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',
61 isa => 'Text::Tradition',
62 writer => '_set_tradition',
70 Text::Tradition::Collation - a software model for a text collation
75 my $t = Text::Tradition->new(
76 'name' => 'this is a text',
78 'file' => '/path/to/tei_parallel_seg_file.xml' );
80 my $c = $t->collation;
81 my @readings = $c->readings;
82 my @paths = $c->paths;
83 my @relationships = $c->relationships;
85 my $svg_variant_graph = $t->collation->as_svg();
89 Text::Tradition is a library for representation and analysis of collated
90 texts, particularly medieval ones. The Collation is the central feature of
91 a Tradition, where the text, its sequence of readings, and its relationships
92 between readings are actually kept.
98 The constructor. Takes a hash or hashref of the following arguments:
102 =item * tradition - The Text::Tradition object to which the collation
105 =item * linear - Whether the collation should be linear; that is, whether
106 transposed readings should be treated as two linked readings rather than one,
107 and therefore whether the collation graph is acyclic. Defaults to true.
109 =item * baselabel - The default label for the path taken by a base text
110 (if any). Defaults to 'base text'.
112 =item * wit_list_separator - The string to join a list of witnesses for
113 purposes of making labels in display graphs. Defaults to ', '.
115 =item * ac_label - The extra label to tack onto a witness sigil when
116 representing another layer of path for the given witness - that is, when
117 a text has more than one possible reading due to scribal corrections or
118 the like. Defaults to ' (a.c.)'.
120 =item * wordsep - The string used to separate words in the original text.
131 =head2 wit_list_separator
139 Simple accessors for collation attributes.
143 The meta-reading at the start of every witness path.
147 The meta-reading at the end of every witness path.
151 Returns all Reading objects in the graph.
153 =head2 reading( $id )
155 Returns the Reading object corresponding to the given ID.
157 =head2 add_reading( $reading_args )
159 Adds a new reading object to the collation.
160 See L<Text::Tradition::Collation::Reading> for the available arguments.
162 =head2 del_reading( $object_or_id )
164 Removes the given reading from the collation, implicitly removing its
165 paths and relationships.
167 =head2 has_reading( $id )
169 Predicate to see whether a given reading ID is in the graph.
171 =head2 reading_witnesses( $object_or_id )
173 Returns a list of sigils whose witnesses contain the reading.
177 Returns all reading paths within the document - that is, all edges in the
178 collation graph. Each path is an arrayref of [ $source, $target ] reading IDs.
180 =head2 add_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 del_path( $source, $target, $sigil )
187 Links the given readings in the collation in sequence, under the given witness
188 sigil. The readings may be specified by object or ID.
190 =head2 has_path( $source, $target );
192 Returns true if the two readings are linked in sequence in any witness.
193 The readings may be specified by object or ID.
197 Returns all Relationship objects in the collation.
199 =head2 add_relationship( $reading, $other_reading, $options, $changed_readings )
201 Adds a new relationship of the type given in $options between the two readings,
202 which may be specified by object or ID. Returns a value of ( $status, @vectors)
203 where $status is true on success, and @vectors is a list of relationship edges
204 that were ultimately added. If an array reference is passed in as $changed_readings,
205 then any readings that were altered due to the relationship creation are added to
208 See L<Text::Tradition::Collation::Relationship> for the available options.
213 my ( $class, @args ) = @_;
214 my %args = @args == 1 ? %{ $args[0] } : @args;
215 # TODO determine these from the Moose::Meta object
216 my @delegate_attrs = qw(sequence relations readings wit_list_separator baselabel
217 linear wordsep direction start end cached_table _graphcalc_done);
219 for my $attr (@delegate_attrs) {
220 $data_args{$attr} = delete $args{$attr} if exists $args{$attr};
222 $args{_data} = Text::Tradition::Collation::Data->new(%data_args);
228 $self->_set_relations( Text::Tradition::Collation::RelationshipStore->new( 'collation' => $self ) );
229 $self->_set_start( $self->add_reading(
230 { 'collation' => $self, 'is_start' => 1, 'init' => 1 } ) );
231 $self->_set_end( $self->add_reading(
232 { 'collation' => $self, 'is_end' => 1, 'init' => 1 } ) );
235 =head2 register_relationship_type( %relationship_definition )
237 Add a relationship type definition to this collation. The argument can be either a
238 hash or a hashref, defining the properties of the relationship. For relationship types
239 and their properties, see L<Text::Tradition::Collation::RelationshipType>.
241 =head2 get_relationship_type( $relationship_name )
243 Retrieve the RelationshipType object for the relationship with the given name.
247 sub register_relationship_type {
249 my %args = @_ == 1 ? %{$_[0]} : @_;
250 if( $self->relations->has_type( $args{name} ) ) {
251 throw( 'Relationship type ' . $args{name} . ' already registered' );
253 $self->relations->add_type( %args );
256 sub get_relationship_type {
257 my( $self, $name ) = @_;
258 return $self->relations->has_type( $name )
259 ? $self->relations->type( $name ) : undef;
262 ### Reading construct/destruct functions
265 my( $self, $reading ) = @_;
266 unless( ref( $reading ) eq 'Text::Tradition::Collation::Reading' ) {
267 my %args = %$reading;
268 if( $args{'init'} ) {
269 # If we are initializing an empty collation, don't assume that we
270 # have set a tradition.
271 delete $args{'init'};
272 } elsif( $self->tradition->can('language') && $self->tradition->has_language
273 && !exists $args{'language'} ) {
274 $args{'language'} = $self->tradition->language;
276 $reading = Text::Tradition::Collation::Reading->new(
277 'collation' => $self,
280 # First check to see if a reading with this ID exists.
281 if( $self->reading( $reading->id ) ) {
282 throw( "Collation already has a reading with id " . $reading->id );
284 $self->_graphcalc_done(0);
285 $self->_add_reading( $reading->id => $reading );
286 # Once the reading has been added, put it in both graphs.
287 $self->sequence->add_vertex( $reading->id );
288 $self->relations->add_reading( $reading->id );
292 around del_reading => sub {
297 if( ref( $arg ) eq 'Text::Tradition::Collation::Reading' ) {
300 # Remove the reading from the graphs.
301 $self->_graphcalc_done(0);
302 $self->_clear_cache; # Explicitly clear caches to GC the reading
303 $self->sequence->delete_vertex( $arg );
304 $self->relations->delete_reading( $arg );
307 $self->$orig( $arg );
310 =head2 merge_readings( $main, $second, $concatenate, $with_str )
312 Merges the $second reading into the $main one. If $concatenate is true, then
313 the merged node will carry the text of both readings, concatenated with either
314 $with_str (if specified) or a sensible default (the empty string if the
315 appropriate 'join_*' flag is set on either reading, or else $self->wordsep.)
