1 package Text::Tradition::Collation;
4 use Encode qw( decode_utf8 );
8 use IPC::Run qw( run binary );
9 use JSON qw/ to_json /;
11 use Text::Tradition::Collation::Data;
12 use Text::Tradition::Collation::Reading;
13 use Text::Tradition::Collation::RelationshipStore;
14 use Text::Tradition::Error;
15 use XML::Easy::Syntax qw( $xml10_namestartchar_rx $xml10_namechar_rx );
17 use XML::LibXML::XPathContext;
21 isa => 'Text::Tradition::Collation::Data',
60 isa => 'Text::Tradition',
61 writer => '_set_tradition',
69 Text::Tradition::Collation - a software model for a text collation
74 my $t = Text::Tradition->new(
75 'name' => 'this is a text',
77 'file' => '/path/to/tei_parallel_seg_file.xml' );
79 my $c = $t->collation;
80 my @readings = $c->readings;
81 my @paths = $c->paths;
82 my @relationships = $c->relationships;
84 my $svg_variant_graph = $t->collation->as_svg();
88 Text::Tradition is a library for representation and analysis of collated
89 texts, particularly medieval ones. The Collation is the central feature of
90 a Tradition, where the text, its sequence of readings, and its relationships
91 between readings are actually kept.
97 The constructor. Takes a hash or hashref of the following arguments:
101 =item * tradition - The Text::Tradition object to which the collation
104 =item * linear - Whether the collation should be linear; that is, whether
105 transposed readings should be treated as two linked readings rather than one,
106 and therefore whether the collation graph is acyclic. Defaults to true.
108 =item * baselabel - The default label for the path taken by a base text
109 (if any). Defaults to 'base text'.
111 =item * wit_list_separator - The string to join a list of witnesses for
112 purposes of making labels in display graphs. Defaults to ', '.
114 =item * ac_label - The extra label to tack onto a witness sigil when
115 representing another layer of path for the given witness - that is, when
116 a text has more than one possible reading due to scribal corrections or
117 the like. Defaults to ' (a.c.)'.
119 =item * wordsep - The string used to separate words in the original text.
130 =head2 wit_list_separator
138 Simple accessors for collation attributes.
142 The meta-reading at the start of every witness path.
146 The meta-reading at the end of every witness path.
150 Returns all Reading objects in the graph.
152 =head2 reading( $id )
154 Returns the Reading object corresponding to the given ID.
156 =head2 add_reading( $reading_args )
158 Adds a new reading object to the collation.
159 See L<Text::Tradition::Collation::Reading> for the available arguments.
161 =head2 del_reading( $object_or_id )
163 Removes the given reading from the collation, implicitly removing its
164 paths and relationships.
166 =head2 has_reading( $id )
168 Predicate to see whether a given reading ID is in the graph.
170 =head2 reading_witnesses( $object_or_id )
172 Returns a list of sigils whose witnesses contain the reading.
176 Returns all reading paths within the document - that is, all edges in the
177 collation graph. Each path is an arrayref of [ $source, $target ] reading IDs.
179 =head2 add_path( $source, $target, $sigil )
181 Links the given readings in the collation in sequence, under the given witness
182 sigil. The readings may be specified by object or ID.
184 =head2 del_path( $source, $target, $sigil )
186 Links the given readings in the collation in sequence, under the given witness
187 sigil. The readings may be specified by object or ID.
189 =head2 has_path( $source, $target );
191 Returns true if the two readings are linked in sequence in any witness.
192 The readings may be specified by object or ID.
196 Returns all Relationship objects in the collation.
198 =head2 add_relationship( $reading, $other_reading, $options )
200 Adds a new relationship of the type given in $options between the two readings,
201 which may be specified by object or ID. Returns a value of ( $status, @vectors)
202 where $status is true on success, and @vectors is a list of relationship edges
203 that were ultimately added.
204 See L<Text::Tradition::Collation::Relationship> for the available options.
209 my ( $class, @args ) = @_;
210 my %args = @args == 1 ? %{ $args[0] } : @args;
211 # TODO determine these from the Moose::Meta object
212 my @delegate_attrs = qw(sequence relations readings wit_list_separator baselabel
213 linear wordsep start end cached_table _graphcalc_done);
215 for my $attr (@delegate_attrs) {
216 $data_args{$attr} = delete $args{$attr} if exists $args{$attr};
218 $args{_data} = Text::Tradition::Collation::Data->new(%data_args);
224 $self->_set_relations( Text::Tradition::Collation::RelationshipStore->new( 'collation' => $self ) );
225 $self->_set_start( $self->add_reading(
226 { 'collation' => $self, 'is_start' => 1, 'init' => 1 } ) );
227 $self->_set_end( $self->add_reading(
228 { 'collation' => $self, 'is_end' => 1, 'init' => 1 } ) );
231 =head2 register_relationship_type( %relationship_definition )
233 Add a relationship type definition to this collation. The argument can be either a
234 hash or a hashref, defining the properties of the relationship. For relationship types
235 and their properties, see L<Text::Tradition::Collation::RelationshipType>.
237 =head2 get_relationship_type( $relationship_name )
239 Retrieve the RelationshipType object for the relationship with the given name.
243 sub register_relationship_type {
245 my %args = @_ == 1 ? %{$_[0]} : @_;
246 if( $self->relations->has_type( $args{name} ) ) {
247 throw( 'Relationship type ' . $args{name} . ' already registered' );
249 $self->relations->add_type( %args );
252 sub get_relationship_type {
253 my( $self, $name ) = @_;
254 return $self->relations->has_type( $name )
255 ? $self->relations->type( $name ) : undef;
258 ### Reading construct/destruct functions
261 my( $self, $reading ) = @_;
262 unless( ref( $reading ) eq 'Text::Tradition::Collation::Reading' ) {
263 my %args = %$reading;
264 if( $args{'init'} ) {
265 # If we are initializing an empty collation, don't assume that we
266 # have set a tradition.
267 delete $args{'init'};
268 } elsif( $self->tradition->can('language') && $self->tradition->has_language
269 && !exists $args{'language'} ) {
270 $args{'language'} = $self->tradition->language;
272 $reading = Text::Tradition::Collation::Reading->new(
273 'collation' => $self,
276 # First check to see if a reading with this ID exists.
277 if( $self->reading( $reading->id ) ) {
278 throw( "Collation already has a reading with id " . $reading->id );
280 $self->_graphcalc_done(0);
281 $self->_add_reading( $reading->id => $reading );
282 # Once the reading has been added, put it in both graphs.
283 $self->sequence->add_vertex( $reading->id );
284 $self->relations->add_reading( $reading->id );
288 around del_reading => sub {
293 if( ref( $arg ) eq 'Text::Tradition::Collation::Reading' ) {
296 # Remove the reading from the graphs.
297 $self->_graphcalc_done(0);
298 $self->_clear_cache; # Explicitly clear caches to GC the reading
299 $self->sequence->delete_vertex( $arg );
300 $self->relations->delete_reading( $arg );
303 $self->$orig( $arg );
306 =head2 merge_readings( $main, $second, $concatenate, $with_str )
308 Merges the $second reading into the $main one. If $concatenate is true, then
309 the merged node will carry the text of both readings, concatenated with either
310 $with_str (if specified) or a sensible default (the empty string if the
311 appropriate 'join_*' flag is set on either reading, or else $self->wordsep.)
313 The first two arguments may be either readings or reading IDs.
320 my $cxfile = 't/data/Collatex-16.xml';
321 my $t = Text::Tradition->new(
323 'input' => 'CollateX',
326 my $c = $t->collation;
328 my $rno = scalar $c->readings;
329 # Split n21 ('unto') for testing purposes
330 my $new_r = $c->add_reading( { 'id' => 'n21p0', 'text' => 'un', 'join_next' => 1 } );
331 my $old_r = $c->reading( 'n21' );
332 $old_r->alter_text( 'to' );
333 $c->del_path( 'n20', 'n21', 'A' );
334 $c->add_path( 'n20', 'n21p0', 'A' );
335 $c->add_path( 'n21p0', 'n21', 'A' );
336 $c->add_relationship( 'n21', 'n22', { type => 'collated', scope => 'local' } );
338 ok( $c->reading( 'n21p0' ), "New reading exists" );
339 is( scalar $c->readings, $rno, "Reading add offset by flatten_ranks" );
341 # Combine n3 and n4 ( with his )
342 $c->merge_readings( 'n3', 'n4', 1 );
343 ok( !$c->reading('n4'), "Reading n4 is gone" );
344 is( $c->reading('n3')->text, 'with his', "Reading n3 has both words" );
346 # Collapse n9 and n10 ( rood / root )
347 $c->merge_readings( 'n9', 'n10' );
348 ok( !$c->reading('n10'), "Reading n10 is gone" );
349 is( $c->reading('n9')->text, 'rood', "Reading n9 has an unchanged word" );
351 # Try to combine n21 and n21p0. This should break.
352 my $remaining = $c->reading('n21');
353 $remaining ||= $c->reading('n22'); # one of these should still exist
355 $c->merge_readings( 'n21p0', $remaining, 1 );
356 ok( 0, "Bad reading merge changed the graph" );
357 } catch( Text::Tradition::Error $e ) {
358 like( $e->message, qr/neither concatenated nor collated/, "Expected exception from bad concatenation" );
360 ok( 0, "Unexpected error on bad reading merge: $@" );
364 $c->calculate_ranks();
365 ok( 1, "Graph is still evidently whole" );
366 } catch( Text::Tradition::Error $e ) {
367 ok( 0, "Caught a rank exception: " . $e->message );
378 my( $kept_obj, $del_obj, $combine, $combine_char ) = $self->_objectify_args( @_ );
379 my $mergemeta = $kept_obj->is_meta;
380 throw( "Cannot merge meta and non-meta reading" )
381 unless ( $mergemeta && $del_obj->is_meta )
382 || ( !$mergemeta && !$del_obj->is_meta );
384 throw( "Cannot merge with start or end node" )
385 if( $kept_obj eq $self->start || $kept_obj eq $self->end
386 || $del_obj eq $self->start || $del_obj eq $self->end );
387 throw( "Cannot combine text of meta readings" ) if $combine;
389 # We can only merge readings in a linear graph if:
390 # - they are contiguous with only one edge between them, OR
391 # - they are at equivalent ranks in the graph.
392 if( $self->linear ) {
393 my @delpred = $del_obj->predecessors;
394 my @keptsuc = $kept_obj->successors;
395 unless ( @delpred == 1 && $delpred[0] eq $kept_obj
396 && @keptsuc == 1 && $keptsuc[0] eq $del_obj ) {
397 my( $is_ok, $msg ) = $self->relations->relationship_valid(
398 $kept_obj, $del_obj, 'collated' );
400 throw( "Readings $kept_obj and $del_obj can be neither concatenated nor collated" );
405 # We only need the IDs for adding paths to the graph, not the reading
406 # objects themselves.
407 my $kept = $kept_obj->id;
408 my $deleted = $del_obj->id;
409 $self->_graphcalc_done(0);
411 # The kept reading should inherit the paths and the relationships
412 # of the deleted reading.
