sig ( @sigils ) {
793 $self->del_path( $e, $sig );
797 # Clear out the newly unused readings
798 foreach my $r ( $self->readings ) {
799 unless( $self->reading_witnesses( $r ) ) {
800 $self->del_reading( $r );
805 sub add_relationship {
807 my( $source, $target, $opts ) = $self->_stringify_args( @_ );
808 my( @vectors ) = $self->relations->add_relationship( $source, $target, $opts );
809 foreach my $v ( @vectors ) {
810 next unless $self->get_relationship( $v )->colocated;
811 if( $self->reading( $v->[0] )->has_rank && $self->reading( $v->[1] )->has_rank
812 && $self->reading( $v->[0] )->rank ne $self->reading( $v->[1] )->rank ) {
813 $self->_graphcalc_done(0);
821 around qw/ get_relationship del_relationship / => sub {
825 if( @args == 1 && ref( $args[0] ) eq 'ARRAY' ) {
828 my @stringargs = $self->_stringify_args( @args );
829 $self->$orig( @stringargs );
832 =head2 reading_witnesses( $reading )
834 Return a list of sigils corresponding to the witnesses in which the reading appears.
838 sub reading_witnesses {
839 my( $self, $reading ) = @_;
840 # We need only check either the incoming or the outgoing edges; I have
841 # arbitrarily chosen "incoming". Thus, special-case the start node.
842 if( $reading eq $self->start ) {
843 return map { $_->sigil } grep { $_->is_collated } $self->tradition->witnesses;
846 foreach my $e ( $self->sequence->edges_to( $reading ) ) {
847 my $wits = $self->sequence->get_edge_attributes( @$e );
848 @all_witnesses{ keys %$wits } = 1;
850 my $acstr = $self->ac_label;
851 foreach my $acwit ( grep { $_ =~ s/^(.*)\Q$acstr\E$/$1/ } keys %all_witnesses ) {
852 delete $all_witnesses{$acwit.$acstr} if exists $all_witnesses{$acwit};
854 return keys %all_witnesses;
857 =head1 OUTPUT METHODS
859 =head2 as_svg( \%options )
861 Returns an SVG string that represents the graph, via as_dot and graphviz.
862 See as_dot for a list of options. Must have GraphViz (dot) installed to run.
867 my( $self, $opts ) = @_;
868 throw( "Need GraphViz installed to output SVG" )
869 unless File::Which::which( 'dot' );
870 my $want_subgraph = exists $opts->{'from'} || exists $opts->{'to'};
871 $self->calculate_ranks()
872 unless( $self->_graphcalc_done || $opts->{'nocalc'} || !$self->linear );
873 my @cmd = qw/dot -Tsvg/;
875 my $dotfile = File::Temp->new();
877 # $dotfile->unlink_on_destroy(0);
878 binmode $dotfile, ':utf8';
879 print $dotfile $self->as_dot( $opts );
880 push( @cmd, $dotfile->filename );
881 run( \@cmd, ">", binary(), \$svg );
882 $svg = decode_utf8( $svg );
887 =head2 as_dot( \%options )
889 Returns a string that is the collation graph expressed in dot
890 (i.e. GraphViz) format. Options include:
905 my( $self, $opts ) = @_;
906 my $startrank = $opts->{'from'} if $opts;
907 my $endrank = $opts->{'to'} if $opts;
908 my $color_common = $opts->{'color_common'} if $opts;
909 my $STRAIGHTENHACK = !$startrank && !$endrank && $self->end->rank
910 && $self->end->rank > 100;
911 $STRAIGHTENHACK = 1 if $opts->{'straight'}; # even for subgraphs or small graphs
913 # Check the arguments
915 return if $endrank && $startrank > $endrank;
916 return if $startrank > $self->end->rank;
918 if( defined $endrank ) {
919 return if $endrank < 0;
920 $endrank = undef if $endrank == $self->end->rank;
923 my $graph_name = $self->tradition->name;
924 $graph_name =~ s/[^\w\s]//g;
925 $graph_name = join( '_', split( /\s+/, $graph_name ) );
933 'fillcolor' => 'white',
938 'arrowhead' => 'open',
939 'color' => '#000000',
940 'fontcolor' => '#000000',
943 my $dot = sprintf( "digraph %s {\n", $graph_name );
944 $dot .= "\tgraph " . _dot_attr_string( \%graph_attrs ) . ";\n";
945 $dot .= "\tnode " . _dot_attr_string( \%node_attrs ) . ";\n";
947 # Output substitute start/end readings if necessary
949 $dot .= "\t\"__SUBSTART__\" [ label=\"...\",id=\"__START__\" ];\n";
952 $dot .= "\t\"__SUBEND__\" [ label=\"...\",id=\"__END__\" ];\n";
954 if( $STRAIGHTENHACK ) {
956 my $startlabel = $startrank ? '__SUBSTART__' : '__START__';
957 $dot .= "\tsubgraph { rank=same \"$startlabel\" \"#SILENT#\" }\n";
958 $dot .= "\t\"#SILENT#\" [ shape=diamond,color=white,penwidth=0,label=\"\" ];"
960 my %used; # Keep track of the readings that actually appear in the graph
961 # Sort the readings by rank if we have ranks; this speeds layout.
962 my @all_readings = $self->end->has_rank
963 ? sort { $a->rank <=> $b->rank } $self->readings
965 # TODO Refrain from outputting lacuna nodes - just grey out the edges.
966 foreach my $reading ( @all_readings ) {
967 # Only output readings within our rank range.
968 next if $startrank && $reading->rank < $startrank;
969 next if $endrank && $reading->rank > $endrank;
970 $used{$reading->id} = 1;
971 # Need not output nodes without separate labels
972 next if $reading->id eq $reading->text;
974 my $label = $reading->text;
975 $label .= '-' if $reading->join_next;
976 $label = "-$label" if $reading->join_prior;
977 $label =~ s/\"/\\\"/g;
978 $rattrs->{'label'} = $label;
979 $rattrs->{'id'} = $reading->id;
980 $rattrs->{'fillcolor'} = '#b3f36d' if $reading->is_common && $color_common;
981 $dot .= sprintf( "\t\"%s\" %s;\n", $reading->id, _dot_attr_string( $rattrs ) );
984 # Add the real edges. Need to weight one edge per rank jump, in a
986 # my $weighted = $self->_add_edge_weights;
987 my @edges = $self->paths;
988 my( %substart, %subend );
989 foreach my $edge ( @edges ) {
990 # Do we need to output this edge?