317 The first two arguments may be either readings or reading IDs.
324 my $cxfile = 't/data/Collatex-16.xml';
325 my $t = Text::Tradition->new(
327 'input' => 'CollateX',
330 my $c = $t->collation;
332 my $rno = scalar $c->readings;
333 # Split n21 ('unto') for testing purposes
334 my $new_r = $c->add_reading( { 'id' => 'n21p0', 'text' => 'un', 'join_next' => 1 } );
335 my $old_r = $c->reading( 'n21' );
336 $old_r->alter_text( 'to' );
337 $c->del_path( 'n20', 'n21', 'A' );
338 $c->add_path( 'n20', 'n21p0', 'A' );
339 $c->add_path( 'n21p0', 'n21', 'A' );
340 $c->add_relationship( 'n21', 'n22', { type => 'collated', scope => 'local' } );
342 ok( $c->reading( 'n21p0' ), "New reading exists" );
343 is( scalar $c->readings, $rno, "Reading add offset by flatten_ranks" );
345 # Combine n3 and n4 ( with his )
346 $c->merge_readings( 'n3', 'n4', 1 );
347 ok( !$c->reading('n4'), "Reading n4 is gone" );
348 is( $c->reading('n3')->text, 'with his', "Reading n3 has both words" );
350 # Collapse n9 and n10 ( rood / root )
351 $c->merge_readings( 'n9', 'n10' );
352 ok( !$c->reading('n10'), "Reading n10 is gone" );
353 is( $c->reading('n9')->text, 'rood', "Reading n9 has an unchanged word" );
355 # Try to combine n21 and n21p0. This should break.
356 my $remaining = $c->reading('n21');
357 $remaining ||= $c->reading('n22'); # one of these should still exist
359 $c->merge_readings( 'n21p0', $remaining, 1 );
360 ok( 0, "Bad reading merge changed the graph" );
361 } catch( Text::Tradition::Error $e ) {
362 like( $e->message, qr/neither concatenated nor collated/, "Expected exception from bad concatenation" );
364 ok( 0, "Unexpected error on bad reading merge: $@" );
368 $c->calculate_ranks();
369 ok( 1, "Graph is still evidently whole" );
370 } catch( Text::Tradition::Error $e ) {
371 ok( 0, "Caught a rank exception: " . $e->message );
382 my( $kept_obj, $del_obj, $combine, $combine_char ) = $self->_objectify_args( @_ );
383 my $mergemeta = $kept_obj->is_meta;
384 throw( "Cannot merge meta and non-meta reading" )
385 unless ( $mergemeta && $del_obj->is_meta )
386 || ( !$mergemeta && !$del_obj->is_meta );
388 throw( "Cannot merge with start or end node" )
389 if( $kept_obj eq $self->start || $kept_obj eq $self->end
390 || $del_obj eq $self->start || $del_obj eq $self->end );
391 throw( "Cannot combine text of meta readings" ) if $combine;
393 # We can only merge readings in a linear graph if:
394 # - they are contiguous with only one edge between them, OR
395 # - they are at equivalent ranks in the graph.
396 if( $self->linear ) {
397 my @delpred = $del_obj->predecessors;
398 my @keptsuc = $kept_obj->successors;
399 unless ( @delpred == 1 && $delpred[0] eq $kept_obj
400 && @keptsuc == 1 && $keptsuc[0] eq $del_obj ) {
401 my( $is_ok, $msg ) = $self->relations->relationship_valid(
402 $kept_obj, $del_obj, 'collated' );
404 throw( "Readings $kept_obj and $del_obj can be neither concatenated nor collated" );
409 # We only need the IDs for adding paths to the graph, not the reading
410 # objects themselves.
411 my $kept = $kept_obj->id;
412 my $deleted = $del_obj->id;
413 $self->_graphcalc_done(0);
415 # The kept reading should inherit the paths and the relationships
416 # of the deleted reading.
417 foreach my $path ( $self->sequence->edges_at( $deleted ) ) {
418 my @vector = ( $kept );
419 push( @vector, $path->[1] ) if $path->[0] eq $deleted;
420 unshift( @vector, $path->[0] ) if $path->[1] eq $deleted;
421 next if $vector[0] eq $vector[1]; # Don't add a self loop
422 my %wits = %{$self->sequence->get_edge_attributes( @$path )};
423 $self->sequence->add_edge( @vector );
424 my $fwits = $self->sequence->get_edge_attributes( @vector );
425 @wits{keys %$fwits} = values %$fwits;
426 $self->sequence->set_edge_attributes( @vector, \%wits );
428 $self->relations->merge_readings( $kept, $deleted, $combine );
430 # Do the deletion deed.
432 # Combine the text of the readings
433 my $joinstr = $combine_char;
434 unless( defined $joinstr ) {
435 $joinstr = '' if $kept_obj->join_next || $del_obj->join_prior;
436 $joinstr = $self->wordsep unless defined $joinstr;
438 $kept_obj->_combine( $del_obj, $joinstr );
440 $self->del_reading( $deleted );
443 =head2 merge_related( @relationship_types )
445 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.
447 WARNING: This operation cannot be undone.
459 $t = Text::Tradition->new( 'input' => 'Self', 'file' => 't/data/legendfrag.xml' );
460 } [qr/Cannot set relationship on a meta reading/],
461 "Got expected relationship drop warning on parse";
463 my $c = $t->collation;
464 # Force the transitive propagation of all existing relationships.
465 $c->relations->propagate_all_relationships();
468 map { $rdg_ids{$_} = 1 } $c->readings;
469 $c->merge_related( 'orthographic' );
470 is( scalar( $c->readings ), keys( %rdg_ids ) - 9,
471 "Successfully collapsed orthographic variation" );
472 map { $rdg_ids{$_} = undef } qw/ r13.3 r11.4 r8.5 r8.2 r7.7 r7.5 r7.4 r7.3 r7.1 /;
473 foreach my $rid ( keys %rdg_ids ) {
474 my $exp = $rdg_ids{$rid};
475 is( !$c->reading( $rid ), !$exp, "Reading $rid correctly " .
476 ( $exp ? "retained" : "removed" ) );
478 ok( $c->linear, "Graph is still linear" );
480 $c->calculate_ranks; # This should succeed
481 ok( 1, "Can still calculate ranks on the new graph" );
483 ok( 0, "Rank calculation on merged graph failed: $@" );
486 # Now add some transpositions
487 $c->add_relationship( 'r8.4', 'r10.4', { type => 'transposition' } );
488 $c->merge_related( 'transposition' );
489 is( scalar( $c->readings ), keys( %rdg_ids ) - 10,
490 "Transposed relationship is merged away" );
491 ok( !$c->reading('r8.4'), "Correct transposed reading removed" );
492 ok( !$c->linear, "Graph is no longer linear" );
494 $c->calculate_ranks; # This should fail
495 ok( 0, "Rank calculation happened on nonlinear graph?!" );
496 } catch ( Text::Tradition::Error $e ) {
497 is( $e->message, 'Cannot calculate ranks on a non-linear graph',
498 "Rank calculation on merged graph threw an error" );
505 # TODO: there should be a way to display merged without affecting the underlying data!