413 foreach my $path ( $self->sequence->edges_at( $deleted ) ) {
414 my @vector = ( $kept );
415 push( @vector, $path->[1] ) if $path->[0] eq $deleted;
416 unshift( @vector, $path->[0] ) if $path->[1] eq $deleted;
417 next if $vector[0] eq $vector[1]; # Don't add a self loop
418 my %wits = %{$self->sequence->get_edge_attributes( @$path )};
419 $self->sequence->add_edge( @vector );
420 my $fwits = $self->sequence->get_edge_attributes( @vector );
421 @wits{keys %$fwits} = values %$fwits;
422 $self->sequence->set_edge_attributes( @vector, \%wits );
424 $self->relations->merge_readings( $kept, $deleted, $combine );
426 # Do the deletion deed.
428 # Combine the text of the readings
429 my $joinstr = $combine_char;
430 unless( defined $joinstr ) {
431 $joinstr = '' if $kept_obj->join_next || $del_obj->join_prior;
432 $joinstr = $self->wordsep unless defined $joinstr;
434 $kept_obj->_combine( $del_obj, $joinstr );
436 $self->del_reading( $deleted );
439 =head2 merge_related( @relationship_types )
441 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.
443 WARNING: This operation cannot be undone.
455 $t = Text::Tradition->new( 'input' => 'Self', 'file' => 't/data/legendfrag.xml' );
456 } [qr/Cannot set relationship on a meta reading/],
457 "Got expected relationship drop warning on parse";
459 my $c = $t->collation;
460 # Force the transitive propagation of all existing relationships.
461 $c->relations->propagate_all_relationships();
464 map { $rdg_ids{$_} = 1 } $c->readings;
465 $c->merge_related( 'orthographic' );
466 is( scalar( $c->readings ), keys( %rdg_ids ) - 9,
467 "Successfully collapsed orthographic variation" );
468 map { $rdg_ids{$_} = undef } qw/ r13.3 r11.4 r8.5 r8.2 r7.7 r7.5 r7.4 r7.3 r7.1 /;
469 foreach my $rid ( keys %rdg_ids ) {
470 my $exp = $rdg_ids{$rid};
471 is( !$c->reading( $rid ), !$exp, "Reading $rid correctly " .
472 ( $exp ? "retained" : "removed" ) );
474 ok( $c->linear, "Graph is still linear" );
476 $c->calculate_ranks; # This should succeed
477 ok( 1, "Can still calculate ranks on the new graph" );
479 ok( 0, "Rank calculation on merged graph failed: $@" );
482 # Now add some transpositions
483 $c->add_relationship( 'r8.4', 'r10.4', { type => 'transposition' } );
484 $c->merge_related( 'transposition' );
485 is( scalar( $c->readings ), keys( %rdg_ids ) - 10,
486 "Transposed relationship is merged away" );
487 ok( !$c->reading('r8.4'), "Correct transposed reading removed" );
488 ok( !$c->linear, "Graph is no longer linear" );
490 $c->calculate_ranks; # This should fail
491 ok( 0, "Rank calculation happened on nonlinear graph?!" );
492 } catch ( Text::Tradition::Error $e ) {
493 is( $e->message, 'Cannot calculate ranks on a non-linear graph',
494 "Rank calculation on merged graph threw an error" );
501 # TODO: there should be a way to display merged without affecting the underlying data!
506 map { $reltypehash{$_} = 1 } @_;
508 # Set up the filter for finding related readings
510 exists $reltypehash{$_[0]->type};
513 # Go through all readings looking for related ones
514 foreach my $r ( $self->readings ) {
515 next unless $self->reading( "$r" ); # might have been deleted meanwhile
516 while( my @related = $self->related_readings( $r, $filter ) ) {
517 push( @related, $r );
519 scalar $b->witnesses <=> scalar $a->witnesses
521 my $keep = shift @related;
522 foreach my $delr ( @related ) {
524 unless( $self->get_relationship( $keep, $delr )->colocated );
525 $self->merge_readings( $keep, $delr );
531 =head2 compress_readings
533 Where possible in the graph, compresses plain sequences of readings into a
534 single reading. The sequences must consist of readings with no
535 relationships to other readings, with only a single witness path between
536 them and no other witness paths from either that would skip the other. The
537 readings must also not be marked as nonsense or bad grammar.
539 WARNING: This operation cannot be undone.
545 my $t = Text::Tradition->new( input => 'CollateX', file => 't/data/Collatex-16.xml' );
546 my $c = $t->collation;
547 my $n = scalar $c->readings;
548 $c->compress_readings();
549 is( scalar $c->readings, $n - 6, "Compressing readings seems to work" );
551 # Now put in a join-word and make sure the thing still works.
552 my $t2 = Text::Tradition->new( input => 'CollateX', file => 't/data/Collatex-16.xml' );
553 my $c2 = $t2->collation;
554 # Split n21 ('unto') for testing purposes
555 my $new_r = $c2->add_reading( { 'id' => 'n21p0', 'text' => 'un', 'join_next' => 1 } );
556 my $old_r = $c2->reading( 'n21' );
557 $old_r->alter_text( 'to' );
558 $c2->del_path( 'n20', 'n21', 'A' );
559 $c2->add_path( 'n20', 'n21p0', 'A' );
560 $c2->add_path( 'n21p0', 'n21', 'A' );
561 $c2->calculate_ranks();
562 is( scalar $c2->readings, $n + 1, "We have our extra test reading" );
563 $c2->compress_readings();
564 is( scalar $c2->readings, $n - 6, "Compressing readings also works with join_next" );
565 is( $c2->reading( 'n21p0' )->text, 'unto', "The joined word has no space" );
572 sub compress_readings {
574 # Sanity check: first save the original text of each witness.
576 foreach my $wit ( $self->tradition->witnesses ) {
577 $origtext{$wit->sigil} = $self->path_text( $wit->sigil );
578 if( $wit->is_layered ) {
579 my $acsig = $wit->sigil . $self->ac_label;
580 $origtext{$acsig} = $self->path_text( $acsig );
585 # Anywhere in the graph that there is a reading that joins only to a single
586 # successor, and neither of these have any relationships, just join the two
588 foreach my $rdg ( sort { $a->rank <=> $b->rank } $self->readings ) {
589 # Now look for readings that can be joined to their successors.
590 next unless $rdg->is_combinable;
592 while( $self->sequence->successors( $rdg ) == 1 ) {
593 my( $next ) = $self->reading( $self->sequence->successors( $rdg ) );
594 throw( "Infinite loop" ) if $seen{$next->id};
595 $seen{$next->id} = 1;
596 last if $self->sequence->predecessors( $next ) > 1;
597 last unless $next->is_combinable;
598 say "Joining readings $rdg and $next";
599 $self->merge_readings( $rdg, $next, 1 );
603 # Finally, make sure we haven't screwed anything up.
604 foreach my $wit ( $self->tradition->witnesses ) {
605 my $pathtext = $self->path_text( $wit->sigil );
606 throw( "Text differs for witness " . $wit->sigil )
607 unless $pathtext eq $origtext{$wit->sigil};
608 if( $wit->is_layered ) {
609 my $acsig = $wit->sigil . $self->ac_label;
610 $pathtext = $self->path_text( $acsig );
611 throw( "Layered text differs for witness " . $wit->sigil )
612 unless $pathtext eq $origtext{$acsig};
616 $self->relations->rebuild_equivalence();
617 $self->calculate_ranks();
620 # Helper function for manipulating the graph.
621 sub _stringify_args {
622 my( $self, $first, $second, @args ) = @_;
624 if ref( $first ) eq 'Text::Tradition::Collation::Reading';
625 $second = $second->id
626 if ref( $second ) eq 'Text::Tradition::Collation::Reading';
627 return( $first, $second, @args );
630 # Helper function for manipulating the graph.
631 sub _objectify_args {
632 my( $self, $first, $second, $arg ) = @_;
633 $first = $self->reading( $first )
634 unless ref( $first ) eq 'Text::Tradition::Collation::Reading';
635 $second = $self->reading( $second )
636 unless ref( $second ) eq 'Text::Tradition::Collation::Reading';
637 return( $first, $second, $arg );
640 =head2 duplicate_reading( $reading, @witlist )
642 Split the given reading into two, so that the new reading is in the path for
643 the witnesses given in @witlist. If the result is that certain non-colocated
644 relationships (e.g. transpositions) are no longer valid, these will be removed.
645 Returns the newly-created reading.
649 use Test::More::UTF8;
653 my $st = Text::Tradition->new( 'input' => 'Self', 'file' => 't/data/collatecorr.xml' );
654 is( ref( $st ), 'Text::Tradition', "Got a tradition from test file" );
655 ok( $st->has_witness('Ba96'), "Tradition has the affected witness" );
657 my $sc = $st->collation;
659 ok( $sc->reading('n131'), "Tradition has the affected reading" );
660 is( scalar( $sc->readings ), $numr, "There are $numr readings in the graph" );
661 is( $sc->end->rank, 14, "There are fourteen ranks in the graph" );
663 # Detach the erroneously collated reading
664 my( $newr, @del_rdgs ) = $sc->duplicate_reading( 'n131', 'Ba96' );
665 ok( $newr, "New reading was created" );
666 ok( $sc->reading('n131_0'), "Detached the bad collation with a new reading" );
667 is( scalar( $sc->readings ), $numr + 1, "A reading was added to the graph" );
668 is( $sc->end->rank, 10, "There are now only ten ranks in the graph" );
669 my $csucc = $sc->common_successor( 'n131', 'n131_0' );
670 is( $csucc->id, 'n136', "Found correct common successor to duped reading" );
672 # Check that the bad transposition is gone
673 is( scalar @del_rdgs, 1, "Deleted reading was returned by API call" );
674 is( $sc->get_relationship( 'n130', 'n135' ), undef, "Bad transposition relationship is gone" );
676 # The collation should not be fixed
677 my @pairs = $sc->identical_readings();
678 is( scalar @pairs, 0, "Not re-collated yet" );
680 ok( $sc->merge_readings( 'n124', 'n131_0' ), "Collated the readings correctly" );
681 @pairs = $sc->identical_readings( start => 'n124', end => $csucc->id );
682 is( scalar @pairs, 3, "Found three more identical readings" );
683 is( $sc->end->rank, 11, "The ranks shifted appropriately" );
684 $sc->flatten_ranks();
685 is( scalar( $sc->readings ), $numr - 3, "Now we are collated correctly" );
687 # Check that we can't "duplicate" a reading with no wits or with all wits
689 my( $badr, @del_rdgs ) = $sc->duplicate_reading( 'n124' );
690 ok( 0, "Reading duplication without witnesses throws an error" );
691 } catch( Text::Tradition::Error $e ) {
692 like( $e->message, qr/Must specify one or more witnesses/,
693 "Reading duplication without witnesses throws the expected error" );
695 ok( 0, "Reading duplication without witnesses threw the wrong error" );
699 my( $badr, @del_rdgs ) = $sc->duplicate_reading( 'n124', 'Ba96', 'Mü11475' );
700 ok( 0, "Reading duplication with all witnesses throws an error" );
701 } catch( Text::Tradition::Error $e ) {
702 like( $e->message, qr/Cannot join all witnesses/,
703 "Reading duplication with all witnesses throws the expected error" );
705 ok( 0, "Reading duplication with all witnesses threw the wrong error" );
709 $sc->calculate_ranks();
710 ok( 1, "Graph is still evidently whole" );
711 } catch( Text::Tradition::Error $e ) {
712 ok( 0, "Caught a rank exception: " . $e->message );
719 sub duplicate_reading {
720 my( $self, $r, @wits ) = @_;
721 # Check that we are not doing anything unwise.