991 if( $used{$edge->[0]} && $used{$edge->[1]} ) {
992 my $label = $self->_path_display_label( $opts,
993 $self->path_witnesses( $edge ) );
994 my $variables = { %edge_attrs, 'label' => $label };
996 # Account for the rank gap if necessary
997 my $rank0 = $self->reading( $edge->[0] )->rank
998 if $self->reading( $edge->[0] )->has_rank;
999 my $rank1 = $self->reading( $edge->[1] )->rank
1000 if $self->reading( $edge->[1] )->has_rank;
1001 if( defined $rank0 && defined $rank1 && $rank1 - $rank0 > 1 ) {
1002 $variables->{'minlen'} = $rank1 - $rank0;
1005 # Add the calculated edge weights
1006 # if( exists $weighted->{$edge->[0]}
1007 # && $weighted->{$edge->[0]} eq $edge->[1] ) {
1008 # # $variables->{'color'} = 'red';
1009 # $variables->{'weight'} = 3.0;
1012 # EXPERIMENTAL: make edge width reflect no. of witnesses
1013 my $extrawidth = scalar( $self->path_witnesses( $edge ) ) * 0.2;
1014 $variables->{'penwidth'} = $extrawidth + 0.8; # gives 1 for a single wit
1016 my $varopts = _dot_attr_string( $variables );
1017 $dot .= sprintf( "\t\"%s\" -> \"%s\" %s;\n",
1018 $edge->[0], $edge->[1], $varopts );
1019 } elsif( $used{$edge->[0]} ) {
1020 $subend{$edge->[0]} = $edge->[1];
1021 } elsif( $used{$edge->[1]} ) {
1022 $substart{$edge->[1]} = $edge->[0];
1026 # If we are asked to, add relationship links
1027 if( exists $opts->{show_relations} ) {
1028 my $filter = $opts->{show_relations}; # can be 'transposition' or 'all'
1029 if( $filter eq 'transposition' ) {
1030 $filter =~ qr/^transposition$/;
1032 foreach my $redge ( $self->relationships ) {
1033 if( $used{$redge->[0]} && $used{$redge->[1]} ) {
1034 if( $filter ne 'all' ) {
1035 my $rel = $self->get_relationship( $redge );
1036 next unless $rel->type =~ /$filter/;
1038 arrowhead => 'none',
1040 constraint => 'false',
1041 label => uc( substr( $rel->type, 0, 4 ) ),
1044 $dot .= sprintf( "\t\"%s\" -> \"%s\" %s;\n",
1045 $redge->[0], $redge->[1], _dot_attr_string( $variables ) );
1051 # Add substitute start and end edges if necessary
1052 foreach my $node ( keys %substart ) {
1053 my $witstr = $self->_path_display_label( $opts,
1054 $self->path_witnesses( $substart{$node}, $node ) );
1055 my $variables = { %edge_attrs, 'label' => $witstr };
1056 my $nrdg = $self->reading( $node );
1057 if( $nrdg->has_rank && $nrdg->rank > $startrank ) {
1058 # Substart is actually one lower than $startrank
1059 $variables->{'minlen'} = $nrdg->rank - ( $startrank - 1 );
1061 my $varopts = _dot_attr_string( $variables );
1062 $dot .= "\t\"__SUBSTART__\" -> \"$node\" $varopts;\n";
1064 foreach my $node ( keys %subend ) {
1065 my $witstr = $self->_path_display_label( $opts,
1066 $self->path_witnesses( $node, $subend{$node} ) );
1067 my $variables = { %edge_attrs, 'label' => $witstr };
1068 my $varopts = _dot_attr_string( $variables );
1069 $dot .= "\t\"$node\" -> \"__SUBEND__\" $varopts;\n";
1072 if( $STRAIGHTENHACK ) {
1073 my $endlabel = $endrank ? '__SUBEND__' : '__END__';
1074 $dot .= "\t\"$endlabel\" -> \"#SILENT#\" [ color=white,penwidth=0 ];\n";
1081 sub _dot_attr_string {
1084 foreach my $k ( sort keys %$hash ) {
1085 my $v = $hash->{$k};
1086 push( @attrs, $k.'="'.$v.'"' );
1088 return( '[ ' . join( ', ', @attrs ) . ' ]' );
1091 sub _add_edge_weights {
1093 # Walk the graph from START to END, choosing the successor node with
1094 # the largest number of witness paths each time.
1096 my $curr = $self->start->id;
1097 my $ranked = $self->end->has_rank;
1098 while( $curr ne $self->end->id ) {
1099 my $rank = $ranked ? $self->reading( $curr )->rank : 0;
1100 my @succ = sort { $self->path_witnesses( $curr, $a )
1101 <=> $self->path_witnesses( $curr, $b ) }
1102 $self->sequence->successors( $curr );
1103 my $next = pop @succ;
1104 my $nextrank = $ranked ? $self->reading( $next )->rank : 0;
1105 # Try to avoid lacunae in the weighted path.
1107 ( $self->reading( $next )->is_lacuna ||
1108 $nextrank - $rank > 1 ) ){
1111 $weighted->{$curr} = $next;
1117 =head2 path_witnesses( $edge )
1119 Returns the list of sigils whose witnesses are associated with the given edge.
1120 The edge can be passed as either an array or an arrayref of ( $source, $target ).
1124 sub path_witnesses {
1125 my( $self, @edge ) = @_;
1126 # If edge is an arrayref, cope.
1127 if( @edge == 1 && ref( $edge[0] ) eq 'ARRAY' ) {
1128 my $e = shift @edge;
1131 my @wits = keys %{$self->sequence->get_edge_attributes( @edge )};
1135 # Helper function. Make a display label for the given witnesses, showing a.c.
1136 # witnesses only where the main witness is not also in the list.
1137 sub _path_display_label {
1141 map { $wits{$_} = 1 } @_;
1143 # If an a.c. wit is listed, remove it if the main wit is also listed.
1144 # Otherwise keep it for explicit listing.
1145 my $aclabel = $self->ac_label;
1147 foreach my $w ( sort keys %wits ) {
1148 if( $w =~ /^(.*)\Q$aclabel\E$/ ) {
1149 if( exists $wits{$1} ) {
1152 push( @disp_ac, $w );
1157 if( $opts->{'explicit_wits'} ) {
1158 return join( ', ', sort keys %wits );
1160 # See if we are in a majority situation.
1161 my $maj = scalar( $self->tradition->witnesses ) * 0.6;
1162 $maj = $maj > 5 ? $maj : 5;
1163 if( scalar keys %wits > $maj ) {
1164 unshift( @disp_ac, 'majority' );
1165 return join( ', ', @disp_ac );
1167 return join( ', ', sort keys %wits );
1172 =head2 readings_at_rank( $rank )
1174 Returns a list of readings at a given rank, taken from the alignment table.
1178 sub readings_at_rank {
1179 my( $self, $rank ) = @_;
1180 my $table = $self->alignment_table;
1181 # Table rank is real rank - 1.