510 map { $reltypehash{$_} = 1 } @_;
512 # Set up the filter for finding related readings
514 exists $reltypehash{$_[0]->type};
517 # Go through all readings looking for related ones
518 foreach my $r ( $self->readings ) {
519 next unless $self->reading( "$r" ); # might have been deleted meanwhile
520 while( my @related = $self->related_readings( $r, $filter ) ) {
521 push( @related, $r );
523 scalar $b->witnesses <=> scalar $a->witnesses
525 my $keep = shift @related;
526 foreach my $delr ( @related ) {
528 unless( $self->get_relationship( $keep, $delr )->colocated );
529 $self->merge_readings( $keep, $delr );
535 =head2 compress_readings
537 Where possible in the graph, compresses plain sequences of readings into a
538 single reading. The sequences must consist of readings with no
539 relationships to other readings, with only a single witness path between
540 them and no other witness paths from either that would skip the other. The
541 readings must also not be marked as nonsense or bad grammar.
543 WARNING: This operation cannot be undone.
549 my $t = Text::Tradition->new( input => 'CollateX', file => 't/data/Collatex-16.xml' );
550 my $c = $t->collation;
551 my $n = scalar $c->readings;
552 $c->compress_readings();
553 is( scalar $c->readings, $n - 6, "Compressing readings seems to work" );
555 # Now put in a join-word and make sure the thing still works.
556 my $t2 = Text::Tradition->new( input => 'CollateX', file => 't/data/Collatex-16.xml' );
557 my $c2 = $t2->collation;
558 # Split n21 ('unto') for testing purposes
559 my $new_r = $c2->add_reading( { 'id' => 'n21p0', 'text' => 'un', 'join_next' => 1 } );
560 my $old_r = $c2->reading( 'n21' );
561 $old_r->alter_text( 'to' );
562 $c2->del_path( 'n20', 'n21', 'A' );
563 $c2->add_path( 'n20', 'n21p0', 'A' );
564 $c2->add_path( 'n21p0', 'n21', 'A' );
565 $c2->calculate_ranks();
566 is( scalar $c2->readings, $n + 1, "We have our extra test reading" );
567 $c2->compress_readings();
568 is( scalar $c2->readings, $n - 6, "Compressing readings also works with join_next" );
569 is( $c2->reading( 'n21p0' )->text, 'unto', "The joined word has no space" );
576 sub compress_readings {
578 # Sanity check: first save the original text of each witness.
580 foreach my $wit ( $self->tradition->witnesses ) {
581 $origtext{$wit->sigil} = $self->path_text( $wit->sigil );
582 if( $wit->is_layered ) {
583 my $acsig = $wit->sigil . $self->ac_label;
584 $origtext{$acsig} = $self->path_text( $acsig );
589 # Anywhere in the graph that there is a reading that joins only to a single
590 # successor, and neither of these have any relationships, just join the two
592 foreach my $rdg ( sort { $a->rank <=> $b->rank } $self->readings ) {
593 # Now look for readings that can be joined to their successors.
594 next unless $rdg->is_combinable;
596 while( $self->sequence->successors( $rdg ) == 1 ) {
597 my( $next ) = $self->reading( $self->sequence->successors( $rdg ) );
598 throw( "Infinite loop" ) if $seen{$next->id};
599 $seen{$next->id} = 1;
600 last if $self->sequence->predecessors( $next ) > 1;
601 last unless $next->is_combinable;
602 say "Joining readings $rdg and $next";
603 $self->merge_readings( $rdg, $next, 1 );
607 # Finally, make sure we haven't screwed anything up.
608 foreach my $wit ( $self->tradition->witnesses ) {
609 my $pathtext = $self->path_text( $wit->sigil );
610 throw( "Text differs for witness " . $wit->sigil )
611 unless $pathtext eq $origtext{$wit->sigil};
612 if( $wit->is_layered ) {
613 my $acsig = $wit->sigil . $self->ac_label;
614 $pathtext = $self->path_text( $acsig );
615 throw( "Layered text differs for witness " . $wit->sigil )
616 unless $pathtext eq $origtext{$acsig};
620 $self->relations->rebuild_equivalence();
621 $self->calculate_ranks();
624 # Helper function for manipulating the graph.
625 sub _stringify_args {
626 my( $self, $first, $second, @args ) = @_;
628 if ref( $first ) eq 'Text::Tradition::Collation::Reading';
629 $second = $second->id
630 if ref( $second ) eq 'Text::Tradition::Collation::Reading';
631 return( $first, $second, @args );
634 # Helper function for manipulating the graph.
635 sub _objectify_args {
636 my( $self, $first, $second, $arg ) = @_;
637 $first = $self->reading( $first )
638 unless ref( $first ) eq 'Text::Tradition::Collation::Reading';
639 $second = $self->reading( $second )
640 unless ref( $second ) eq 'Text::Tradition::Collation::Reading';
641 return( $first, $second, $arg );
644 =head2 duplicate_reading( $reading, @witlist )
646 Split the given reading into two, so that the new reading is in the path for
647 the witnesses given in @witlist. If the result is that certain non-colocated
648 relationships (e.g. transpositions) are no longer valid, these will be removed.
649 Returns the newly-created reading.
653 use Test::More::UTF8;
657 my $st = Text::Tradition->new( 'input' => 'Self', 'file' => 't/data/collatecorr.xml' );
658 is( ref( $st ), 'Text::Tradition', "Got a tradition from test file" );
659 ok( $st->has_witness('Ba96'), "Tradition has the affected witness" );
661 my $sc = $st->collation;
663 ok( $sc->reading('n131'), "Tradition has the affected reading" );
664 is( scalar( $sc->readings ), $numr, "There are $numr readings in the graph" );
665 is( $sc->end->rank, 14, "There are fourteen ranks in the graph" );
667 # Detach the erroneously collated reading
668 my( $newr, @del_rdgs ) = $sc->duplicate_reading( 'n131', 'Ba96' );
669 ok( $newr, "New reading was created" );
670 ok( $sc->reading('n131_0'), "Detached the bad collation with a new reading" );
671 is( scalar( $sc->readings ), $numr + 1, "A reading was added to the graph" );
672 is( $sc->end->rank, 10, "There are now only ten ranks in the graph" );
673 my $csucc = $sc->common_successor( 'n131', 'n131_0' );
674 is( $csucc->id, 'n136', "Found correct common successor to duped reading" );
676 # Check that the bad transposition is gone
677 is( scalar @del_rdgs, 1, "Deleted reading was returned by API call" );
678 is( $sc->get_relationship( 'n130', 'n135' ), undef, "Bad transposition relationship is gone" );
680 # The collation should not be fixed
681 my @pairs = $sc->identical_readings();
682 is( scalar @pairs, 0, "Not re-collated yet" );
684 ok( $sc->merge_readings( 'n124', 'n131_0' ), "Collated the readings correctly" );
685 @pairs = $sc->identical_readings( start => 'n124', end => $csucc->id );
686 is( scalar @pairs, 3, "Found three more identical readings" );
687 is( $sc->end->rank, 11, "The ranks shifted appropriately" );
688 $sc->flatten_ranks();
689 is( scalar( $sc->readings ), $numr - 3, "Now we are collated correctly" );
691 # Check that we can't "duplicate" a reading with no wits or with all wits
693 my( $badr, @del_rdgs ) = $sc->duplicate_reading( 'n124' );
694 ok( 0, "Reading duplication without witnesses throws an error" );
695 } catch( Text::Tradition::Error $e ) {
696 like( $e->message, qr/Must specify one or more witnesses/,
697 "Reading duplication without witnesses throws the expected error" );
699 ok( 0, "Reading duplication without witnesses threw the wrong error" );
703 my( $badr, @del_rdgs ) = $sc->duplicate_reading( 'n124', 'Ba96', 'Mü11475' );
704 ok( 0, "Reading duplication with all witnesses throws an error" );
705 } catch( Text::Tradition::Error $e ) {
706 like( $e->message, qr/Cannot join all witnesses/,
707 "Reading duplication with all witnesses throws the expected error" );
709 ok( 0, "Reading duplication with all witnesses threw the wrong error" );
713 $sc->calculate_ranks();
714 ok( 1, "Graph is still evidently whole" );
715 } catch( Text::Tradition::Error $e ) {
716 ok( 0, "Caught a rank exception: " . $e->message );
723 sub duplicate_reading {
724 my( $self, $r, @wits ) = @_;
725 # Check that we are not doing anything unwise.