722 throw( "Must specify one or more witnesses for the duplicated reading" )
724 unless( ref( $r ) eq 'Text::Tradition::Collation::Reading' ) {
725 $r = $self->reading( $r );
727 throw( "Cannot duplicate a meta-reading" )
729 throw( "Cannot join all witnesses to the new reading" )
730 if scalar( @wits ) == scalar( $r->witnesses );
732 # Get all the reading attributes and duplicate them.
733 my $rmeta = Text::Tradition::Collation::Reading->meta;
735 foreach my $attr( $rmeta->get_all_attributes ) {
736 next if $attr->name =~ /^_/;
737 my $acc = $attr->get_read_method;
738 if( !$acc && $attr->has_applied_traits ) {
739 my $tr = $attr->applied_traits;
740 if( $tr->[0] =~ /::(Array|Hash)$/ ) {
742 my %methods = reverse %{$attr->handles};
743 $acc = $methods{elements};
744 $args{$attr->name} = $which eq 'Array'
745 ? [ $r->$acc ] : { $r->$acc };
748 $args{$attr->name} = $r->$acc if $acc;
751 # By definition the new reading will no longer be common.
752 $args{is_common} = 0;
753 # The new reading also needs its own ID.
754 $args{id} = $self->_generate_dup_id( $r->id );
756 # Try to make the new reading.
757 my $newr = $self->add_reading( \%args );
758 # The old reading is also no longer common.
761 # For each of the witnesses, dissociate from the old reading and
762 # associate with the new.
763 foreach my $wit ( @wits ) {
764 my $prior = $self->prior_reading( $r, $wit );
765 my $next = $self->next_reading( $r, $wit );
766 $self->del_path( $prior, $r, $wit );
767 $self->add_path( $prior, $newr, $wit );
768 $self->del_path( $r, $next, $wit );
769 $self->add_path( $newr, $next, $wit );
772 # If the graph is ranked, we need to look for relationships that are now
773 # invalid (i.e. 'non-colocation' types that might now be colocated) and
774 # remove them. If not, we can skip it.
777 my @deleted_relations;
778 if( $self->end->has_rank ) {
779 # Find the point where we can stop checking
780 $succ = $self->common_successor( $r, $newr );
782 # Hash the existing ranks
783 foreach my $rdg ( $self->readings ) {
784 $rrk{$rdg->id} = $rdg->rank;
786 # Calculate the new ranks
787 $self->calculate_ranks();
789 # Check for invalid non-colocated relationships among changed-rank readings
790 # from where the ranks start changing up to $succ
791 my $lastrank = $succ->rank;
792 foreach my $rdg ( $self->readings ) {
793 next if $rdg->rank > $lastrank;
794 next if $rdg->rank == $rrk{$rdg->id};
795 my @noncolo = $rdg->related_readings( sub { !$_[0]->colocated } );
796 next unless @noncolo;
797 foreach my $nc ( @noncolo ) {
798 unless( $self->relations->verify_or_delete( $rdg, $nc ) ) {
799 push( @deleted_relations, [ $rdg->id, $nc->id ] );
804 return ( $newr, @deleted_relations );
807 sub _generate_dup_id {
808 my( $self, $rid ) = @_;
813 if( $self->has_reading( $newid ) ) {
826 # We only need the IDs for adding paths to the graph, not the reading
827 # objects themselves.
828 my( $source, $target, $wit ) = $self->_stringify_args( @_ );
830 $self->_graphcalc_done(0);
831 # Connect the readings
832 unless( $self->sequence->has_edge( $source, $target ) ) {
833 $self->sequence->add_edge( $source, $target );
834 $self->relations->add_equivalence_edge( $source, $target );
836 # Note the witness in question
837 $self->sequence->set_edge_attribute( $source, $target, $wit, 1 );
843 if( ref( $_[0] ) eq 'ARRAY' ) {
850 # We only need the IDs for removing paths from the graph, not the reading
851 # objects themselves.
852 my( $source, $target, $wit ) = $self->_stringify_args( @args );
854 $self->_graphcalc_done(0);
855 if( $self->sequence->has_edge_attribute( $source, $target, $wit ) ) {
856 $self->sequence->delete_edge_attribute( $source, $target, $wit );
858 unless( $self->sequence->has_edge_attributes( $source, $target ) ) {
859 $self->sequence->delete_edge( $source, $target );
860 $self->relations->delete_equivalence_edge( $source, $target );
865 # Extra graph-alike utility
868 my( $source, $target, $wit ) = $self->_stringify_args( @_ );
869 return undef unless $self->sequence->has_edge( $source, $target );
870 return $self->sequence->has_edge_attribute( $source, $target, $wit );
873 =head2 clear_witness( @sigil_list )
875 Clear the given witnesses out of the collation entirely, removing references
876 to them in paths, and removing readings that belong only to them. Should only
877 be called via $tradition->del_witness.
882 my( $self, @sigils ) = @_;
884 $self->_graphcalc_done(0);
885 # Clear the witness(es) out of the paths
886 foreach my $e ( $self->paths ) {
887 foreach my $sig ( @sigils ) {
888 $self->del_path( $e, $sig );
892 # Clear out the newly unused readings
893 foreach my $r ( $self->readings ) {
894 unless( $self->reading_witnesses( $r ) ) {
895 $self->del_reading( $r );
900 sub add_relationship {
902 my( $source, $target, $opts ) = $self->_stringify_args( @_ );
903 my( @vectors ) = $self->relations->add_relationship( $source, $target, $opts );
904 foreach my $v ( @vectors ) {
905 next unless $self->get_relationship( $v )->colocated;
906 if( $self->reading( $v->[0] )->has_rank && $self->reading( $v->[1] )->has_rank
907 && $self->reading( $v->[0] )->rank ne $self->reading( $v->[1] )->rank ) {
908 $self->_graphcalc_done(0);
916 around qw/ get_relationship del_relationship / => sub {
920 if( @args == 1 && ref( $args[0] ) eq 'ARRAY' ) {
923 my @stringargs = $self->_stringify_args( @args );
924 $self->$orig( @stringargs );
927 =head2 reading_witnesses( $reading )
929 Return a list of sigils corresponding to the witnesses in which the reading appears.
933 sub reading_witnesses {
934 my( $self, $reading ) = @_;
935 # We need only check either the incoming or the outgoing edges; I have
936 # arbitrarily chosen "incoming". Thus, special-case the start node.
937 if( $reading eq $self->start ) {
938 return map { $_->sigil } grep { $_->is_collated } $self->tradition->witnesses;
941 foreach my $e ( $self->sequence->edges_to( $reading ) ) {
942 my $wits = $self->sequence->get_edge_attributes( @$e );
943 @all_witnesses{ keys %$wits } = 1;
945 my $acstr = $self->ac_label;
946 foreach my $acwit ( grep { $_ =~ s/^(.*)\Q$acstr\E$/$1/ } keys %all_witnesses ) {
947 delete $all_witnesses{$acwit.$acstr} if exists $all_witnesses{$acwit};
949 return keys %all_witnesses;
952 =head1 OUTPUT METHODS
954 =head2 as_svg( \%options )
956 Returns an SVG string that represents the graph, via as_dot and graphviz.
957 See as_dot for a list of options. Must have GraphViz (dot) installed to run.
962 my( $self, $opts ) = @_;
963 throw( "Need GraphViz installed to output SVG" )
964 unless File::Which::which( 'dot' );
965 my $want_subgraph = exists $opts->{'from'} || exists $opts->{'to'};
966 $self->calculate_ranks()
967 unless( $self->_graphcalc_done || $opts->{'nocalc'} || !$self->linear );
968 my @cmd = qw/dot -Tsvg/;
970 my $dotfile = File::Temp->new();
972 # $dotfile->unlink_on_destroy(0);
973 binmode $dotfile, ':utf8';
974 print $dotfile $self->as_dot( $opts );
975 push( @cmd, $dotfile->filename );
976 run( \@cmd, ">", binary(), \$svg );
977 $svg = decode_utf8( $svg );
982 =head2 as_dot( \%options )
984 Returns a string that is the collation graph expressed in dot
985 (i.e. GraphViz) format. Options include:
1000 my( $self, $opts ) = @_;
1001 my $startrank = $opts->{'from'} if $opts;
1002 my $endrank = $opts->{'to'} if $opts;
1003 my $color_common = $opts->{'color_common'} if $opts;
1004 my $STRAIGHTENHACK = !$startrank && !$endrank && $self->end->rank
1005 && $self->end->rank > 100;
1006 $STRAIGHTENHACK = 1 if $opts->{'straight'}; # even for subgraphs or small graphs
1008 # Check the arguments
1010 return if $endrank && $startrank > $endrank;
1011 return if $startrank > $self->end->rank;
1013 if( defined $endrank ) {
1014 return if $endrank < 0;
1015 $endrank = undef if $endrank == $self->end->rank;
1018 my $graph_name = $self->tradition->name;
1019 $graph_name =~ s/[^\w\s]//g;
1020 $graph_name = join( '_', split( /\s+/, $graph_name ) );
1024 'bgcolor' => 'none',
1028 'fillcolor' => 'white',
1029 'style' => 'filled',
1030 'shape' => 'ellipse'
1033 'arrowhead' => 'open',
1034 'color' => '#000000',
1035 'fontcolor' => '#000000',
1038 my $dot = sprintf( "digraph %s {\n", $graph_name );
1039 $dot .= "\tgraph " . _dot_attr_string( \%graph_attrs ) . ";\n";
1040 $dot .= "\tnode " . _dot_attr_string( \%node_attrs ) . ";\n";
1042 # Output substitute start/end readings if necessary
1044 $dot .= "\t\"__SUBSTART__\" [ label=\"...\",id=\"__START__\" ];\n";
1047 $dot .= "\t\"__SUBEND__\" [ label=\"...\",id=\"__END__\" ];\n";
1049 if( $STRAIGHTENHACK ) {
1051 my $startlabel = $startrank ? '__SUBSTART__' : '__START__';
1052 $dot .= "\tsubgraph { rank=same \"$startlabel\" \"#SILENT#\" }\n";
1053 $dot .= "\t\"#SILENT#\" [ shape=diamond,color=white,penwidth=0,label=\"\" ];"
1055 my %used; # Keep track of the readings that actually appear in the graph
1056 # Sort the readings by rank if we have ranks; this speeds layout.