1182 my @elements = map { $_->{'tokens'}->[$rank-1] } @{$table->{'alignment'}};
1184 foreach my $e ( @elements ) {
1185 next unless ref( $e ) eq 'HASH';
1186 next unless exists $e->{'t'};
1187 $readings{$e->{'t'}->id} = $e->{'t'};
1189 return values %readings;
1194 Returns a GraphML representation of the collation. The GraphML will contain
1195 two graphs. The first expresses the attributes of the readings and the witness
1196 paths that link them; the second expresses the relationships that link the
1197 readings. This is the native transfer format for a tradition.
1201 use Text::Tradition;
1207 my $datafile = 't/data/florilegium_tei_ps.xml';
1208 my $tradition = Text::Tradition->new( 'input' => 'TEI',
1210 'file' => $datafile,
1213 ok( $tradition, "Got a tradition object" );
1214 is( scalar $tradition->witnesses, 13, "Found all witnesses" );
1215 ok( $tradition->collation, "Tradition has a collation" );
1217 my $c = $tradition->collation;
1218 is( scalar $c->readings, $READINGS, "Collation has all readings" );
1219 is( scalar $c->paths, $PATHS, "Collation has all paths" );
1220 is( scalar $c->relationships, 0, "Collation has all relationships" );
1222 # Add a few relationships
1223 $c->add_relationship( 'w123', 'w125', { 'type' => 'collated' } );
1224 $c->add_relationship( 'w193', 'w196', { 'type' => 'collated' } );
1225 $c->add_relationship( 'w257', 'w262', { 'type' => 'transposition' } );
1227 # Now write it to GraphML and parse it again.
1229 my $graphml = $c->as_graphml;
1230 my $st = Text::Tradition->new( 'input' => 'Self', 'string' => $graphml );
1231 is( scalar $st->collation->readings, $READINGS, "Reparsed collation has all readings" );
1232 is( scalar $st->collation->paths, $PATHS, "Reparsed collation has all paths" );
1233 is( scalar $st->collation->relationships, 3, "Reparsed collation has new relationships" );
1235 # Now add a stemma, write to GraphML, and look at the output.
1237 skip "Analysis module not present", 3 unless $tradition->can( 'add_stemma' );
1238 my $stemma = $tradition->add_stemma( 'dotfile' => 't/data/florilegium.dot' );
1239 is( ref( $stemma ), 'Text::Tradition::Stemma', "Parsed dotfile into stemma" );
1240 is( $tradition->stemmata, 1, "Tradition now has the stemma" );
1241 $graphml = $c->as_graphml;
1242 like( $graphml, qr/digraph/, "Digraph declaration exists in GraphML" );
1249 ## TODO MOVE this to Tradition.pm and modularize it better
1251 my( $self, $options ) = @_;
1252 $self->calculate_ranks unless $self->_graphcalc_done;
1254 my $start = $options->{'from'}
1255 ? $self->reading( $options->{'from'} ) : $self->start;
1256 my $end = $options->{'to'}
1257 ? $self->reading( $options->{'to'} ) : $self->end;
1258 if( $start->has_rank && $end->has_rank && $end->rank < $start->rank ) {
1259 throw( 'Start node must be before end node' );
1261 # The readings need to be ranked for this to work.
1262 $start = $self->start unless $start->has_rank;
1263 $end = $self->end unless $end->has_rank;
1265 unless( $start eq $self->start ) {
1266 $rankoffset = $start->rank - 1;
1271 my $graphml_ns = 'http://graphml.graphdrawing.org/xmlns';
1272 my $xsi_ns = 'http://www.w3.org/2001/XMLSchema-instance';
1273 my $graphml_schema = 'http://graphml.graphdrawing.org/xmlns ' .
1274 'http://graphml.graphdrawing.org/xmlns/1.0/graphml.xsd';
1276 # Create the document and root node
1277 require XML::LibXML;
1278 my $graphml = XML::LibXML->createDocument( "1.0", "UTF-8" );
1279 my $root = $graphml->createElementNS( $graphml_ns, 'graphml' );
1280 $graphml->setDocumentElement( $root );
1281 $root->setNamespace( $xsi_ns, 'xsi', 0 );
1282 $root->setAttributeNS( $xsi_ns, 'schemaLocation', $graphml_schema );
1284 # List of attribute types to save on our objects and their corresponding
1289 'Bool' => 'boolean',
1290 'ReadingID' => 'string',
1291 'RelationshipType' => 'string',
1292 'RelationshipScope' => 'string',
1295 # Add the data keys for the graph. Include an extra key 'version' for the
1296 # GraphML output version.
1297 my %graph_data_keys;
1299 my %graph_attributes = ( 'version' => 'string' );
1300 # Graph attributes include those of Tradition and those of Collation.
1302 # TODO Use meta introspection method from duplicate_reading to do this
1303 # instead of naming custom keys.
1304 my $tmeta = $self->tradition->meta;
1305 my $cmeta = $self->meta;
1306 map { $gattr_from{$_->name} = 'Tradition' } $tmeta->get_all_attributes;
1307 map { $gattr_from{$_->name} = 'Collation' } $cmeta->get_all_attributes;
1308 foreach my $attr ( ( $tmeta->get_all_attributes, $cmeta->get_all_attributes ) ) {
1309 next if $attr->name =~ /^_/;
1310 next unless $save_types{$attr->type_constraint->name};
1311 $graph_attributes{$attr->name} = $save_types{$attr->type_constraint->name};
1313 # Extra custom keys for complex objects that should be saved in some form.
1314 # The subroutine should return a string, or undef/empty.
1315 if( $tmeta->has_method('stemmata') ) {
1316 $graph_attributes{'stemmata'} = sub {
1318 map { push( @stemstrs, $_->editable( {linesep => ''} ) ) }
1319 $self->tradition->stemmata;
1320 join( "\n", @stemstrs );
1324 if( $tmeta->has_method('user') ) {
1325 $graph_attributes{'user'} = sub {
1326 $self->tradition->user ? $self->tradition->user->id : undef
1330 foreach my $datum ( sort keys %graph_attributes ) {
1331 $graph_data_keys{$datum} = 'dg'.$gdi++;
1332 my $key = $root->addNewChild( $graphml_ns, 'key' );
1333 my $dtype = ref( $graph_attributes{$datum} ) ? 'string'
1334 : $graph_attributes{$datum};
1335 $key->setAttribute( 'attr.name', $datum );
1336 $key->setAttribute( 'attr.type', $dtype );
1337 $key->setAttribute( 'for', 'graph' );
1338 $key->setAttribute( 'id', $graph_data_keys{$datum} );
1341 # Add the data keys for reading nodes
1342 my %reading_attributes;
1343 my $rmeta = Text::Tradition::Collation::Reading->meta;
1344 foreach my $attr( $rmeta->get_all_attributes ) {
1345 next if $attr->name =~ /^_/;
1346 next unless $save_types{$attr->type_constraint->name};
1347 $reading_attributes{$attr->name} = $save_types{$attr->type_constraint->name};
1349 if( $self->start->does('Text::Tradition::Morphology' ) ) {
1350 # Extra custom key for the reading morphology
1351 $reading_attributes{'lexemes'} = 'string';
1356 foreach my $datum ( sort keys %reading_attributes ) {
1357 $node_data_keys{$datum} = 'dn'.$ndi++;
1358 my $key = $root->addNewChild( $graphml_ns, 'key' );
1359 $key->setAttribute( 'attr.name', $datum );
1360 $key->setAttribute( 'attr.type', $reading_attributes{$datum} );
1361 $key->setAttribute( 'for', 'node' );
1362 $key->setAttribute( 'id', $node_data_keys{$datum} );
1365 # Add the data keys for edges, that is, paths and relationships. Path
1366 # data does not come from a Moose class so is here manually.