726 throw( "Must specify one or more witnesses for the duplicated reading" )
728 unless( ref( $r ) eq 'Text::Tradition::Collation::Reading' ) {
729 $r = $self->reading( $r );
731 throw( "Cannot duplicate a meta-reading" )
733 throw( "Cannot join all witnesses to the new reading" )
734 if scalar( @wits ) == scalar( $r->witnesses );
736 # Get all the reading attributes and duplicate them.
737 my $rmeta = Text::Tradition::Collation::Reading->meta;
739 foreach my $attr( $rmeta->get_all_attributes ) {
740 next if $attr->name =~ /^_/;
741 my $acc = $attr->get_read_method;
742 if( !$acc && $attr->has_applied_traits ) {
743 my $tr = $attr->applied_traits;
744 if( $tr->[0] =~ /::(Array|Hash)$/ ) {
746 my %methods = reverse %{$attr->handles};
747 $acc = $methods{elements};
748 $args{$attr->name} = $which eq 'Array'
749 ? [ $r->$acc ] : { $r->$acc };
752 my $attrval = $r->$acc;
753 if( defined $attrval ) {
754 $args{$attr->name} = $attrval;
758 # By definition the new reading will no longer be common.
759 $args{is_common} = 0;
760 # The new reading also needs its own ID.
761 $args{id} = $self->_generate_dup_id( $r->id );
763 # Try to make the new reading.
764 my $newr = $self->add_reading( \%args );
765 # The old reading is also no longer common.
768 # For each of the witnesses, dissociate from the old reading and
769 # associate with the new.
770 foreach my $wit ( @wits ) {
771 my $prior = $self->prior_reading( $r, $wit );
772 my $next = $self->next_reading( $r, $wit );
773 $self->del_path( $prior, $r, $wit );
774 $self->add_path( $prior, $newr, $wit );
775 $self->del_path( $r, $next, $wit );
776 $self->add_path( $newr, $next, $wit );
779 # If the graph is ranked, we need to look for relationships that are now
780 # invalid (i.e. 'non-colocation' types that might now be colocated) and
781 # remove them. If not, we can skip it.
784 my @deleted_relations;
785 if( $self->end->has_rank ) {
786 # Find the point where we can stop checking
787 $succ = $self->common_successor( $r, $newr );
789 # Hash the existing ranks
790 foreach my $rdg ( $self->readings ) {
791 $rrk{$rdg->id} = $rdg->rank;
793 # Calculate the new ranks
794 $self->calculate_ranks();
796 # Check for invalid non-colocated relationships among changed-rank readings
797 # from where the ranks start changing up to $succ
798 my $lastrank = $succ->rank;
799 foreach my $rdg ( $self->readings ) {
800 next if $rdg->rank > $lastrank;
801 next if $rdg->rank == $rrk{$rdg->id};
802 my @noncolo = $rdg->related_readings( sub { !$_[0]->colocated } );
803 next unless @noncolo;
804 foreach my $nc ( @noncolo ) {
805 unless( $self->relations->verify_or_delete( $rdg, $nc ) ) {
806 push( @deleted_relations, [ $rdg->id, $nc->id ] );
811 return ( $newr, @deleted_relations );
814 sub _generate_dup_id {
815 my( $self, $rid ) = @_;
820 if( $self->has_reading( $newid ) ) {
833 # We only need the IDs for adding paths to the graph, not the reading
834 # objects themselves.
835 my( $source, $target, $wit ) = $self->_stringify_args( @_ );
837 $self->_graphcalc_done(0);
838 # Connect the readings
839 unless( $self->sequence->has_edge( $source, $target ) ) {
840 $self->sequence->add_edge( $source, $target );
841 $self->relations->add_equivalence_edge( $source, $target );
843 # Note the witness in question
844 $self->sequence->set_edge_attribute( $source, $target, $wit, 1 );
850 if( ref( $_[0] ) eq 'ARRAY' ) {
857 # We only need the IDs for removing paths from the graph, not the reading
858 # objects themselves.
859 my( $source, $target, $wit ) = $self->_stringify_args( @args );
861 $self->_graphcalc_done(0);
862 if( $self->sequence->has_edge_attribute( $source, $target, $wit ) ) {
863 $self->sequence->delete_edge_attribute( $source, $target, $wit );
865 unless( $self->sequence->has_edge_attributes( $source, $target ) ) {
866 $self->sequence->delete_edge( $source, $target );
867 $self->relations->delete_equivalence_edge( $source, $target );
872 # Extra graph-alike utility
875 my( $source, $target, $wit ) = $self->_stringify_args( @_ );
876 return undef unless $self->sequence->has_edge( $source, $target );
877 return $self->sequence->has_edge_attribute( $source, $target, $wit );
880 =head2 clear_witness( @sigil_list )
882 Clear the given witnesses out of the collation entirely, removing references
883 to them in paths, and removing readings that belong only to them. Should only
884 be called via $tradition->del_witness.
889 my( $self, @sigils ) = @_;
891 $self->_graphcalc_done(0);
892 # Clear the witness(es) out of the paths
893 foreach my $e ( $self->paths ) {
894 foreach my $sig ( @sigils ) {
895 $self->del_path( $e, $sig );
899 # Clear out the newly unused readings
900 foreach my $r ( $self->readings ) {
901 unless( $self->reading_witnesses( $r ) ) {
902 $self->del_reading( $r );
907 sub add_relationship {
909 my( $source, $target, $opts, $altered_readings ) = $self->_stringify_args( @_ );
910 my( @vectors ) = $self->relations->add_relationship( $source, $target, $opts );
912 foreach my $v ( @vectors ) {
913 my $rel = $self->get_relationship( $v );
914 next unless $rel->colocated;
915 my $r1 = $self->reading( $v->[0] );
916 my $r2 = $self->reading( $v->[1] );
917 # If it's a spelling or orthographic relationship, and one is marked
918 # as a lemma, set the normal form on the non-lemma to reflect that.