1057 my @all_readings = $self->end->has_rank
1058 ? sort { $a->rank <=> $b->rank } $self->readings
1060 # TODO Refrain from outputting lacuna nodes - just grey out the edges.
1061 foreach my $reading ( @all_readings ) {
1062 # Only output readings within our rank range.
1063 next if $startrank && $reading->rank < $startrank;
1064 next if $endrank && $reading->rank > $endrank;
1065 $used{$reading->id} = 1;
1066 # Need not output nodes without separate labels
1067 next if $reading->id eq $reading->text;
1069 my $label = $reading->text;
1070 unless( $label =~ /^[[:punct:]]+$/ ) {
1071 $label .= '-' if $reading->join_next;
1072 $label = "-$label" if $reading->join_prior;
1074 $label =~ s/\"/\\\"/g;
1075 $rattrs->{'label'} = $label;
1076 $rattrs->{'id'} = $reading->id;
1077 $rattrs->{'fillcolor'} = '#b3f36d' if $reading->is_common && $color_common;
1078 $dot .= sprintf( "\t\"%s\" %s;\n", $reading->id, _dot_attr_string( $rattrs ) );
1081 # Add the real edges.
1082 my @edges = $self->paths;
1083 my( %substart, %subend );
1084 foreach my $edge ( @edges ) {
1085 # Do we need to output this edge?
1086 if( $used{$edge->[0]} && $used{$edge->[1]} ) {
1087 my $label = $self->_path_display_label( $opts,
1088 $self->path_witnesses( $edge ) );
1089 my $variables = { %edge_attrs, 'label' => $label };
1091 # Account for the rank gap if necessary
1092 my $rank0 = $self->reading( $edge->[0] )->rank
1093 if $self->reading( $edge->[0] )->has_rank;
1094 my $rank1 = $self->reading( $edge->[1] )->rank
1095 if $self->reading( $edge->[1] )->has_rank;
1096 if( defined $rank0 && defined $rank1 && $rank1 - $rank0 > 1 ) {
1097 $variables->{'minlen'} = $rank1 - $rank0;
1100 # EXPERIMENTAL: make edge width reflect no. of witnesses
1101 my $extrawidth = scalar( $self->path_witnesses( $edge ) ) * 0.2;
1102 $variables->{'penwidth'} = $extrawidth + 0.8; # gives 1 for a single wit
1104 my $varopts = _dot_attr_string( $variables );
1105 $dot .= sprintf( "\t\"%s\" -> \"%s\" %s;\n",
1106 $edge->[0], $edge->[1], $varopts );
1107 } elsif( $used{$edge->[0]} ) {
1108 $subend{$edge->[0]} = $edge->[1];
1109 } elsif( $used{$edge->[1]} ) {
1110 $substart{$edge->[1]} = $edge->[0];
1114 # If we are asked to, add relationship links
1115 if( exists $opts->{show_relations} ) {
1116 my $filter = $opts->{show_relations}; # can be 'transposition' or 'all'
1117 if( $filter eq 'transposition' ) {
1118 $filter =~ qr/^transposition$/;
1121 my @types = sort( map { $_->name } $self->relations->types );
1122 if( exists $opts->{graphcolors} ) {
1123 foreach my $tdx ( 0 .. $#types ) {
1124 $typecolors{$types[$tdx]} = $opts->{graphcolors}->[$tdx];
1127 map { $typecolors{$_} = '#FFA14F' } @types;
1129 foreach my $redge ( $self->relationships ) {
1130 if( $used{$redge->[0]} && $used{$redge->[1]} ) {
1131 my $rel = $self->get_relationship( $redge );
1132 next unless $filter eq 'all' || $rel->type =~ /$filter/;
1134 arrowhead => 'none',
1135 color => $typecolors{$rel->type},
1136 constraint => 'false',
1139 unless( exists $opts->{graphcolors} ) {
1140 $variables->{label} = uc( substr( $rel->type, 0, 4 ) ),
1142 $dot .= sprintf( "\t\"%s\" -> \"%s\" %s;\n",
1143 $redge->[0], $redge->[1], _dot_attr_string( $variables ) );
1148 # Add substitute start and end edges if necessary
1149 foreach my $node ( keys %substart ) {
1150 my $witstr = $self->_path_display_label( $opts,
1151 $self->path_witnesses( $substart{$node}, $node ) );
1152 my $variables = { %edge_attrs, 'label' => $witstr };
1153 my $nrdg = $self->reading( $node );
1154 if( $nrdg->has_rank && $nrdg->rank > $startrank ) {
1155 # Substart is actually one lower than $startrank
1156 $variables->{'minlen'} = $nrdg->rank - ( $startrank - 1 );
1158 my $varopts = _dot_attr_string( $variables );
1159 $dot .= "\t\"__SUBSTART__\" -> \"$node\" $varopts;\n";
1161 foreach my $node ( keys %subend ) {
1162 my $witstr = $self->_path_display_label( $opts,
1163 $self->path_witnesses( $node, $subend{$node} ) );
1164 my $variables = { %edge_attrs, 'label' => $witstr };
1165 my $varopts = _dot_attr_string( $variables );
1166 $dot .= "\t\"$node\" -> \"__SUBEND__\" $varopts;\n";
1169 if( $STRAIGHTENHACK ) {
1170 my $endlabel = $endrank ? '__SUBEND__' : '__END__';
1171 $dot .= "\t\"$endlabel\" -> \"#SILENT#\" [ color=white,penwidth=0 ];\n";
1178 sub _dot_attr_string {
1181 foreach my $k ( sort keys %$hash ) {
1182 my $v = $hash->{$k};
1183 push( @attrs, $k.'="'.$v.'"' );
1185 return( '[ ' . join( ', ', @attrs ) . ' ]' );
1188 =head2 path_witnesses( $edge )
1190 Returns the list of sigils whose witnesses are associated with the given edge.
1191 The edge can be passed as either an array or an arrayref of ( $source, $target ).
1195 sub path_witnesses {
1196 my( $self, @edge ) = @_;
1197 # If edge is an arrayref, cope.
1198 if( @edge == 1 && ref( $edge[0] ) eq 'ARRAY' ) {
1199 my $e = shift @edge;
1202 my @wits = keys %{$self->sequence->get_edge_attributes( @edge )};
1206 # Helper function. Make a display label for the given witnesses, showing a.c.
1207 # witnesses only where the main witness is not also in the list.
1208 sub _path_display_label {
1212 map { $wits{$_} = 1 } @_;
1214 # If an a.c. wit is listed, remove it if the main wit is also listed.
1215 # Otherwise keep it for explicit listing.
1216 my $aclabel = $self->ac_label;
1218 foreach my $w ( sort keys %wits ) {
1219 if( $w =~ /^(.*)\Q$aclabel\E$/ ) {
1220 if( exists $wits{$1} ) {
1223 push( @disp_ac, $w );
1228 if( $opts->{'explicit_wits'} ) {
1229 return join( ', ', sort keys %wits );
1231 # See if we are in a majority situation.
1232 my $maj = scalar( $self->tradition->witnesses ) * 0.6;
1233 $maj = $maj > 5 ? $maj : 5;
1234 if( scalar keys %wits > $maj ) {
1235 unshift( @disp_ac, 'majority' );
1236 return join( ', ', @disp_ac );
1238 return join( ', ', sort keys %wits );
1243 =head2 as_adjacency_list
1245 Returns a JSON structure that represents the collation sequence graph.
1249 use JSON qw/ from_json /;
1250 use Text::Tradition;
1252 my $t = Text::Tradition->new(
1254 'file' => 't/data/florilegium_graphml.xml' );
1255 my $c = $t->collation;
1257 # Make a connection so we can test rank preservation
1258 $c->add_relationship( 'w91', 'w92', { type => 'grammatical' } );
1260 # Create an adjacency list of the whole thing; test the output.
1261 my $adj_whole = from_json( $c->as_adjacency_list() );
1262 is( scalar @$adj_whole, scalar $c->readings(),
1263 "Same number of nodes in graph and adjacency list" );
1264 my @adj_whole_edges;
1265 map { push( @adj_whole_edges, @{$_->{adjacent}} ) } @$adj_whole;
1266 is( scalar @adj_whole_edges, scalar $c->sequence->edges,
1267 "Same number of edges in graph and adjacency list" );
1268 # Find the reading whose rank should be preserved
1269 my( $test_rdg ) = grep { $_->{id} eq 'w89' } @$adj_whole;
1270 my( $test_edge ) = grep { $_->{id} eq 'w92' } @{$test_rdg->{adjacent}};
1271 is( $test_edge->{minlen}, 2, "Rank of test reading is preserved" );
1273 # Now create an adjacency list of just a portion. w76 to w122
1274 my $adj_part = from_json( $c->as_adjacency_list(
1275 { from => $c->reading('w76')->rank,
1276 to => $c->reading('w122')->rank }));
1277 is( scalar @$adj_part, 48, "Correct number of nodes in partial graph" );
1279 map { push( @adj_part_edges, @{$_->{adjacent}} ) } @$adj_part;
1280 is( scalar @adj_part_edges, 58,
1281 "Same number of edges in partial graph and adjacency list" );
1282 # Check for consistency
1284 map { $part_nodes{$_->{id}} = 1 } @$adj_part;
1285 foreach my $edge ( @adj_part_edges ) {
1286 my $testid = $edge->{id};
1287 ok( $part_nodes{$testid}, "ID $testid referenced in edge is given as node" );
1294 sub as_adjacency_list {
1295 my( $self, $opts ) = @_;
1296 # Make a structure that contains all the nodes, the nodes they point to,
1297 # and the attributes of the edges that connect them.
1298 # [ { id: 'n0', label: 'Gallia', adjacent: [
1299 # { id: 'n1', label: 'P Q' } ,
1300 # { id: 'n2', label: 'R S', minlen: 2 } ] },
1301 # { id: 'n1', label: 'est', adjacent: [ ... ] },
1303 my $startrank = $opts->{'from'} || 0;
1304 my $endrank = $opts->{'to'} || $self->end->rank;
1306 $self->calculate_ranks()
1307 unless( $self->_graphcalc_done || $opts->{'nocalc'} || !$self->linear );
1309 foreach my $rdg ( $self->readings ) {
1312 # Figure out what the node's successors should be.
1313 if( $rdg eq $self->start && $startrank > 0 ) {
1314 # Connect the start node with all the nodes at startrank.
1315 # Lacunas should be included only if the node really has that rank.
1316 @successors = $self->readings_at_rank( $startrank, 1 );
1318 } elsif( $rdg->rank < $startrank
1319 || $rdg->rank > $endrank && $rdg ne $self->end ) {
1322 @successors = $rdg->successors;
1324 # Make sure that the end node is at the end of the successors
1325 # list if it is needed.