1369 my %edge_attributes = (
1370 witness => 'string', # ID/label for a path
1371 extra => 'boolean', # Path key
1373 my @path_attributes = keys %edge_attributes; # track our manual additions
1374 my $pmeta = Text::Tradition::Collation::Relationship->meta;
1375 foreach my $attr( $pmeta->get_all_attributes ) {
1376 next if $attr->name =~ /^_/;
1377 next unless $save_types{$attr->type_constraint->name};
1378 $edge_attributes{$attr->name} = $save_types{$attr->type_constraint->name};
1380 foreach my $datum ( sort keys %edge_attributes ) {
1381 $edge_data_keys{$datum} = 'de'.$edi++;
1382 my $key = $root->addNewChild( $graphml_ns, 'key' );
1383 $key->setAttribute( 'attr.name', $datum );
1384 $key->setAttribute( 'attr.type', $edge_attributes{$datum} );
1385 $key->setAttribute( 'for', 'edge' );
1386 $key->setAttribute( 'id', $edge_data_keys{$datum} );
1389 # Add the collation graph itself. First, sanitize the name to a valid XML ID.
1390 my $xmlidname = $self->tradition->name;
1391 $xmlidname =~ s/(?!$xml10_namechar_rx)./_/g;
1392 if( $xmlidname !~ /^$xml10_namestartchar_rx/ ) {
1393 $xmlidname = '_'.$xmlidname;
1395 my $sgraph = $root->addNewChild( $graphml_ns, 'graph' );
1396 $sgraph->setAttribute( 'edgedefault', 'directed' );
1397 $sgraph->setAttribute( 'id', $xmlidname );
1398 $sgraph->setAttribute( 'parse.edgeids', 'canonical' );
1399 $sgraph->setAttribute( 'parse.edges', 0 ); # fill in later
1400 $sgraph->setAttribute( 'parse.nodeids', 'canonical' );
1401 $sgraph->setAttribute( 'parse.nodes', 0 ); # fill in later
1402 $sgraph->setAttribute( 'parse.order', 'nodesfirst' );
1404 # Tradition/collation attribute data
1405 foreach my $datum ( keys %graph_attributes ) {
1407 if( $datum eq 'version' ) {
1409 } elsif( ref( $graph_attributes{$datum} ) ) {
1410 my $sub = $graph_attributes{$datum};
1412 } elsif( $gattr_from{$datum} eq 'Tradition' ) {
1413 $value = $self->tradition->$datum;
1415 $value = $self->$datum;
1417 _add_graphml_data( $sgraph, $graph_data_keys{$datum}, $value );
1422 # Add our readings to the graph
1423 foreach my $n ( sort { $a->id cmp $b->id } $self->readings ) {
1424 next if $n->has_rank && $n ne $self->start && $n ne $self->end &&
1425 ( $n->rank < $start->rank || $n->rank > $end->rank );
1426 $use_readings{$n->id} = 1;
1427 # Add to the main graph
1428 my $node_el = $sgraph->addNewChild( $graphml_ns, 'node' );
1429 my $node_xmlid = 'n' . $node_ctr++;
1430 $node_hash{ $n->id } = $node_xmlid;
1431 $node_el->setAttribute( 'id', $node_xmlid );
1432 foreach my $d ( keys %reading_attributes ) {
1434 # Custom serialization
1435 if( $d eq 'lexemes' ) {
1436 # If nval is a true value, we have lexemes so we need to
1437 # serialize them. Otherwise set nval to undef so that the
1438 # key is excluded from this reading.
1439 $nval = $nval ? $n->_serialize_lexemes : undef;
1440 } elsif( $d eq 'normal_form' && $n->normal_form eq $n->text ) {
1443 if( $rankoffset && $d eq 'rank' && $n ne $self->start ) {
1444 # Adjust the ranks within the subgraph.
1445 $nval = $n eq $self->end ? $end->rank - $rankoffset + 1
1446 : $nval - $rankoffset;
1448 _add_graphml_data( $node_el, $node_data_keys{$d}, $nval )
1453 # Add the path edges to the sequence graph
1455 foreach my $e ( sort { $a->[0] cmp $b->[0] } $self->sequence->edges() ) {
1456 # We add an edge in the graphml for every witness in $e.
1457 next unless( $use_readings{$e->[0]} || $use_readings{$e->[1]} );
1458 my @edge_wits = sort $self->path_witnesses( $e );
1459 $e->[0] = $self->start->id unless $use_readings{$e->[0]};
1460 $e->[1] = $self->end->id unless $use_readings{$e->[1]};
1461 # Skip any path from start to end; that witness is not in the subgraph.
1462 next if ( $e->[0] eq $self->start->id && $e->[1] eq $self->end->id );
1463 foreach my $wit ( @edge_wits ) {
1464 my( $id, $from, $to ) = ( 'e'.$edge_ctr++,
1465 $node_hash{ $e->[0] },
1466 $node_hash{ $e->[1] } );
1467 my $edge_el = $sgraph->addNewChild( $graphml_ns, 'edge' );
1468 $edge_el->setAttribute( 'source', $from );
1469 $edge_el->setAttribute( 'target', $to );
1470 $edge_el->setAttribute( 'id', $id );
1472 # It's a witness path, so add the witness
1474 my $key = $edge_data_keys{'witness'};
1475 # Is this an ante-corr witness?