919 if( $r1->does( 'Text::Tradition::Morphology' ) ) {
920 my @changed = $r1->relationship_added( $r2, $rel );
921 if( ref( $altered_readings ) eq 'ARRAY' ) {
922 push( @$altered_readings, @changed );
926 if( $r1->has_rank && $r2->has_rank && $r1->rank ne $r2->rank ) {
927 $self->_graphcalc_done(0);
935 around qw/ get_relationship del_relationship / => sub {
939 if( @args == 1 && ref( $args[0] ) eq 'ARRAY' ) {
942 my @stringargs = $self->_stringify_args( @args );
943 $self->$orig( @stringargs );
946 =head2 reading_witnesses( $reading )
948 Return a list of sigils corresponding to the witnesses in which the reading appears.
952 sub reading_witnesses {
953 my( $self, $reading ) = @_;
954 # We need only check either the incoming or the outgoing edges; I have
955 # arbitrarily chosen "incoming". Thus, special-case the start node.
956 if( $reading eq $self->start ) {
957 return map { $_->sigil } grep { $_->is_collated } $self->tradition->witnesses;
960 foreach my $e ( $self->sequence->edges_to( $reading ) ) {
961 my $wits = $self->sequence->get_edge_attributes( @$e );
962 @all_witnesses{ keys %$wits } = 1;
964 my $acstr = $self->ac_label;
965 foreach my $acwit ( grep { $_ =~ s/^(.*)\Q$acstr\E$/$1/ } keys %all_witnesses ) {
966 delete $all_witnesses{$acwit.$acstr} if exists $all_witnesses{$acwit};
968 return keys %all_witnesses;
971 =head1 OUTPUT METHODS
973 =head2 as_svg( \%options )
975 Returns an SVG string that represents the graph, via as_dot and graphviz.
976 See as_dot for a list of options. Must have GraphViz (dot) installed to run.
981 my( $self, $opts ) = @_;
982 throw( "Need GraphViz installed to output SVG" )
983 unless File::Which::which( 'dot' );
984 my $want_subgraph = exists $opts->{'from'} || exists $opts->{'to'};
985 $self->calculate_ranks()
986 unless( $self->_graphcalc_done || $opts->{'nocalc'} || !$self->linear );
987 my @cmd = qw/dot -Tsvg/;
989 my $dotfile = File::Temp->new();
991 # $dotfile->unlink_on_destroy(0);
992 binmode $dotfile, ':utf8';
993 print $dotfile $self->as_dot( $opts );
994 push( @cmd, $dotfile->filename );
995 run( \@cmd, ">", binary(), \$svg );
996 $svg = decode_utf8( $svg );
1001 =head2 as_dot( \%options )
1003 Returns a string that is the collation graph expressed in dot
1004 (i.e. GraphViz) format. Options include:
1012 =item * color_common
1019 my( $self, $opts ) = @_;
1020 my $startrank = $opts->{'from'} if $opts;
1021 my $endrank = $opts->{'to'} if $opts;
1022 my $color_common = $opts->{'color_common'} if $opts;
1023 my $STRAIGHTENHACK = !$startrank && !$endrank && $self->end->rank
1024 && $self->end->rank > 100;
1025 $STRAIGHTENHACK = 1 if $opts->{'straight'}; # even for subgraphs or small graphs
1027 # Check the arguments
1029 return if $endrank && $startrank > $endrank;
1030 return if $startrank > $self->end->rank;
1032 if( defined $endrank ) {
1033 return if $endrank < 0;
1034 $endrank = undef if $endrank == $self->end->rank;
1037 my $graph_name = $self->tradition->name;
1038 $graph_name =~ s/[^\w\s]//g;
1039 $graph_name = join( '_', split( /\s+/, $graph_name ) );
1042 'bgcolor' => 'none',
1044 unless( $self->direction eq 'BI' ) {
1045 $graph_attrs{rankdir} = $self->direction;
1049 'fillcolor' => 'white',
1050 'style' => 'filled',
1051 'shape' => 'ellipse'
1054 'arrowhead' => 'open',
1055 'color' => '#000000',
1056 'fontcolor' => '#000000',
1059 my $dot = sprintf( "digraph %s {\n", $graph_name );
1060 $dot .= "\tgraph " . _dot_attr_string( \%graph_attrs ) . ";\n";
1061 $dot .= "\tnode " . _dot_attr_string( \%node_attrs ) . ";\n";
1063 # Output substitute start/end readings if necessary
1065 $dot .= "\t\"__SUBSTART__\" [ label=\"...\",id=\"__START__\" ];\n";
1068 $dot .= "\t\"__SUBEND__\" [ label=\"...\",id=\"__END__\" ];\n";
1070 if( $STRAIGHTENHACK ) {
1072 my $startlabel = $startrank ? '__SUBSTART__' : '__START__';
1073 $dot .= "\tsubgraph { rank=same \"$startlabel\" \"#SILENT#\" }\n";
1074 $dot .= "\t\"#SILENT#\" [ shape=diamond,color=white,penwidth=0,label=\"\" ];"
1076 my %used; # Keep track of the readings that actually appear in the graph
1077 # Sort the readings by rank if we have ranks; this speeds layout.
1078 my @all_readings = $self->end->has_rank
1079 ? sort { $a->rank <=> $b->rank } $self->readings
1081 # TODO Refrain from outputting lacuna nodes - just grey out the edges.
1082 foreach my $reading ( @all_readings ) {
1083 # Only output readings within our rank range.
1084 next if $startrank && $reading->rank < $startrank;
1085 next if $endrank && $reading->rank > $endrank;
1086 $used{$reading->id} = 1;
1087 # Need not output nodes without separate labels
1088 next if $reading->id eq $reading->text;
1090 my $label = $reading->text;
1091 unless( $label =~ /^[[:punct:]]+$/ ) {
1092 $label .= '-' if $reading->join_next;
1093 $label = "-$label" if $reading->join_prior;
1095 $label =~ s/\"/\\\"/g;
1096 $rattrs->{'label'} = $label;
1097 $rattrs->{'id'} = $reading->id;
1098 $rattrs->{'fillcolor'} = '#b3f36d' if $reading->is_common && $color_common;
1099 $dot .= sprintf( "\t\"%s\" %s;\n", $reading->id, _dot_attr_string( $rattrs ) );
1102 # Add the real edges.
1103 my @edges = $self->paths;
1104 my( %substart, %subend );
1105 foreach my $edge ( @edges ) {
1106 # Do we need to output this edge?