1326 if( grep { $_ eq $self->end } @successors ) {
1327 my @ts = grep { $_ ne $self->end } @successors;
1328 @successors = ( @ts, $self->end );
1329 } elsif ( grep { $_->rank > $endrank } @successors ) {
1330 push( @successors, $self->end );
1333 my $listitem = { id => $rdg->id, label => $rdg->text };
1336 foreach my $succ ( @successors ) {
1338 if( $phony eq 'start' ) {
1339 @edgewits = $succ->witnesses;
1340 } elsif( $self->sequence->has_edge( $rdg->id, $succ->id ) ) {
1341 @edgewits = $self->path_witnesses( $rdg->id, $succ->id );
1344 if( $succ eq $self->end ) {
1345 @edgewits = @endwits;
1346 } elsif( $succ->rank > $endrank ) {
1347 # These witnesses will point to 'end' instead, not to the
1349 push( @endwits, @edgewits );
1352 my $edgelabel = $self->_path_display_label( $opts, @edgewits );
1353 my $edgedef = { id => $succ->id, label => $edgelabel };
1354 my $rankoffset = $succ->rank - $rdg->rank;
1355 if( $rankoffset > 1 and $succ ne $self->end ) {
1356 $edgedef->{minlen} = $rankoffset;
1358 push( @$adjacent, $edgedef );
1360 $listitem->{adjacent} = $adjacent;
1361 push( @$list, $listitem );
1363 return to_json( $list );
1368 Returns a GraphML representation of the collation. The GraphML will contain
1369 two graphs. The first expresses the attributes of the readings and the witness
1370 paths that link them; the second expresses the relationships that link the
1371 readings. This is the native transfer format for a tradition.
1375 use Text::Tradition;
1381 my $datafile = 't/data/florilegium_tei_ps.xml';
1382 my $tradition = Text::Tradition->new( 'input' => 'TEI',
1384 'file' => $datafile,
1387 ok( $tradition, "Got a tradition object" );
1388 is( scalar $tradition->witnesses, 13, "Found all witnesses" );
1389 ok( $tradition->collation, "Tradition has a collation" );
1391 my $c = $tradition->collation;
1392 is( scalar $c->readings, $READINGS, "Collation has all readings" );
1393 is( scalar $c->paths, $PATHS, "Collation has all paths" );
1394 is( scalar $c->relationships, 0, "Collation has all relationships" );
1396 # Add a few relationships
1397 $c->add_relationship( 'w123', 'w125', { 'type' => 'collated' } );
1398 $c->add_relationship( 'w193', 'w196', { 'type' => 'collated' } );
1399 $c->add_relationship( 'w257', 'w262', { 'type' => 'transposition',
1400 'is_significant' => 'yes' } );
1402 # Now write it to GraphML and parse it again.
1404 my $graphml = $c->as_graphml;
1405 my $st = Text::Tradition->new( 'input' => 'Self', 'string' => $graphml );
1406 is( scalar $st->collation->readings, $READINGS, "Reparsed collation has all readings" );
1407 is( scalar $st->collation->paths, $PATHS, "Reparsed collation has all paths" );
1408 is( scalar $st->collation->relationships, 3, "Reparsed collation has new relationships" );
1409 my $sigrel = $st->collation->get_relationship( 'w257', 'w262' );
1410 is( $sigrel->is_significant, 'yes', "Ternary attribute value was restored" );
1412 # Now add a stemma, write to GraphML, and look at the output.
1414 skip "Analysis module not present", 3 unless $tradition->can( 'add_stemma' );
1415 my $stemma = $tradition->add_stemma( 'dotfile' => 't/data/florilegium.dot' );
1416 is( ref( $stemma ), 'Text::Tradition::Stemma', "Parsed dotfile into stemma" );
1417 is( $tradition->stemmata, 1, "Tradition now has the stemma" );
1418 $graphml = $c->as_graphml;
1419 like( $graphml, qr/digraph/, "Digraph declaration exists in GraphML" );
1426 ## TODO MOVE this to Tradition.pm and modularize it better
1428 my( $self, $options ) = @_;
1429 $self->calculate_ranks unless $self->_graphcalc_done;
1431 my $start = $options->{'from'}
1432 ? $self->reading( $options->{'from'} ) : $self->start;
1433 my $end = $options->{'to'}
1434 ? $self->reading( $options->{'to'} ) : $self->end;
1435 if( $start->has_rank && $end->has_rank && $end->rank < $start->rank ) {
1436 throw( 'Start node must be before end node' );
1438 # The readings need to be ranked for this to work.
1439 $start = $self->start unless $start->has_rank;
1440 $end = $self->end unless $end->has_rank;
1442 unless( $start eq $self->start ) {
1443 $rankoffset = $start->rank - 1;
1448 my $graphml_ns = 'http://graphml.graphdrawing.org/xmlns';
1449 my $xsi_ns = 'http://www.w3.org/2001/XMLSchema-instance';
1450 my $graphml_schema = 'http://graphml.graphdrawing.org/xmlns ' .
1451 'http://graphml.graphdrawing.org/xmlns/1.0/graphml.xsd';
1453 # Create the document and root node
1454 require XML::LibXML;
1455 my $graphml = XML::LibXML->createDocument( "1.0", "UTF-8" );
1456 my $root = $graphml->createElementNS( $graphml_ns, 'graphml' );
1457 $graphml->setDocumentElement( $root );
1458 $root->setNamespace( $xsi_ns, 'xsi', 0 );
1459 $root->setAttributeNS( $xsi_ns, 'schemaLocation', $graphml_schema );
1461 # List of attribute types to save on our objects and their corresponding
1466 'Bool' => 'boolean',
1467 'ReadingID' => 'string',
1468 'RelationshipType' => 'string',
1469 'RelationshipScope' => 'string',
1470 'Ternary' => 'string',
1473 # Add the data keys for the graph. Include an extra key 'version' for the
1474 # GraphML output version.
1475 my %graph_data_keys;
1477 my %graph_attributes = ( 'version' => 'string' );
1478 # Graph attributes include those of Tradition and those of Collation.
1480 # TODO Use meta introspection method from duplicate_reading to do this
1481 # instead of naming custom keys.
1482 my $tmeta = $self->tradition->meta;
1483 my $cmeta = $self->meta;
1484 map { $gattr_from{$_->name} = 'Tradition' } $tmeta->get_all_attributes;
1485 map { $gattr_from{$_->name} = 'Collation' } $cmeta->get_all_attributes;
1486 foreach my $attr ( ( $tmeta->get_all_attributes, $cmeta->get_all_attributes ) ) {
1487 next if $attr->name =~ /^_/;
1488 next unless $save_types{$attr->type_constraint->name};
1489 $graph_attributes{$attr->name} = $save_types{$attr->type_constraint->name};
1491 # Extra custom keys for complex objects that should be saved in some form.
1492 # The subroutine should return a string, or undef/empty.
1493 if( $tmeta->has_method('stemmata') ) {
1494 $graph_attributes{'stemmata'} = sub {
1496 map { push( @stemstrs, $_->editable( {linesep => ''} ) ) }
1497 $self->tradition->stemmata;
1498 join( "\n", @stemstrs );
1502 if( $tmeta->has_method('user') ) {
1503 $graph_attributes{'user'} = sub {
1504 $self->tradition->user ? $self->tradition->user->id : undef
1508 foreach my $datum ( sort keys %graph_attributes ) {
1509 $graph_data_keys{$datum} = 'dg'.$gdi++;
1510 my $key = $root->addNewChild( $graphml_ns, 'key' );
1511 my $dtype = ref( $graph_attributes{$datum} ) ? 'string'
1512 : $graph_attributes{$datum};
1513 $key->setAttribute( 'attr.name', $datum );
1514 $key->setAttribute( 'attr.type', $dtype );
1515 $key->setAttribute( 'for', 'graph' );
1516 $key->setAttribute( 'id', $graph_data_keys{$datum} );
1519 # Add the data keys for reading nodes
1520 my %reading_attributes;
1521 my $rmeta = Text::Tradition::Collation::Reading->meta;
1522 foreach my $attr( $rmeta->get_all_attributes ) {
1523 next if $attr->name =~ /^_/;
1524 next unless $save_types{$attr->type_constraint->name};
1525 $reading_attributes{$attr->name} = $save_types{$attr->type_constraint->name};
1527 if( $self->start->does('Text::Tradition::Morphology' ) ) {
1528 # Extra custom key for the reading morphology
1529 $reading_attributes{'lexemes'} = 'string';
1534 foreach my $datum ( sort keys %reading_attributes ) {
1535 $node_data_keys{$datum} = 'dn'.$ndi++;
1536 my $key = $root->addNewChild( $graphml_ns, 'key' );
1537 $key->setAttribute( 'attr.name', $datum );
1538 $key->setAttribute( 'attr.type', $reading_attributes{$datum} );
1539 $key->setAttribute( 'for', 'node' );
1540 $key->setAttribute( 'id', $node_data_keys{$datum} );
1543 # Add the data keys for edges, that is, paths and relationships. Path
1544 # data does not come from a Moose class so is here manually.
1547 my %edge_attributes = (
1548 witness => 'string', # ID/label for a path
1549 extra => 'boolean', # Path key
1551 my @path_attributes = keys %edge_attributes; # track our manual additions
1552 my $pmeta = Text::Tradition::Collation::Relationship->meta;
1553 foreach my $attr( $pmeta->get_all_attributes ) {
1554 next if $attr->name =~ /^_/;
1555 next unless $save_types{$attr->type_constraint->name};
1556 $edge_attributes{$attr->name} = $save_types{$attr->type_constraint->name};
1558 foreach my $datum ( sort keys %edge_attributes ) {
1559 $edge_data_keys{$datum} = 'de'.$edi++;
1560 my $key = $root->addNewChild( $graphml_ns, 'key' );
1561 $key->setAttribute( 'attr.name', $datum );
1562 $key->setAttribute( 'attr.type', $edge_attributes{$datum} );
1563 $key->setAttribute( 'for', 'edge' );
1564 $key->setAttribute( 'id', $edge_data_keys{$datum} );
1567 # Add the collation graph itself. First, sanitize the name to a valid XML ID.
1568 my $xmlidname = $self->tradition->name;
1569 $xmlidname =~ s/(?!$xml10_namechar_rx)./_/g;
1570 if( $xmlidname !~ /^$xml10_namestartchar_rx/ ) {
1571 $xmlidname = '_'.$xmlidname;
1573 my $sgraph = $root->addNewChild( $graphml_ns, 'graph' );
1574 $sgraph->setAttribute( 'edgedefault', 'directed' );
1575 $sgraph->setAttribute( 'id', $xmlidname );
1576 $sgraph->setAttribute( 'parse.edgeids', 'canonical' );
1577 $sgraph->setAttribute( 'parse.edges', 0 ); # fill in later
1578 $sgraph->setAttribute( 'parse.nodeids', 'canonical' );
1579 $sgraph->setAttribute( 'parse.nodes', 0 ); # fill in later
1580 $sgraph->setAttribute( 'parse.order', 'nodesfirst' );
1582 # Tradition/collation attribute data
1583 foreach my $datum ( keys %graph_attributes ) {
1585 if( $datum eq 'version' ) {
1587 } elsif( ref( $graph_attributes{$datum} ) ) {
1588 my $sub = $graph_attributes{$datum};
1590 } elsif( $gattr_from{$datum} eq 'Tradition' ) {
1591 $value = $self->tradition->$datum;
1593 $value = $self->$datum;
1595 _add_graphml_data( $sgraph, $graph_data_keys{$datum}, $value );
1600 # Add our readings to the graph
1601 foreach my $n ( sort { $a->id cmp $b->id } $self->readings ) {
1602 next if $n->has_rank && $n ne $self->start && $n ne $self->end &&
1603 ( $n->rank < $start->rank || $n->rank > $end->rank );
1604 $use_readings{$n->id} = 1;
1605 # Add to the main graph
1606 my $node_el = $sgraph->addNewChild( $graphml_ns, 'node' );
1607 my $node_xmlid = 'n' . $node_ctr++;
1608 $node_hash{ $n->id } = $node_xmlid;
1609 $node_el->setAttribute( 'id', $node_xmlid );
1610 foreach my $d ( keys %reading_attributes ) {
1612 # Custom serialization
1613 if( $d eq 'lexemes' ) {
1614 # If nval is a true value, we have lexemes so we need to
1615 # serialize them. Otherwise set nval to undef so that the
1616 # key is excluded from this reading.