1476 my $aclabel = $self->ac_label;
1477 if( $wit =~ /^(.*)\Q$aclabel\E$/ ) {
1478 # Keep the base witness
1480 # ...and record that this is an 'extra' reading path
1481 _add_graphml_data( $edge_el, $edge_data_keys{'extra'}, $aclabel );
1483 _add_graphml_data( $edge_el, $edge_data_keys{'witness'}, $base );
1487 # Report the actual number of nodes and edges that went in
1488 $sgraph->setAttribute( 'parse.edges', $edge_ctr );
1489 $sgraph->setAttribute( 'parse.nodes', $node_ctr );
1491 # Add the relationship graph to the XML
1492 map { delete $edge_data_keys{$_} } @path_attributes;
1493 $self->relations->_as_graphml( $graphml_ns, $root, \%node_hash,
1494 $node_data_keys{'id'}, \%edge_data_keys );
1496 # Save and return the thing
1497 my $result = decode_utf8( $graphml->toString(1) );
1501 sub _add_graphml_data {
1502 my( $el, $key, $value ) = @_;
1503 return unless defined $value;
1504 my $data_el = $el->addNewChild( $el->namespaceURI, 'data' );
1505 $data_el->setAttribute( 'key', $key );
1506 $data_el->appendText( $value );
1511 Returns a CSV alignment table representation of the collation graph, one
1512 row per witness (or witness uncorrected.)
1516 Returns a tab-separated alignment table representation of the collation graph,
1517 one row per witness (or witness uncorrected.)
1521 use Text::Tradition;
1529 my $datafile = 't/data/florilegium_tei_ps.xml';
1530 my $tradition = Text::Tradition->new( 'input' => 'TEI',
1532 'file' => $datafile,
1535 my $c = $tradition->collation;
1536 # Export the thing to CSV
1537 my $csvstr = $c->as_csv();
1539 my $csv = Text::CSV->new({ sep_char => ',', binary => 1 });
1540 my @lines = split(/\n/, $csvstr );
1541 ok( $csv->parse( $lines[0] ), "Successfully parsed first line of CSV" );
1542 is( scalar( $csv->fields ), $WITS + $WITAC, "CSV has correct number of witness columns" );
1543 my @q_ac = grep { $_ eq 'Q'.$c->ac_label } $csv->fields;
1544 ok( @q_ac, "Found a layered witness" );
1546 my $t2 = Text::Tradition->new( input => 'Tabular',
1550 is( scalar $t2->collation->readings, $READINGS, "Reparsed CSV collation has all readings" );
1551 is( scalar $t2->collation->paths, $PATHS, "Reparsed CSV collation has all paths" );
1553 # Now do it with TSV
1554 my $tsvstr = $c->as_tsv();
1555 my $t3 = Text::Tradition->new( input => 'Tabular',
1559 is( scalar $t3->collation->readings, $READINGS, "Reparsed TSV collation has all readings" );
1560 is( scalar $t3->collation->paths, $PATHS, "Reparsed TSV collation has all paths" );
1562 my $table = $c->alignment_table;
1563 my $noaccsv = $c->as_csv({ noac => 1 });
1564 my @noaclines = split(/\n/, $noaccsv );
1565 ok( $csv->parse( $noaclines[0] ), "Successfully parsed first line of no-ac CSV" );
1566 is( scalar( $csv->fields ), $WITS, "CSV has correct number of witness columns" );
1567 is( $c->alignment_table, $table, "Request for CSV did not alter the alignment table" );
1569 my $safecsv = $c->as_csv({ safe_ac => 1});
1570 my @safelines = split(/\n/, $safecsv );
1571 ok( $csv->parse( $safelines[0] ), "Successfully parsed first line of safe CSV" );
1572 is( scalar( $csv->fields ), $WITS + $WITAC, "CSV has correct number of witness columns" );
1573 @q_ac = grep { $_ eq 'Q__L' } $csv->fields;
1574 ok( @q_ac, "Found a sanitized layered witness" );
1575 is( $c->alignment_table, $table, "Request for CSV did not alter the alignment table" );
1582 my( $self, $opts ) = @_;
1583 my $table = $self->alignment_table( $opts );
1584 my $csv_options = { binary => 1, quote_null => 0 };
1585 $csv_options->{'sep_char'} = $opts->{fieldsep};
1586 if( $opts->{fieldsep} eq "\t" ) {
1587 # If it is really tab separated, nothing is an escape char.
1588 $csv_options->{'quote_char'} = undef;
1589 $csv_options->{'escape_char'} = '';
1591 my $csv = Text::CSV->new( $csv_options );
1593 # Make the header row
1594 $csv->combine( map { $_->{'witness'} } @{$table->{'alignment'}} );
1595 push( @result, $csv->string );
1596 # Make the rest of the rows
1597 foreach my $idx ( 0 .. $table->{'length'} - 1 ) {
1598 my @rowobjs = map { $_->{'tokens'}->[$idx] } @{$table->{'alignment'}};
1599 my @row = map { $_ ? $_->{'t'}->text : $_ } @rowobjs;
1600 $csv->combine( @row );
1601 push( @result, $csv->string );
1603 return join( "\n", @result );
1608 my $opts = shift || {};
1609 $opts->{fieldsep} = ',';
1610 return $self->_tabular( $opts );
1615 my $opts = shift || {};
1616 $opts->{fieldsep} = "\t";
1617 return $self->_tabular( $opts );
1620 =head2 alignment_table
1622 Return a reference to an alignment table, in a slightly enhanced CollateX
1623 format which looks like this:
1625 $table = { alignment => [ { witness => "SIGIL",
1626 tokens => [ { t => "TEXT" }, ... ] },
1627 { witness => "SIG2",
1628 tokens => [ { t => "TEXT" }, ... ] },
1630 length => TEXTLEN };
1634 sub alignment_table {
1635 my( $self, $opts ) = @_;
1636 if( $self->has_cached_table ) {
1637 return $self->cached_table
1638 unless $opts->{noac} || $opts->{safe_ac};
1641 # Make sure we can do this
1642 throw( "Need a linear graph in order to make an alignment table" )
1643 unless $self->linear;
1644 $self->calculate_ranks()
1645 unless $self->_graphcalc_done && $self->end->has_rank;
1647 my $table = { 'alignment' => [], 'length' => $self->end->rank - 1 };
1648 my @all_pos = ( 1 .. $self->end->rank - 1 );
1649 foreach my $wit ( sort { $a->sigil cmp $b->sigil } $self->tradition->witnesses ) {
1650 # say STDERR "Making witness row(s) for " . $wit->sigil;
1651 my @wit_path = $self->reading_sequence( $self->start, $self->end, $wit->sigil );
1652 my @row = _make_witness_row( \@wit_path, \@all_pos );
1653 my $witobj = { 'witness' => $wit->sigil, 'tokens' => \@row };
1654 $witobj->{'identifier'} = $wit->identifier if $wit->identifier;
1655 push( @{$table->{'alignment'}}, $witobj );
1656 if( $wit->is_layered && !$opts->{noac} ) {
1657 my @wit_ac_path = $self->reading_sequence( $self->start, $self->end,
1658 $wit->sigil.$self->ac_label );
1659 my @ac_row = _make_witness_row( \@wit_ac_path, \@all_pos );
1660 my $witlabel = $opts->{safe_ac}
1661 ? $wit->sigil . '__L' : $wit->sigil.$self->ac_label;
1662 my $witacobj = { 'witness' => $witlabel,
1663 'tokens' => \@ac_row };
1664 $witacobj->{'identifier'} = $wit->identifier if $wit->identifier;
1665 push( @{$table->{'alignment'}}, $witacobj );
1668 unless( $opts->{noac} || $opts->{safe_ac} ) {
1669 $self->cached_table( $table );
1674 sub _make_witness_row {
1675 my( $path, $positions ) = @_;
1677 map { $char_hash{$_} = undef } @$positions;
1679 foreach my $rdg ( @$path ) {
1680 say STDERR "rank " . $rdg->rank if $debug;
1681 # say STDERR "No rank for " . $rdg->id unless defined $rdg->rank;
1682 $char_hash{$rdg->rank} = { 't' => $rdg };
1684 my @row = map { $char_hash{$_} } @$positions;
1685 # Fill in lacuna markers for undef spots in the row
1686 my $last_el = shift @row;
1687 my @filled_row = ( $last_el );
1688 foreach my $el ( @row ) {
1689 # If we are using node reference, make the lacuna node appear many times
1690 # in the table. If not, use the lacuna tag.