1107 if( $used{$edge->[0]} && $used{$edge->[1]} ) {
1108 my $label = $self->_path_display_label( $opts,
1109 $self->path_witnesses( $edge ) );
1110 my $variables = { %edge_attrs, 'label' => $label };
1112 # Account for the rank gap if necessary
1113 my $rank0 = $self->reading( $edge->[0] )->rank
1114 if $self->reading( $edge->[0] )->has_rank;
1115 my $rank1 = $self->reading( $edge->[1] )->rank
1116 if $self->reading( $edge->[1] )->has_rank;
1117 if( defined $rank0 && defined $rank1 && $rank1 - $rank0 > 1 ) {
1118 $variables->{'minlen'} = $rank1 - $rank0;
1121 # EXPERIMENTAL: make edge width reflect no. of witnesses
1122 my $extrawidth = scalar( $self->path_witnesses( $edge ) ) * 0.2;
1123 $variables->{'penwidth'} = $extrawidth + 0.8; # gives 1 for a single wit
1125 my $varopts = _dot_attr_string( $variables );
1126 $dot .= sprintf( "\t\"%s\" -> \"%s\" %s;\n",
1127 $edge->[0], $edge->[1], $varopts );
1128 } elsif( $used{$edge->[0]} ) {
1129 $subend{$edge->[0]} = $edge->[1];
1130 } elsif( $used{$edge->[1]} ) {
1131 $substart{$edge->[1]} = $edge->[0];
1135 # If we are asked to, add relationship links
1136 if( exists $opts->{show_relations} ) {
1137 my $filter = $opts->{show_relations}; # can be 'transposition' or 'all'
1138 if( $filter eq 'transposition' ) {
1139 $filter =~ qr/^transposition$/;
1142 my @types = sort( map { $_->name } $self->relations->types );
1143 if( exists $opts->{graphcolors} ) {
1144 foreach my $tdx ( 0 .. $#types ) {
1145 $typecolors{$types[$tdx]} = $opts->{graphcolors}->[$tdx];
1148 map { $typecolors{$_} = '#FFA14F' } @types;
1150 foreach my $redge ( $self->relationships ) {
1151 if( $used{$redge->[0]} && $used{$redge->[1]} ) {
1152 my $rel = $self->get_relationship( $redge );
1153 next unless $filter eq 'all' || $rel->type =~ /$filter/;
1155 arrowhead => 'none',
1156 color => $typecolors{$rel->type},
1157 constraint => 'false',
1160 unless( exists $opts->{graphcolors} ) {
1161 $variables->{label} = uc( substr( $rel->type, 0, 4 ) ),
1163 $dot .= sprintf( "\t\"%s\" -> \"%s\" %s;\n",
1164 $redge->[0], $redge->[1], _dot_attr_string( $variables ) );
1169 # Add substitute start and end edges if necessary
1170 foreach my $node ( keys %substart ) {
1171 my $witstr = $self->_path_display_label( $opts,
1172 $self->path_witnesses( $substart{$node}, $node ) );
1173 my $variables = { %edge_attrs, 'label' => $witstr };
1174 my $nrdg = $self->reading( $node );
1175 if( $nrdg->has_rank && $nrdg->rank > $startrank ) {
1176 # Substart is actually one lower than $startrank
1177 $variables->{'minlen'} = $nrdg->rank - ( $startrank - 1 );
1179 my $varopts = _dot_attr_string( $variables );
1180 $dot .= "\t\"__SUBSTART__\" -> \"$node\" $varopts;\n";
1182 foreach my $node ( keys %subend ) {
1183 my $witstr = $self->_path_display_label( $opts,
1184 $self->path_witnesses( $node, $subend{$node} ) );
1185 my $variables = { %edge_attrs, 'label' => $witstr };
1186 my $varopts = _dot_attr_string( $variables );
1187 $dot .= "\t\"$node\" -> \"__SUBEND__\" $varopts;\n";
1190 if( $STRAIGHTENHACK ) {
1191 my $endlabel = $endrank ? '__SUBEND__' : '__END__';
1192 $dot .= "\t\"$endlabel\" -> \"#SILENT#\" [ color=white,penwidth=0 ];\n";
1199 sub _dot_attr_string {
1202 foreach my $k ( sort keys %$hash ) {
1203 my $v = $hash->{$k};
1204 push( @attrs, $k.'="'.$v.'"' );
1206 return( '[ ' . join( ', ', @attrs ) . ' ]' );
1209 =head2 path_witnesses( $edge )
1211 Returns the list of sigils whose witnesses are associated with the given edge.
1212 The edge can be passed as either an array or an arrayref of ( $source, $target ).
1216 sub path_witnesses {
1217 my( $self, @edge ) = @_;
1218 # If edge is an arrayref, cope.
1219 if( @edge == 1 && ref( $edge[0] ) eq 'ARRAY' ) {
1220 my $e = shift @edge;
1223 my @wits = keys %{$self->sequence->get_edge_attributes( @edge )};
1227 # Helper function. Make a display label for the given witnesses, showing a.c.
1228 # witnesses only where the main witness is not also in the list.
1229 sub _path_display_label {
1233 map { $wits{$_} = 1 } @_;
1235 # If an a.c. wit is listed, remove it if the main wit is also listed.
1236 # Otherwise keep it for explicit listing.
1237 my $aclabel = $self->ac_label;
1239 foreach my $w ( sort keys %wits ) {
1240 if( $w =~ /^(.*)\Q$aclabel\E$/ ) {
1241 if( exists $wits{$1} ) {
1244 push( @disp_ac, $w );
1249 if( $opts->{'explicit_wits'} ) {
1250 return join( ', ', sort keys %wits );
1252 # See if we are in a majority situation.
1253 my $maj = scalar( $self->tradition->witnesses ) * 0.6;
1254 $maj = $maj > 5 ? $maj : 5;
1255 if( scalar keys %wits > $maj ) {
1256 unshift( @disp_ac, 'majority' );
1257 return join( ', ', @disp_ac );
1259 return join( ', ', sort keys %wits );
1264 =head2 as_adjacency_list
1266 Returns a JSON structure that represents the collation sequence graph.
1270 use JSON qw/ from_json /;
1271 use Text::Tradition;
1273 my $t = Text::Tradition->new(
1275 'file' => 't/data/florilegium_graphml.xml' );
1276 my $c = $t->collation;
1278 # Make a connection so we can test rank preservation
1279 $c->add_relationship( 'w91', 'w92', { type => 'grammatical' } );
1281 # Create an adjacency list of the whole thing; test the output.
1282 my $adj_whole = from_json( $c->as_adjacency_list() );
1283 is( scalar @$adj_whole, scalar $c->readings(),
1284 "Same number of nodes in graph and adjacency list" );
1285 my @adj_whole_edges;
1286 map { push( @adj_whole_edges, @{$_->{adjacent}} ) } @$adj_whole;
1287 is( scalar @adj_whole_edges, scalar $c->sequence->edges,
1288 "Same number of edges in graph and adjacency list" );
1289 # Find the reading whose rank should be preserved
1290 my( $test_rdg ) = grep { $_->{id} eq 'w89' } @$adj_whole;
1291 my( $test_edge ) = grep { $_->{id} eq 'w92' } @{$test_rdg->{adjacent}};
1292 is( $test_edge->{minlen}, 2, "Rank of test reading is preserved" );
1294 # Now create an adjacency list of just a portion. w76 to w122
1295 my $adj_part = from_json( $c->as_adjacency_list(
1296 { from => $c->reading('w76')->rank,
1297 to => $c->reading('w122')->rank }));
1298 is( scalar @$adj_part, 48, "Correct number of nodes in partial graph" );
1300 map { push( @adj_part_edges, @{$_->{adjacent}} ) } @$adj_part;
1301 is( scalar @adj_part_edges, 58,
1302 "Same number of edges in partial graph and adjacency list" );
1303 # Check for consistency
1305 map { $part_nodes{$_->{id}} = 1 } @$adj_part;
1306 foreach my $edge ( @adj_part_edges ) {
1307 my $testid = $edge->{id};
1308 ok( $part_nodes{$testid}, "ID $testid referenced in edge is given as node" );
1315 sub as_adjacency_list {
1316 my( $self, $opts ) = @_;
1317 # Make a structure that contains all the nodes, the nodes they point to,
1318 # and the attributes of the edges that connect them.