1617 $nval = $nval ? $n->_serialize_lexemes : undef;
1618 } elsif( $d eq 'normal_form' && $n->normal_form eq $n->text ) {
1621 if( $rankoffset && $d eq 'rank' && $n ne $self->start ) {
1622 # Adjust the ranks within the subgraph.
1623 $nval = $n eq $self->end ? $end->rank - $rankoffset + 1
1624 : $nval - $rankoffset;
1626 _add_graphml_data( $node_el, $node_data_keys{$d}, $nval )
1631 # Add the path edges to the sequence graph
1633 foreach my $e ( sort { $a->[0] cmp $b->[0] } $self->sequence->edges() ) {
1634 # We add an edge in the graphml for every witness in $e.
1635 next unless( $use_readings{$e->[0]} || $use_readings{$e->[1]} );
1636 my @edge_wits = sort $self->path_witnesses( $e );
1637 $e->[0] = $self->start->id unless $use_readings{$e->[0]};
1638 $e->[1] = $self->end->id unless $use_readings{$e->[1]};
1639 # Skip any path from start to end; that witness is not in the subgraph.
1640 next if ( $e->[0] eq $self->start->id && $e->[1] eq $self->end->id );
1641 foreach my $wit ( @edge_wits ) {
1642 my( $id, $from, $to ) = ( 'e'.$edge_ctr++,
1643 $node_hash{ $e->[0] },
1644 $node_hash{ $e->[1] } );
1645 my $edge_el = $sgraph->addNewChild( $graphml_ns, 'edge' );
1646 $edge_el->setAttribute( 'source', $from );
1647 $edge_el->setAttribute( 'target', $to );
1648 $edge_el->setAttribute( 'id', $id );
1650 # It's a witness path, so add the witness
1652 my $key = $edge_data_keys{'witness'};
1653 # Is this an ante-corr witness?
1654 my $aclabel = $self->ac_label;
1655 if( $wit =~ /^(.*)\Q$aclabel\E$/ ) {
1656 # Keep the base witness
1658 # ...and record that this is an 'extra' reading path
1659 _add_graphml_data( $edge_el, $edge_data_keys{'extra'}, $aclabel );
1661 _add_graphml_data( $edge_el, $edge_data_keys{'witness'}, $base );
1665 # Report the actual number of nodes and edges that went in
1666 $sgraph->setAttribute( 'parse.edges', $edge_ctr );
1667 $sgraph->setAttribute( 'parse.nodes', $node_ctr );
1669 # Add the relationship graph to the XML
1670 map { delete $edge_data_keys{$_} } @path_attributes;
1671 $self->relations->_as_graphml( $graphml_ns, $root, \%node_hash,
1672 $node_data_keys{'id'}, \%edge_data_keys );
1674 # Save and return the thing
1675 my $result = decode_utf8( $graphml->toString(1) );
1679 sub _add_graphml_data {
1680 my( $el, $key, $value ) = @_;
1681 return unless defined $value;
1682 my $data_el = $el->addNewChild( $el->namespaceURI, 'data' );
1683 $data_el->setAttribute( 'key', $key );
1684 $data_el->appendText( $value );
1689 Returns a CSV alignment table representation of the collation graph, one
1690 row per witness (or witness uncorrected.)
1694 Returns a tab-separated alignment table representation of the collation graph,
1695 one row per witness (or witness uncorrected.)
1699 use Text::Tradition;
1707 my $datafile = 't/data/florilegium_tei_ps.xml';
1708 my $tradition = Text::Tradition->new( 'input' => 'TEI',
1710 'file' => $datafile,
1713 my $c = $tradition->collation;
1714 # Export the thing to CSV
1715 my $csvstr = $c->as_csv();
1717 my $csv = Text::CSV->new({ sep_char => ',', binary => 1 });
1718 my @lines = split(/\n/, $csvstr );
1719 ok( $csv->parse( $lines[0] ), "Successfully parsed first line of CSV" );
1720 is( scalar( $csv->fields ), $WITS + $WITAC, "CSV has correct number of witness columns" );
1721 my @q_ac = grep { $_ eq 'Q'.$c->ac_label } $csv->fields;
1722 ok( @q_ac, "Found a layered witness" );
1724 my $t2 = Text::Tradition->new( input => 'Tabular',
1728 is( scalar $t2->collation->readings, $READINGS, "Reparsed CSV collation has all readings" );
1729 is( scalar $t2->collation->paths, $PATHS, "Reparsed CSV collation has all paths" );
1731 # Now do it with TSV
1732 my $tsvstr = $c->as_tsv();
1733 my $t3 = Text::Tradition->new( input => 'Tabular',
1737 is( scalar $t3->collation->readings, $READINGS, "Reparsed TSV collation has all readings" );
1738 is( scalar $t3->collation->paths, $PATHS, "Reparsed TSV collation has all paths" );
1740 my $table = $c->alignment_table;
1741 my $noaccsv = $c->as_csv({ noac => 1 });
1742 my @noaclines = split(/\n/, $noaccsv );
1743 ok( $csv->parse( $noaclines[0] ), "Successfully parsed first line of no-ac CSV" );
1744 is( scalar( $csv->fields ), $WITS, "CSV has correct number of witness columns" );
1745 is( $c->alignment_table, $table, "Request for CSV did not alter the alignment table" );
1747 my $safecsv = $c->as_csv({ safe_ac => 1});
1748 my @safelines = split(/\n/, $safecsv );
1749 ok( $csv->parse( $safelines[0] ), "Successfully parsed first line of safe CSV" );
1750 is( scalar( $csv->fields ), $WITS + $WITAC, "CSV has correct number of witness columns" );
1751 @q_ac = grep { $_ eq 'Q__L' } $csv->fields;
1752 ok( @q_ac, "Found a sanitized layered witness" );
1753 is( $c->alignment_table, $table, "Request for CSV did not alter the alignment table" );
1755 # Test relationship collapse
1756 $c->add_relationship( $c->readings_at_rank( 37 ), { type => 'spelling' } );
1757 $c->add_relationship( $c->readings_at_rank( 60 ), { type => 'spelling' } );
1759 my $mergedtsv = $c->as_tsv({mergetypes => [ 'spelling', 'orthographic' ] });
1760 my $t4 = Text::Tradition->new( input => 'Tabular',
1762 string => $mergedtsv,
1764 is( scalar $t4->collation->readings, $READINGS - 2, "Reparsed TSV merge collation has fewer readings" );
1765 is( scalar $t4->collation->paths, $PATHS - 4, "Reparsed TSV merge collation has fewer paths" );
1767 # Test non-ASCII sigla
1768 my $t5 = Text::Tradition->new( input => 'Tabular',
1770 file => 't/data/armexample.xlsx',
1772 my $awittsv = $t5->collation->as_tsv({ noac => 1, ascii => 1 });
1773 my @awitlines = split( /\n/, $awittsv );
1774 like( $awitlines[0], qr/_A_5315622/, "Found ASCII sigil variant in TSV" );
1781 my( $self, $opts ) = @_;
1782 my $table = $self->alignment_table( $opts );
1783 my $csv_options = { binary => 1, quote_null => 0 };
1784 $csv_options->{'sep_char'} = $opts->{fieldsep};
1785 if( $opts->{fieldsep} eq "\t" ) {
1786 # If it is really tab separated, nothing is an escape char.
1787 $csv_options->{'quote_char'} = undef;
1788 $csv_options->{'escape_char'} = '';
1790 my $csv = Text::CSV->new( $csv_options );
1793 # Make the header row
1794 my @witnesses = map { $_->{'witness'} } @{$table->{'alignment'}};
1795 if( $opts->{ascii} ) {
1796 # TODO think of a fix for this
1797 throw( "Cannot currently produce ASCII sigla with witness layers" )
1798 unless $opts->{noac};
1799 my @awits = map { $self->tradition->witness( $_ )->ascii_sigil } @witnesses;
1800 @witnesses = @awits;
1802 $csv->combine( @witnesses );
1803 push( @result, $csv->string );
1805 # Make the rest of the rows
1806 foreach my $idx ( 0 .. $table->{'length'} - 1 ) {
1807 my @rowobjs = map { $_->{'tokens'}->[$idx] } @{$table->{'alignment'}};
1808 my @row = map { $_ ? $_->{'t'}->text : $_ } @rowobjs;
1809 # Quick and dirty collapse of requested relationship types
1810 if( ref( $opts->{mergetypes} ) eq 'ARRAY' ) {
1811 # Now substitute the reading in the relevant index of @row
1812 # for its merge-related reading
1815 my $thisr = shift @rowobjs;
1817 next if exists $substitutes{$thisr->{t}->text};
1818 # Make sure we don't have A <-> B substitutions.
1819 $substitutes{$thisr->{t}->text} = $thisr->{t}->text;
1820 foreach my $thatr ( @rowobjs ) {
1822 next if exists $substitutes{$thatr->{t}->text};
1823 my $ttrel = $self->get_relationship( $thisr->{t}, $thatr->{t} );
1825 next unless grep { $ttrel->type eq $_ } @{$opts->{mergetypes}};
1826 # If we have got this far then we need to merge them.