1691 if( $last_el && $last_el->{'t'}->is_lacuna && !defined $el ) {
1694 push( @filled_row, $el );
1701 =head1 NAVIGATION METHODS
1703 =head2 reading_sequence( $first, $last, $sigil, $backup )
1705 Returns the ordered list of readings, starting with $first and ending
1706 with $last, for the witness given in $sigil. If a $backup sigil is
1707 specified (e.g. when walking a layered witness), it will be used wherever
1708 no $sigil path exists. If there is a base text reading, that will be
1709 used wherever no path exists for $sigil or $backup.
1713 # TODO Think about returning some lazy-eval iterator.
1714 # TODO Get rid of backup; we should know from what witness is whether we need it.
1716 sub reading_sequence {
1717 my( $self, $start, $end, $witness ) = @_;
1719 $witness = $self->baselabel unless $witness;
1720 my @readings = ( $start );
1723 while( $n && $n->id ne $end->id ) {
1724 if( exists( $seen{$n->id} ) ) {
1725 throw( "Detected loop for $witness at " . $n->id );
1729 my $next = $self->next_reading( $n, $witness );
1731 throw( "Did not find any path for $witness from reading " . $n->id );
1733 push( @readings, $next );
1736 # Check that the last reading is our end reading.
1737 my $last = $readings[$#readings];
1738 throw( "Last reading found from " . $start->text .
1739 " for witness $witness is not the end!" ) # TODO do we get this far?
1740 unless $last->id eq $end->id;
1745 =head2 next_reading( $reading, $sigil );
1747 Returns the reading that follows the given reading along the given witness
1753 # Return the successor via the corresponding path.
1755 my $answer = $self->_find_linked_reading( 'next', @_ );
1756 return undef unless $answer;
1757 return $self->reading( $answer );
1760 =head2 prior_reading( $reading, $sigil )
1762 Returns the reading that precedes the given reading along the given witness
1768 # Return the predecessor via the corresponding path.
1770 my $answer = $self->_find_linked_reading( 'prior', @_ );
1771 return $self->reading( $answer );
1774 sub _find_linked_reading {
1775 my( $self, $direction, $node, $path ) = @_;
1777 # Get a backup if we are dealing with a layered witness
1779 my $aclabel = $self->ac_label;
1780 if( $path && $path =~ /^(.*)\Q$aclabel\E$/ ) {
1784 my @linked_paths = $direction eq 'next'
1785 ? $self->sequence->edges_from( $node )
1786 : $self->sequence->edges_to( $node );
1787 return undef unless scalar( @linked_paths );
1789 # We have to find the linked path that contains all of the
1790 # witnesses supplied in $path.
1791 my( @path_wits, @alt_path_wits );
1792 @path_wits = sort( $self->_witnesses_of_label( $path ) ) if $path;
1793 @alt_path_wits = sort( $self->_witnesses_of_label( $alt_path ) ) if $alt_path;
1796 foreach my $le ( @linked_paths ) {
1797 if( $self->sequence->has_edge_attribute( @$le, $self->baselabel ) ) {
1800 my @le_wits = sort $self->path_witnesses( $le );
1801 if( _is_within( \@path_wits, \@le_wits ) ) {
1802 # This is the right path.
1803 return $direction eq 'next' ? $le->[1] : $le->[0];
1804 } elsif( _is_within( \@alt_path_wits, \@le_wits ) ) {
1808 # Got this far? Return the alternate path if it exists.
1809 return $direction eq 'next' ? $alt_le->[1] : $alt_le->[0]
1812 # Got this far? Return the base path if it exists.
1813 return $direction eq 'next' ? $base_le->[1] : $base_le->[0]
1816 # Got this far? We have no appropriate path.
1817 warn "Could not find $direction node from " . $node->id
1818 . " along path $path";
1824 my( $set1, $set2 ) = @_;
1825 my $ret = @$set1; # will be 0, i.e. false, if set1 is empty
1826 foreach my $el ( @$set1 ) {
1827 $ret = 0 unless grep { /^\Q$el\E$/ } @$set2;
1832 # Return the string that joins together a list of witnesses for
1833 # display on a single path.
1834 sub _witnesses_of_label {
1835 my( $self, $label ) = @_;
1836 my $regex = $self->wit_list_separator;
1837 my @answer = split( /\Q$regex\E/, $label );
1841 =head2 common_readings
1843 Returns the list of common readings in the graph (i.e. those readings that are
1844 shared by all non-lacunose witnesses.)
1848 sub common_readings {
1850 my @common = grep { $_->is_common } $self->readings;
1854 =head2 path_text( $sigil, [, $start, $end ] )
1856 Returns the text of a witness (plus its backup, if we are using a layer)
1857 as stored in the collation. The text is returned as a string, where the
1858 individual readings are joined with spaces and the meta-readings (e.g.
1859 lacunae) are omitted. Optional specification of $start and $end allows
1860 the generation of a subset of the witness text.