1319 # [ { id: 'n0', label: 'Gallia', adjacent: [
1320 # { id: 'n1', label: 'P Q' } ,
1321 # { id: 'n2', label: 'R S', minlen: 2 } ] },
1322 # { id: 'n1', label: 'est', adjacent: [ ... ] },
1324 my $startrank = $opts->{'from'} || 0;
1325 my $endrank = $opts->{'to'} || $self->end->rank;
1327 $self->calculate_ranks()
1328 unless( $self->_graphcalc_done || $opts->{'nocalc'} || !$self->linear );
1330 foreach my $rdg ( $self->readings ) {
1333 # Figure out what the node's successors should be.
1334 if( $rdg eq $self->start && $startrank > 0 ) {
1335 # Connect the start node with all the nodes at startrank.
1336 # Lacunas should be included only if the node really has that rank.
1337 @successors = $self->readings_at_rank( $startrank, 1 );
1339 } elsif( $rdg->rank < $startrank
1340 || $rdg->rank > $endrank && $rdg ne $self->end ) {
1343 @successors = $rdg->successors;
1345 # Make sure that the end node is at the end of the successors
1346 # list if it is needed.
1347 if( grep { $_ eq $self->end } @successors ) {
1348 my @ts = grep { $_ ne $self->end } @successors;
1349 @successors = ( @ts, $self->end );
1350 } elsif ( grep { $_->rank > $endrank } @successors ) {
1351 push( @successors, $self->end );
1354 my $listitem = { id => $rdg->id, label => $rdg->text };
1357 foreach my $succ ( @successors ) {
1359 if( $phony eq 'start' ) {
1360 @edgewits = $succ->witnesses;
1361 } elsif( $self->sequence->has_edge( $rdg->id, $succ->id ) ) {
1362 @edgewits = $self->path_witnesses( $rdg->id, $succ->id );
1365 if( $succ eq $self->end ) {
1366 @edgewits = @endwits;
1367 } elsif( $succ->rank > $endrank ) {
1368 # These witnesses will point to 'end' instead, not to the
1370 push( @endwits, @edgewits );
1373 my $edgelabel = $self->_path_display_label( $opts, @edgewits );
1374 my $edgedef = { id => $succ->id, label => $edgelabel };
1375 my $rankoffset = $succ->rank - $rdg->rank;
1376 if( $rankoffset > 1 and $succ ne $self->end ) {
1377 $edgedef->{minlen} = $rankoffset;
1379 push( @$adjacent, $edgedef );
1381 $listitem->{adjacent} = $adjacent;
1382 push( @$list, $listitem );
1384 return to_json( $list );
1389 Returns a GraphML representation of the collation. The GraphML will contain
1390 two graphs. The first expresses the attributes of the readings and the witness
1391 paths that link them; the second expresses the relationships that link the
1392 readings. This is the native transfer format for a tradition.
1396 use Text::Tradition;
1402 my $datafile = 't/data/florilegium_tei_ps.xml';
1403 my $tradition = Text::Tradition->new( 'input' => 'TEI',
1405 'file' => $datafile,
1408 ok( $tradition, "Got a tradition object" );
1409 is( scalar $tradition->witnesses, 13, "Found all witnesses" );
1410 ok( $tradition->collation, "Tradition has a collation" );
1412 my $c = $tradition->collation;
1413 is( scalar $c->readings, $READINGS, "Collation has all readings" );
1414 is( scalar $c->paths, $PATHS, "Collation has all paths" );
1415 is( scalar $c->relationships, 0, "Collation has all relationships" );
1417 # Add a few relationships
1418 $c->add_relationship( 'w123', 'w125', { 'type' => 'collated' } );
1419 $c->add_relationship( 'w193', 'w196', { 'type' => 'collated' } );
1420 $c->add_relationship( 'w257', 'w262', { 'type' => 'transposition',
1421 'is_significant' => 'yes' } );
1423 # Now write it to GraphML and parse it again.
1425 my $graphml = $c->as_graphml;
1426 my $st = Text::Tradition->new( 'input' => 'Self', 'string' => $graphml );
1427 is( scalar $st->collation->readings, $READINGS, "Reparsed collation has all readings" );
1428 is( scalar $st->collation->paths, $PATHS, "Reparsed collation has all paths" );
1429 is( scalar $st->collation->relationships, 3, "Reparsed collation has new relationships" );
1430 my $sigrel = $st->collation->get_relationship( 'w257', 'w262' );
1431 is( $sigrel->is_significant, 'yes', "Ternary attribute value was restored" );
1433 # Now add a stemma, write to GraphML, and look at the output.
1435 skip "Analysis module not present", 3 unless $tradition->can( 'add_stemma' );
1436 my $stemma = $tradition->add_stemma( 'dotfile' => 't/data/florilegium.dot' );
1437 is( ref( $stemma ), 'Text::Tradition::Stemma', "Parsed dotfile into stemma" );
1438 is( $tradition->stemmata, 1, "Tradition now has the stemma" );
1439 $graphml = $c->as_graphml;
1440 like( $graphml, qr/digraph/, "Digraph declaration exists in GraphML" );
1447 ## TODO MOVE this to Tradition.pm and modularize it better
1449 my( $self, $options ) = @_;
1450 $self->calculate_ranks unless $self->_graphcalc_done;
1452 my $start = $options->{'from'}
1453 ? $self->reading( $options->{'from'} ) : $self->start;
1454 my $end = $options->{'to'}
1455 ? $self->reading( $options->{'to'} ) : $self->end;
1456 if( $start->has_rank && $end->has_rank && $end->rank < $start->rank ) {
1457 throw( 'Start node must be before end node' );
1459 # The readings need to be ranked for this to work.
1460 $start = $self->start unless $start->has_rank;
1461 $end = $self->end unless $end->has_rank;
1463 unless( $start eq $self->start ) {
1464 $rankoffset = $start->rank - 1;
1469 my $graphml_ns = 'http://graphml.graphdrawing.org/xmlns';
1470 my $xsi_ns = 'http://www.w3.org/2001/XMLSchema-instance';
1471 my $graphml_schema = 'http://graphml.graphdrawing.org/xmlns ' .