1827 $substitutes{$thatr->{t}->text} = $thisr->{t}->text;
1830 @row = map { $_ && exists $substitutes{$_} ? $substitutes{$_} : $_ } @row;
1832 $csv->combine( @row );
1833 push( @result, $csv->string );
1835 return join( "\n", @result );
1840 my $opts = shift || {};
1841 $opts->{fieldsep} = ',';
1842 return $self->_tabular( $opts );
1847 my $opts = shift || {};
1848 $opts->{fieldsep} = "\t";
1849 return $self->_tabular( $opts );
1852 =head2 alignment_table
1854 Return a reference to an alignment table, in a slightly enhanced CollateX
1855 format which looks like this:
1857 $table = { alignment => [ { witness => "SIGIL",
1858 tokens => [ { t => "TEXT" }, ... ] },
1859 { witness => "SIG2",
1860 tokens => [ { t => "TEXT" }, ... ] },
1862 length => TEXTLEN };
1866 sub alignment_table {
1867 my( $self, $opts ) = @_;
1868 if( $self->has_cached_table ) {
1869 return $self->cached_table
1870 unless $opts->{noac} || $opts->{safe_ac};
1873 # Make sure we can do this
1874 throw( "Need a linear graph in order to make an alignment table" )
1875 unless $self->linear;
1876 $self->calculate_ranks()
1877 unless $self->_graphcalc_done && $self->end->has_rank;
1879 my $table = { 'alignment' => [], 'length' => $self->end->rank - 1 };
1880 my @all_pos = ( 1 .. $self->end->rank - 1 );
1881 foreach my $wit ( sort { $a->sigil cmp $b->sigil } $self->tradition->witnesses ) {
1882 # say STDERR "Making witness row(s) for " . $wit->sigil;
1883 my @wit_path = $self->reading_sequence( $self->start, $self->end, $wit->sigil );
1884 my @row = _make_witness_row( \@wit_path, \@all_pos );
1885 my $witobj = { 'witness' => $wit->sigil, 'tokens' => \@row };
1886 $witobj->{'identifier'} = $wit->identifier if $wit->identifier;
1887 push( @{$table->{'alignment'}}, $witobj );
1888 if( $wit->is_layered && !$opts->{noac} ) {
1889 my @wit_ac_path = $self->reading_sequence( $self->start, $self->end,
1890 $wit->sigil.$self->ac_label );
1891 my @ac_row = _make_witness_row( \@wit_ac_path, \@all_pos );
1892 my $witlabel = $opts->{safe_ac}
1893 ? $wit->sigil . '__L' : $wit->sigil.$self->ac_label;
1894 my $witacobj = { 'witness' => $witlabel,
1895 'tokens' => \@ac_row };
1896 $witacobj->{'identifier'} = $wit->identifier if $wit->identifier;
1897 push( @{$table->{'alignment'}}, $witacobj );
1900 unless( $opts->{noac} || $opts->{safe_ac} ) {
1901 $self->cached_table( $table );
1906 sub _make_witness_row {
1907 my( $path, $positions ) = @_;
1909 map { $char_hash{$_} = undef } @$positions;
1911 foreach my $rdg ( @$path ) {
1912 say STDERR "rank " . $rdg->rank if $debug;
1913 # say STDERR "No rank for " . $rdg->id unless defined $rdg->rank;
1914 $char_hash{$rdg->rank} = { 't' => $rdg };
1916 my @row = map { $char_hash{$_} } @$positions;
1917 # Fill in lacuna markers for undef spots in the row
1918 my $last_el = shift @row;
1919 my @filled_row = ( $last_el );
1920 foreach my $el ( @row ) {
1921 # If we are using node reference, make the lacuna node appear many times
1922 # in the table. If not, use the lacuna tag.
1923 if( $last_el && $last_el->{'t'}->is_lacuna && !defined $el ) {
1926 push( @filled_row, $el );
1933 =head1 NAVIGATION METHODS
1935 =head2 reading_sequence( $first, $last, $sigil, $backup )
1937 Returns the ordered list of readings, starting with $first and ending
1938 with $last, for the witness given in $sigil. If a $backup sigil is
1939 specified (e.g. when walking a layered witness), it will be used wherever
1940 no $sigil path exists. If there is a base text reading, that will be
1941 used wherever no path exists for $sigil or $backup.
1945 # TODO Think about returning some lazy-eval iterator.
1946 # TODO Get rid of backup; we should know from what witness is whether we need it.
1948 sub reading_sequence {
1949 my( $self, $start, $end, $witness ) = @_;
1951 $witness = $self->baselabel unless $witness;
1952 my @readings = ( $start );
1955 while( $n && $n->id ne $end->id ) {
1956 if( exists( $seen{$n->id} ) ) {
1957 throw( "Detected loop for $witness at " . $n->id );
1961 my $next = $self->next_reading( $n, $witness );
1963 throw( "Did not find any path for $witness from reading " . $n->id );
1965 push( @readings, $next );
1968 # Check that the last reading is our end reading.
1969 my $last = $readings[$#readings];
1970 throw( "Last reading found from " . $start->text .
1971 " for witness $witness is not the end!" ) # TODO do we get this far?
1972 unless $last->id eq $end->id;
1977 =head2 readings_at_rank( $rank )
1979 Returns a list of readings at a given rank, taken from the alignment table.
1983 sub readings_at_rank {
1984 my( $self, $rank, $nolacuna ) = @_;
1985 my $table = $self->alignment_table;
1986 # Table rank is real rank - 1.
1987 my @elements = map { $_->{'tokens'}->[$rank-1] } @{$table->{'alignment'}};
1989 foreach my $e ( @elements ) {
1990 next unless ref( $e ) eq 'HASH';
1991 next unless exists $e->{'t'};
1992 my $rdg = $e->{'t'};
1993 next if $nolacuna && $rdg->is_lacuna && $rdg->rank ne $rank;
1994 $readings{$e->{'t'}->id} = $e->{'t'};
1996 return values %readings;
1999 =head2 next_reading( $reading, $sigil );
2001 Returns the reading that follows the given reading along the given witness
2007 # Return the successor via the corresponding path.
2009 my $answer = $self->_find_linked_reading( 'next', @_ );
2010 return undef unless $answer;
2011 return $self->reading( $answer );
2014 =head2 prior_reading( $reading, $sigil )
2016 Returns the reading that precedes the given reading along the given witness
2022 # Return the predecessor via the corresponding path.
2024 my $answer = $self->_find_linked_reading( 'prior', @_ );
2025 return $self->reading( $answer );
2028 sub _find_linked_reading {
2029 my( $self, $direction, $node, $path ) = @_;
2031 # Get a backup if we are dealing with a layered witness
2033 my $aclabel = $self->ac_label;
2034 if( $path && $path =~ /^(.*)\Q$aclabel\E$/ ) {
2038 my @linked_paths = $direction eq 'next'
2039 ? $self->sequence->edges_from( $node )
2040 : $self->sequence->edges_to( $node );
2041 return undef unless scalar( @linked_paths );
2043 # We have to find the linked path that contains all of the
2044 # witnesses supplied in $path.
2045 my( @path_wits, @alt_path_wits );
2046 @path_wits = sort( $self->_witnesses_of_label( $path ) ) if $path;
2047 @alt_path_wits = sort( $self->_witnesses_of_label( $alt_path ) ) if $alt_path;
2050 foreach my $le ( @linked_paths ) {
2051 if( $self->sequence->has_edge_attribute( @$le, $self->baselabel ) ) {
2054 my @le_wits = sort $self->path_witnesses( $le );
2055 if( _is_within( \@path_wits, \@le_wits ) ) {
2056 # This is the right path.
2057 return $direction eq 'next' ? $le->[1] : $le->[0];
2058 } elsif( _is_within( \@alt_path_wits, \@le_wits ) ) {
2062 # Got this far? Return the alternate path if it exists.
2063 return $direction eq 'next' ? $alt_le->[1] : $alt_le->[0]
2066 # Got this far? Return the base path if it exists.
2067 return $direction eq 'next' ? $base_le->[1] : $base_le->[0]
2070 # Got this far? We have no appropriate path.
2071 warn "Could not find $direction node from " . $node->id
2072 . " along path $path";
2078 my( $set1, $set2 ) = @_;
2079 my $ret = @$set1; # will be 0, i.e. false, if set1 is empty
2080 foreach my $el ( @$set1 ) {
2081 $ret = 0 unless grep { /^\Q$el\E$/ } @$set2;
2086 # Return the string that joins together a list of witnesses for
2087 # display on a single path.
2088 sub _witnesses_of_label {
2089 my( $self, $label ) = @_;
2090 my $regex = $self->wit_list_separator;
2091 my @answer = split( /\Q$regex\E/, $label );
2095 =head2 common_readings
2097 Returns the list of common readings in the graph (i.e. those readings that are
2098 shared by all non-lacunose witnesses.)
2102 sub common_readings {
2104 my @common = grep { $_->is_common } $self->readings;
2108 =head2 path_text( $sigil, [, $start, $end ] )
2110 Returns the text of a witness (plus its backup, if we are using a layer)
2111 as stored in the collation. The text is returned as a string, where the
2112 individual readings are joined with spaces and the meta-readings (e.g.
2113 lacunae) are omitted. Optional specification of $start and $end allows
2114 the generation of a subset of the witness text.
2119 my( $self, $wit, $start, $end ) = @_;
2120 $start = $self->start unless $start;
2121 $end = $self->end unless $end;
2122 my @path = grep { !$_->is_meta } $self->reading_sequence( $start, $end, $wit );
2125 foreach my $r ( @path ) {
2126 unless ( $r->join_prior || !$last || $last->join_next ) {
2129 $pathtext .= $r->text;
2135 =head1 INITIALIZATION METHODS
2137 These are mostly for use by parsers.
2139 =head2 make_witness_path( $witness )
2141 Link the array of readings contained in $witness->path (and in
2142 $witness->uncorrected_path if it exists) into collation paths.
2143 Clear out the arrays when finished.
2145 =head2 make_witness_paths
2147 Call make_witness_path for all witnesses in the tradition.
2151 # For use when a collation is constructed from a base text and an apparatus.
2152 # We have the sequences of readings and just need to add path edges.
2153 # When we are done, clear out the witness path attributes, as they are no
2155 # TODO Find a way to replace the witness path attributes with encapsulated functions?
2157 sub make_witness_paths {
2159 foreach my $wit ( $self->tradition->witnesses ) {
2160 # say STDERR "Making path for " . $wit->sigil;
2161 $self->make_witness_path( $wit );
2165 sub make_witness_path {
2166 my( $self, $wit ) = @_;
2167 my @chain = @{$wit->path};
2168 my $sig = $wit->sigil;
2169 # Add start and end if necessary
2170 unshift( @chain, $self->start ) unless $chain[0] eq $self->start;
2171 push( @chain, $self->end ) unless $chain[-1] eq $self->end;
2172 foreach my $idx ( 0 .. $#chain-1 ) {
2173 $self->add_path( $chain[$idx], $chain[$idx+1], $sig );
2175 if( $wit->is_layered ) {
2176 @chain = @{$wit->uncorrected_path};
2177 unshift( @chain, $self->start ) unless $chain[0] eq $self->start;
2178 push( @chain, $self->end ) unless $chain[-1] eq $self->end;
2179 foreach my $idx( 0 .. $#chain-1 ) {
2180 my $source = $chain[$idx];
2181 my $target = $chain[$idx+1];
2182 $self->add_path( $source, $target, $sig.$self->ac_label )
2183 unless $self->has_path( $source, $target, $sig );
2187 $wit->clear_uncorrected_path;
2190 =head2 calculate_ranks
2192 Calculate the reading ranks (that is, their aligned positions relative
2193 to each other) for the graph. This can only be called on linear collations.