1865 my( $self, $wit, $start, $end ) = @_;
1866 $start = $self->start unless $start;
1867 $end = $self->end unless $end;
1868 my @path = grep { !$_->is_meta } $self->reading_sequence( $start, $end, $wit );
1871 foreach my $r ( @path ) {
1872 unless ( $r->join_prior || !$last || $last->join_next ) {
1875 $pathtext .= $r->text;
1881 =head1 INITIALIZATION METHODS
1883 These are mostly for use by parsers.
1885 =head2 make_witness_path( $witness )
1887 Link the array of readings contained in $witness->path (and in
1888 $witness->uncorrected_path if it exists) into collation paths.
1889 Clear out the arrays when finished.
1891 =head2 make_witness_paths
1893 Call make_witness_path for all witnesses in the tradition.
1897 # For use when a collation is constructed from a base text and an apparatus.
1898 # We have the sequences of readings and just need to add path edges.
1899 # When we are done, clear out the witness path attributes, as they are no
1901 # TODO Find a way to replace the witness path attributes with encapsulated functions?
1903 sub make_witness_paths {
1905 foreach my $wit ( $self->tradition->witnesses ) {
1906 # say STDERR "Making path for " . $wit->sigil;
1907 $self->make_witness_path( $wit );
1911 sub make_witness_path {
1912 my( $self, $wit ) = @_;
1913 my @chain = @{$wit->path};
1914 my $sig = $wit->sigil;
1915 # Add start and end if necessary
1916 unshift( @chain, $self->start ) unless $chain[0] eq $self->start;
1917 push( @chain, $self->end ) unless $chain[-1] eq $self->end;
1918 foreach my $idx ( 0 .. $#chain-1 ) {
1919 $self->add_path( $chain[$idx], $chain[$idx+1], $sig );
1921 if( $wit->is_layered ) {
1922 @chain = @{$wit->uncorrected_path};
1923 unshift( @chain, $self->start ) unless $chain[0] eq $self->start;
1924 push( @chain, $self->end ) unless $chain[-1] eq $self->end;
1925 foreach my $idx( 0 .. $#chain-1 ) {
1926 my $source = $chain[$idx];
1927 my $target = $chain[$idx+1];
1928 $self->add_path( $source, $target, $sig.$self->ac_label )
1929 unless $self->has_path( $source, $target, $sig );
1933 $wit->clear_uncorrected_path;
1936 =head2 calculate_ranks
1938 Calculate the reading ranks (that is, their aligned positions relative
1939 to each other) for the graph. This can only be called on linear collations.
1943 use Text::Tradition;
1945 my $cxfile = 't/data/Collatex-16.xml';
1946 my $t = Text::Tradition->new(
1948 'input' => 'CollateX',
1951 my $c = $t->collation;
1954 my $table = $c->alignment_table;
1955 ok( $c->has_cached_table, "Alignment table was cached" );
1956 is( $c->alignment_table, $table, "Cached table returned upon second call" );
1957 $c->calculate_ranks;
1958 is( $c->alignment_table, $table, "Cached table retained with no rank change" );
1959 $c->add_relationship( 'n13', 'n23', { type => 'repetition' } );
1960 is( $c->alignment_table, $table, "Alignment table unchanged after non-colo relationship add" );
1961 $c->add_relationship( 'n24', 'n23', { type => 'spelling' } );
1962 isnt( $c->alignment_table, $table, "Alignment table changed after colo relationship add" );
1968 sub calculate_ranks {
1970 # Save the existing ranks, in case we need to invalidate the cached SVG.
1971 throw( "Cannot calculate ranks on a non-linear graph" )
1972 unless $self->linear;
1974 map { $existing_ranks{$_} = $_->rank } $self->readings;
1976 # Do the rankings based on the relationship equivalence graph, starting
1977 # with the start node.
1978 my ( $node_ranks, $rank_nodes ) = $self->relations->equivalence_ranks();
1980 # Transfer our rankings from the topological graph to the real one.
1981 foreach my $r ( $self->readings ) {
1982 if( defined $node_ranks->{$self->equivalence( $r->id )} ) {
1983 $r->rank( $node_ranks->{$self->equivalence( $r->id )} );
1985 # Die. Find the last rank we calculated.
1986 my @all_defined = sort { ( $node_ranks->{$self->equivalence( $a->id )}||-1 )
1987 <=> ( $node_ranks->{$self->equivalence( $b->id )}||-1 ) }
1989 my $last = pop @all_defined;
1990 throw( "Ranks not calculated after $last - do you have a cycle in the graph?" );
1993 # Do we need to invalidate the cached data?
1994 if( $self->has_cached_table ) {
1995 foreach my $r ( $self->readings ) {
1996 next if defined( $existing_ranks{$r} )
1997 && $existing_ranks{$r} == $r->rank;
1998 # Something has changed, so clear the cache
1999 $self->_clear_cache;
2000 # ...and recalculate the common readings.
2001 $self->calculate_common_readings();
2005 # The graph calculation information is now up to date.
2006 $self->_graphcalc_done(1);
2011 $self->wipe_table if $self->has_cached_table;
2015 =head2 flatten_ranks
2017 A convenience method for parsing collation data. Searches the graph for readings
2018 with the same text at the same rank, and merges any that are found.
2023 my ( $self, %args ) = shift;
2024 my %unique_rank_rdg;
2026 foreach my $p ( $self->identical_readings( %args ) ) {
2027 # say STDERR "Combining readings at same rank: @$p";
2029 $self->merge_readings( @$p );
2030 # TODO see if this now makes a common point.
2032 # If we merged readings, the ranks are still fine but the alignment
2033 # table is wrong. Wipe it.
2034 $self->wipe_table() if $changed;
2037 =head2 identical_readings
2038 =head2 identical_readings( start => $startnode, end => $endnode )
2039 =head2 identical_readings( startrank => $startrank, endrank => $endrank )
2041 Goes through the graph identifying all pairs of readings that appear to be
2042 identical, and therefore able to be merged into a single reading. Returns the
2043 relevant identical pairs. Can be restricted to run over only a part of the
2044 graph, specified either by node or by rank.
2048 sub identical_readings {
2049 my ( $self, %args ) = @_;
2050 # Find where we should start and end.
2051 my $startrank = $args{startrank} || 0;
2052 if( $args{start} ) {
2053 throw( "Starting reading has no rank" ) unless $self->reading( $args{start} )
2054 && $self->reading( $args{start} )->has_rank;
2055 $startrank = $self->reading( $args{start} )->rank;
2057 my $endrank = $args{endrank} || $self->end->rank;
2059 throw( "Ending reading has no rank" ) unless $self->reading( $args{end} )
2060 && $self->reading( $args{end} )->has_rank;
2061 $endrank = $self->reading( $args{end} )->rank;
2064 # Make sure the ranks are correct.
2065 unless( $self->_graphcalc_done ) {
2066 $self->calculate_ranks;
2068 # Go through the readings looking for duplicates.