2197 use Text::Tradition;
2199 my $cxfile = 't/data/Collatex-16.xml';
2200 my $t = Text::Tradition->new(
2202 'input' => 'CollateX',
2205 my $c = $t->collation;
2208 my $table = $c->alignment_table;
2209 ok( $c->has_cached_table, "Alignment table was cached" );
2210 is( $c->alignment_table, $table, "Cached table returned upon second call" );
2211 $c->calculate_ranks;
2212 is( $c->alignment_table, $table, "Cached table retained with no rank change" );
2213 $c->add_relationship( 'n13', 'n23', { type => 'repetition' } );
2214 is( $c->alignment_table, $table, "Alignment table unchanged after non-colo relationship add" );
2215 $c->add_relationship( 'n24', 'n23', { type => 'spelling' } );
2216 isnt( $c->alignment_table, $table, "Alignment table changed after colo relationship add" );
2222 sub calculate_ranks {
2224 # Save the existing ranks, in case we need to invalidate the cached SVG.
2225 throw( "Cannot calculate ranks on a non-linear graph" )
2226 unless $self->linear;
2228 map { $existing_ranks{$_} = $_->rank } $self->readings;
2230 # Do the rankings based on the relationship equivalence graph, starting
2231 # with the start node.
2232 my ( $node_ranks, $rank_nodes ) = $self->relations->equivalence_ranks();
2234 # Transfer our rankings from the topological graph to the real one.
2235 foreach my $r ( $self->readings ) {
2236 if( defined $node_ranks->{$self->equivalence( $r->id )} ) {
2237 $r->rank( $node_ranks->{$self->equivalence( $r->id )} );
2239 # Die. Find the last rank we calculated.
2240 my @all_defined = sort { ( $node_ranks->{$self->equivalence( $a->id )}||-1 )
2241 <=> ( $node_ranks->{$self->equivalence( $b->id )}||-1 ) }
2243 my $last = pop @all_defined;
2244 throw( "Ranks not calculated after $last - do you have a cycle in the graph?" );
2247 # Do we need to invalidate the cached data?
2248 if( $self->has_cached_table ) {
2249 foreach my $r ( $self->readings ) {
2250 next if defined( $existing_ranks{$r} )
2251 && $existing_ranks{$r} == $r->rank;
2252 # Something has changed, so clear the cache
2253 $self->_clear_cache;
2254 # ...and recalculate the common readings.
2255 $self->calculate_common_readings();
2259 # The graph calculation information is now up to date.
2260 $self->_graphcalc_done(1);
2265 $self->wipe_table if $self->has_cached_table;
2269 =head2 flatten_ranks
2271 A convenience method for parsing collation data. Searches the graph for readings
2272 with the same text at the same rank, and merges any that are found.
2277 my ( $self, %args ) = shift;
2278 my %unique_rank_rdg;
2280 foreach my $p ( $self->identical_readings( %args ) ) {
2281 # say STDERR "Combining readings at same rank: @$p";
2283 $self->merge_readings( @$p );
2284 # TODO see if this now makes a common point.
2286 # If we merged readings, the ranks are still fine but the alignment
2287 # table is wrong. Wipe it.
2288 $self->wipe_table() if $changed;
2291 =head2 identical_readings
2292 =head2 identical_readings( start => $startnode, end => $endnode )
2293 =head2 identical_readings( startrank => $startrank, endrank => $endrank )
2295 Goes through the graph identifying all pairs of readings that appear to be
2296 identical, and therefore able to be merged into a single reading. Returns the
2297 relevant identical pairs. Can be restricted to run over only a part of the
2298 graph, specified either by node or by rank.
2302 sub identical_readings {
2303 my ( $self, %args ) = @_;
2304 # Find where we should start and end.
2305 my $startrank = $args{startrank} || 0;
2306 if( $args{start} ) {
2307 throw( "Starting reading has no rank" ) unless $self->reading( $args{start} )
2308 && $self->reading( $args{start} )->has_rank;
2309 $startrank = $self->reading( $args{start} )->rank;
2311 my $endrank = $args{endrank} || $self->end->rank;
2313 throw( "Ending reading has no rank" ) unless $self->reading( $args{end} )
2314 && $self->reading( $args{end} )->has_rank;
2315 $endrank = $self->reading( $args{end} )->rank;
2318 # Make sure the ranks are correct.
2319 unless( $self->_graphcalc_done ) {
2320 $self->calculate_ranks;
2322 # Go through the readings looking for duplicates.
2323 my %unique_rank_rdg;
2325 foreach my $rdg ( $self->readings ) {
2326 next unless $rdg->has_rank;
2327 my $rk = $rdg->rank;
2328 next if $rk > $endrank || $rk < $startrank;
2329 my $key = $rk . "||" . $rdg->text;
2330 if( exists $unique_rank_rdg{$key} ) {
2331 # Make sure they don't have different grammatical forms
2332 my $ur = $unique_rank_rdg{$key};
2333 if( $rdg->is_identical( $ur ) ) {
2334 push( @pairs, [ $ur, $rdg ] );
2337 $unique_rank_rdg{$key} = $rdg;
2345 =head2 calculate_common_readings
2347 Goes through the graph identifying the readings that appear in every witness
2348 (apart from those with lacunae at that spot.) Marks them as common and returns
2353 use Text::Tradition;
2355 my $cxfile = 't/data/Collatex-16.xml';
2356 my $t = Text::Tradition->new(
2358 'input' => 'CollateX',
2361 my $c = $t->collation;
2363 my @common = $c->calculate_common_readings();
2364 is( scalar @common, 8, "Found correct number of common readings" );
2365 my @marked = sort $c->common_readings();
2366 is( scalar @common, 8, "All common readings got marked as such" );
2367 my @expected = qw/ n1 n11 n16 n19 n20 n5 n6 n7 /;
2368 is_deeply( \@marked, \@expected, "Found correct list of common readings" );
2374 sub calculate_common_readings {
2377 map { $_->is_common( 0 ) } $self->readings;
2378 # Implicitly calls calculate_ranks
2379 my $table = $self->alignment_table;
2380 foreach my $idx ( 0 .. $table->{'length'} - 1 ) {
2381 my @row = map { $_->{'tokens'}->[$idx]
2382 ? $_->{'tokens'}->[$idx]->{'t'} : '' }
2383 @{$table->{'alignment'}};
2385 foreach my $r ( @row ) {
2387 $hash{$r->id} = $r unless $r->is_meta;
2389 $hash{'UNDEF'} = $r;
2392 if( keys %hash == 1 && !exists $hash{'UNDEF'} ) {
2393 my( $r ) = values %hash;
2395 push( @common, $r );
2401 =head2 text_from_paths
2403 Calculate the text array for all witnesses from the path, for later consistency
2404 checking. Only to be used if there is no non-graph-based way to know the
2409 sub text_from_paths {
2411 foreach my $wit ( $self->tradition->witnesses ) {
2412 my @readings = $self->reading_sequence( $self->start, $self->end, $wit->sigil );
2414 foreach my $r ( @readings ) {
2415 next if $r->is_meta;
2416 push( @text, $r->text );
2418 $wit->text( \@text );
2419 if( $wit->is_layered ) {
2420 my @ucrdgs = $self->reading_sequence( $self->start, $self->end,
2421 $wit->sigil.$self->ac_label );
2423 foreach my $r ( @ucrdgs ) {
2424 next if $r->is_meta;
2425 push( @uctext, $r->text );
2427 $wit->layertext( \@uctext );
2432 =head1 UTILITY FUNCTIONS
2434 =head2 common_predecessor( $reading_a, $reading_b )
2436 Find the last reading that occurs in sequence before both the given readings.
2437 At the very least this should be $self->start.
2439 =head2 common_successor( $reading_a, $reading_b )
2441 Find the first reading that occurs in sequence after both the given readings.
2442 At the very least this should be $self->end.
2446 use Text::Tradition;
2448 my $cxfile = 't/data/Collatex-16.xml';
2449 my $t = Text::Tradition->new(
2451 'input' => 'CollateX',
2454 my $c = $t->collation;
2456 is( $c->common_predecessor( 'n24', 'n23' )->id,
2457 'n20', "Found correct common predecessor" );
2458 is( $c->common_successor( 'n24', 'n23' )->id,
2459 '__END__', "Found correct common successor" );
2461 is( $c->common_predecessor( 'n19', 'n17' )->id,
2462 'n16', "Found correct common predecessor for readings on same path" );
2463 is( $c->common_successor( 'n21', 'n10' )->id,
2464 '__END__', "Found correct common successor for readings on same path" );
2470 ## Return the closest reading that is a predecessor of both the given readings.
2471 sub common_predecessor {
2473 my( $r1, $r2 ) = $self->_objectify_args( @_ );
2474 return $self->_common_in_path( $r1, $r2, 'predecessors' );
2477 sub common_successor {
2479 my( $r1, $r2 ) = $self->_objectify_args( @_ );
2480 return $self->_common_in_path( $r1, $r2, 'successors' );
2484 # TODO think about how to do this without ranks...
2485 sub _common_in_path {
2486 my( $self, $r1, $r2, $dir ) = @_;
2487 my $iter = $self->end->rank;
2489 my @last_r1 = ( $r1 );
2490 my @last_r2 = ( $r2 );
2491 # my %all_seen = ( $r1 => 'r1', $r2 => 'r2' );
2493 # say STDERR "Finding common $dir for $r1, $r2";
2494 while( !@candidates ) {
2495 last unless $iter--; # Avoid looping infinitely
2496 # Iterate separately down the graph from r1 and r2
2497 my( @new_lc1, @new_lc2 );
2498 foreach my $lc ( @last_r1 ) {
2499 foreach my $p ( $lc->$dir ) {
2500 if( $all_seen{$p->id} && $all_seen{$p->id} ne 'r1' ) {
2501 # say STDERR "Path candidate $p from $lc";
2502 push( @candidates, $p );
2503 } elsif( !$all_seen{$p->id} ) {
2504 $all_seen{$p->id} = 'r1';
2505 push( @new_lc1, $p );
2509 foreach my $lc ( @last_r2 ) {
2510 foreach my $p ( $lc->$dir ) {
2511 if( $all_seen{$p->id} && $all_seen{$p->id} ne 'r2' ) {
2512 # say STDERR "Path candidate $p from $lc";
2513 push( @candidates, $p );
2514 } elsif( !$all_seen{$p->id} ) {
2515 $all_seen{$p->id} = 'r2';
2516 push( @new_lc2, $p );
2520 @last_r1 = @new_lc1;
2521 @last_r2 = @new_lc2;
2523 my @answer = sort { $a->rank <=> $b->rank } @candidates;
2524 return $dir eq 'predecessors' ? pop( @answer ) : shift ( @answer );
2528 Text::Tradition::Error->throw(
2529 'ident' => 'Collation error',
2535 __PACKAGE__->meta->make_immutable;
2541 =item * Rework XML serialization in a more modular way
2547 This package is free software and is provided "as is" without express
2548 or implied warranty. You can redistribute it and/or modify it under
2549 the same terms as Perl itself.
2553 Tara L Andrews E<lt>aurum@cpan.orgE<gt>