2069 my %unique_rank_rdg;
2071 foreach my $rdg ( $self->readings ) {
2072 next unless $rdg->has_rank;
2073 my $rk = $rdg->rank;
2074 next if $rk > $endrank || $rk < $startrank;
2075 my $key = $rk . "||" . $rdg->text;
2076 if( exists $unique_rank_rdg{$key} ) {
2077 # Make sure they don't have different grammatical forms
2078 my $ur = $unique_rank_rdg{$key};
2079 if( $rdg->is_identical( $ur ) ) {
2080 push( @pairs, [ $ur, $rdg ] );
2083 $unique_rank_rdg{$key} = $rdg;
2091 =head2 calculate_common_readings
2093 Goes through the graph identifying the readings that appear in every witness
2094 (apart from those with lacunae at that spot.) Marks them as common and returns
2099 use Text::Tradition;
2101 my $cxfile = 't/data/Collatex-16.xml';
2102 my $t = Text::Tradition->new(
2104 'input' => 'CollateX',
2107 my $c = $t->collation;
2109 my @common = $c->calculate_common_readings();
2110 is( scalar @common, 8, "Found correct number of common readings" );
2111 my @marked = sort $c->common_readings();
2112 is( scalar @common, 8, "All common readings got marked as such" );
2113 my @expected = qw/ n1 n11 n16 n19 n20 n5 n6 n7 /;
2114 is_deeply( \@marked, \@expected, "Found correct list of common readings" );
2120 sub calculate_common_readings {
2123 map { $_->is_common( 0 ) } $self->readings;
2124 # Implicitly calls calculate_ranks
2125 my $table = $self->alignment_table;
2126 foreach my $idx ( 0 .. $table->{'length'} - 1 ) {
2127 my @row = map { $_->{'tokens'}->[$idx]
2128 ? $_->{'tokens'}->[$idx]->{'t'} : '' }
2129 @{$table->{'alignment'}};
2131 foreach my $r ( @row ) {
2133 $hash{$r->id} = $r unless $r->is_meta;
2135 $hash{'UNDEF'} = $r;
2138 if( keys %hash == 1 && !exists $hash{'UNDEF'} ) {
2139 my( $r ) = values %hash;
2141 push( @common, $r );
2147 =head2 text_from_paths
2149 Calculate the text array for all witnesses from the path, for later consistency
2150 checking. Only to be used if there is no non-graph-based way to know the
2155 sub text_from_paths {
2157 foreach my $wit ( $self->tradition->witnesses ) {
2158 my @readings = $self->reading_sequence( $self->start, $self->end, $wit->sigil );
2160 foreach my $r ( @readings ) {
2161 next if $r->is_meta;
2162 push( @text, $r->text );
2164 $wit->text( \@text );
2165 if( $wit->is_layered ) {
2166 my @ucrdgs = $self->reading_sequence( $self->start, $self->end,
2167 $wit->sigil.$self->ac_label );
2169 foreach my $r ( @ucrdgs ) {
2170 next if $r->is_meta;
2171 push( @uctext, $r->text );
2173 $wit->layertext( \@uctext );
2178 =head1 UTILITY FUNCTIONS
2180 =head2 common_predecessor( $reading_a, $reading_b )
2182 Find the last reading that occurs in sequence before both the given readings.
2183 At the very least this should be $self->start.
2185 =head2 common_successor( $reading_a, $reading_b )
2187 Find the first reading that occurs in sequence after both the given readings.
2188 At the very least this should be $self->end.
2192 use Text::Tradition;
2194 my $cxfile = 't/data/Collatex-16.xml';
2195 my $t = Text::Tradition->new(
2197 'input' => 'CollateX',
2200 my $c = $t->collation;
2202 is( $c->common_predecessor( 'n24', 'n23' )->id,
2203 'n20', "Found correct common predecessor" );
2204 is( $c->common_successor( 'n24', 'n23' )->id,
2205 '__END__', "Found correct common successor" );
2207 is( $c->common_predecessor( 'n19', 'n17' )->id,
2208 'n16', "Found correct common predecessor for readings on same path" );
2209 is( $c->common_successor( 'n21', 'n10' )->id,
2210 '__END__', "Found correct common successor for readings on same path" );
2216 ## Return the closest reading that is a predecessor of both the given readings.
2217 sub common_predecessor {
2219 my( $r1, $r2 ) = $self->_objectify_args( @_ );
2220 return $self->_common_in_path( $r1, $r2, 'predecessors' );
2223 sub common_successor {
2225 my( $r1, $r2 ) = $self->_objectify_args( @_ );
2226 return $self->_common_in_path( $r1, $r2, 'successors' );
2230 # TODO think about how to do this without ranks...
2231 sub _common_in_path {
2232 my( $self, $r1, $r2, $dir ) = @_;
2233 my $iter = $self->end->rank;
2235 my @last_r1 = ( $r1 );
2236 my @last_r2 = ( $r2 );
2237 # my %all_seen = ( $r1 => 'r1', $r2 => 'r2' );
2239 # say STDERR "Finding common $dir for $r1, $r2";
2240 while( !@candidates ) {
2241 last unless $iter--; # Avoid looping infinitely
2242 # Iterate separately down the graph from r1 and r2
2243 my( @new_lc1, @new_lc2 );
2244 foreach my $lc ( @last_r1 ) {
2245 foreach my $p ( $lc->$dir ) {
2246 if( $all_seen{$p->id} && $all_seen{$p->id} ne 'r1' ) {
2247 # say STDERR "Path candidate $p from $lc";
2248 push( @candidates, $p );
2249 } elsif( !$all_seen{$p->id} ) {
2250 $all_seen{$p->id} = 'r1';
2251 push( @new_lc1, $p );
2255 foreach my $lc ( @last_r2 ) {
2256 foreach my $p ( $lc->$dir ) {
2257 if( $all_seen{$p->id} && $all_seen{$p->id} ne 'r2' ) {
2258 # say STDERR "Path candidate $p from $lc";
2259 push( @candidates, $p );
2260 } elsif( !$all_seen{$p->id} ) {
2261 $all_seen{$p->id} = 'r2';
2262 push( @new_lc2, $p );
2266 @last_r1 = @new_lc1;
2267 @last_r2 = @new_lc2;
2269 my @answer = sort { $a->rank <=> $b->rank } @candidates;
2270 return $dir eq 'predecessors' ? pop( @answer ) : shift ( @answer );
2274 Text::Tradition::Error->throw(
2275 'ident' => 'Collation error',
2281 __PACKAGE__->meta->make_immutable;
2287 =item * Rework XML serialization in a more modular way
2293 This package is free software and is provided "as is" without express
2294 or implied warranty. You can redistribute it and/or modify it under
2295 the same terms as Perl itself.
2299 Tara L Andrews E<lt>aurum@cpan.orgE<gt>