1472 'http://graphml.graphdrawing.org/xmlns/1.0/graphml.xsd';
1474 # Create the document and root node
1475 require XML::LibXML;
1476 my $graphml = XML::LibXML->createDocument( "1.0", "UTF-8" );
1477 my $root = $graphml->createElementNS( $graphml_ns, 'graphml' );
1478 $graphml->setDocumentElement( $root );
1479 $root->setNamespace( $xsi_ns, 'xsi', 0 );
1480 $root->setAttributeNS( $xsi_ns, 'schemaLocation', $graphml_schema );
1482 # List of attribute types to save on our objects and their corresponding
1487 'Bool' => 'boolean',
1488 'ReadingID' => 'string',
1489 'RelationshipType' => 'string',
1490 'RelationshipScope' => 'string',
1491 'Ternary' => 'string',
1494 # Add the data keys for the graph. Include an extra key 'version' for the
1495 # GraphML output version.
1496 my %graph_data_keys;
1498 my %graph_attributes = ( 'version' => 'string' );
1499 # Graph attributes include those of Tradition and those of Collation.
1501 # TODO Use meta introspection method from duplicate_reading to do this
1502 # instead of naming custom keys.
1503 my $tmeta = $self->tradition->meta;
1504 my $cmeta = $self->meta;
1505 map { $gattr_from{$_->name} = 'Tradition' } $tmeta->get_all_attributes;
1506 map { $gattr_from{$_->name} = 'Collation' } $cmeta->get_all_attributes;
1507 foreach my $attr ( ( $tmeta->get_all_attributes, $cmeta->get_all_attributes ) ) {
1508 next if $attr->name =~ /^_/;
1509 next unless $save_types{$attr->type_constraint->name};
1510 $graph_attributes{$attr->name} = $save_types{$attr->type_constraint->name};
1512 # Extra custom keys for complex objects that should be saved in some form.
1513 # The subroutine should return a string, or undef/empty.
1514 if( $tmeta->has_method('stemmata') ) {
1515 $graph_attributes{'stemmata'} = sub {
1517 map { push( @stemstrs, $_->editable( {linesep => ''} ) ) }
1518 $self->tradition->stemmata;
1519 join( "\n", @stemstrs );
1523 if( $tmeta->has_method('user') ) {
1524 $graph_attributes{'user'} = sub {
1525 $self->tradition->user ? $self->tradition->user->id : undef
1529 foreach my $datum ( sort keys %graph_attributes ) {
1530 $graph_data_keys{$datum} = 'dg'.$gdi++;
1531 my $key = $root->addNewChild( $graphml_ns, 'key' );
1532 my $dtype = ref( $graph_attributes{$datum} ) ? 'string'
1533 : $graph_attributes{$datum};
1534 $key->setAttribute( 'attr.name', $datum );
1535 $key->setAttribute( 'attr.type', $dtype );
1536 $key->setAttribute( 'for', 'graph' );
1537 $key->setAttribute( 'id', $graph_data_keys{$datum} );
1540 # Add the data keys for reading nodes
1541 my %reading_attributes;
1542 my $rmeta = Text::Tradition::Collation::Reading->meta;
1543 foreach my $attr( $rmeta->get_all_attributes ) {
1544 next if $attr->name =~ /^_/;
1545 next unless $save_types{$attr->type_constraint->name};
1546 $reading_attributes{$attr->name} = $save_types{$attr->type_constraint->name};
1548 if( $self->start->does('Text::Tradition::Morphology' ) ) {
1549 # Extra custom key for the reading morphology
1550 $reading_attributes{'lexemes'} = 'string';
1555 foreach my $datum ( sort keys %reading_attributes ) {
1556 $node_data_keys{$datum} = 'dn'.$ndi++;
1557 my $key = $root->addNewChild( $graphml_ns, 'key' );
1558 $key->setAttribute( 'attr.name', $datum );
1559 $key->setAttribute( 'attr.type', $reading_attributes{$datum} );
1560 $key->setAttribute( 'for', 'node' );
1561 $key->setAttribute( 'id', $node_data_keys{$datum} );
1564 # Add the data keys for edges, that is, paths and relationships. Path
1565 # data does not come from a Moose class so is here manually.
1568 my %edge_attributes = (
1569 witness => 'string', # ID/label for a path
1570 extra => 'boolean', # Path key
1572 my @path_attributes = keys %edge_attributes; # track our manual additions
1573 my $pmeta = Text::Tradition::Collation::Relationship->meta;
1574 foreach my $attr( $pmeta->get_all_attributes ) {
1575 next if $attr->name =~ /^_/;
1576 next unless $save_types{$attr->type_constraint->name};
1577 $edge_attributes{$attr->name} = $save_types{$attr->type_constraint->name};
1579 foreach my $datum ( sort keys %edge_attributes ) {
1580 $edge_data_keys{$datum} = 'de'.$edi++;
1581 my $key = $root->addNewChild( $graphml_ns, 'key' );
1582 $key->setAttribute( 'attr.name', $datum );
1583 $key->setAttribute( 'attr.type', $edge_attributes{$datum} );
1584 $key->setAttribute( 'for', 'edge' );
1585 $key->setAttribute( 'id', $edge_data_keys{$datum} );
1588 # Add the collation graph itself. First, sanitize the name to a valid XML ID.
1589 my $xmlidname = $self->tradition->name;
1590 $xmlidname =~ s/(?!$xml10_namechar_rx)./_/g;
1591 if( $xmlidname !~ /^$xml10_namestartchar_rx/ ) {
1592 $xmlidname = '_'.$xmlidname;
1594 my $sgraph = $root->addNewChild( $graphml_ns, 'graph' );
1595 $sgraph->setAttribute( 'edgedefault', 'directed' );
1596 $sgraph->setAttribute( 'id', $xmlidname );
1597 $sgraph->setAttribute( 'parse.edgeids', 'canonical' );
1598 $sgraph->setAttribute( 'parse.edges', 0 ); # fill in later
1599 $sgraph->setAttribute( 'parse.nodeids', 'canonical' );
1600 $sgraph->setAttribute( 'parse.nodes', 0 ); # fill in later
1601 $sgraph->setAttribute( 'parse.order', 'nodesfirst' );
1603 # Tradition/collation attribute data
1604 foreach my $datum ( keys %graph_attributes ) {
1606 if( $datum eq 'version' ) {
1608 } elsif( ref( $graph_attributes{$datum} ) ) {
1609 my $sub = $graph_attributes{$datum};
1611 } elsif( $gattr_from{$datum} eq 'Tradition' ) {
1612 $value = $self->tradition->$datum;
1614 $value = $self->$datum;
1616 _add_graphml_data( $sgraph, $graph_data_keys{$datum}, $value );
1621 # Add our readings to the graph
1622 foreach my $n ( sort { $a->id cmp $b->id } $self->readings ) {
1623 next if $n->has_rank && $n ne $self->start && $n ne $self->end &&
1624 ( $n->rank < $start->rank || $n->rank > $end->rank );
1625 $use_readings{$n->id} = 1;
1626 # Add to the main graph
1627 my $node_el = $sgraph->addNewChild( $graphml_ns, 'node' );
1628 my $node_xmlid = 'n' . $node_ctr++;
1629 $node_hash{ $n->id } = $node_xmlid;
1630 $node_el->setAttribute( 'id', $node_xmlid );
1631 foreach my $d ( keys %reading_attributes ) {
1633 # Custom serialization
1634 if( $d eq 'lexemes' ) {
1635 # If nval is a true value, we have lexemes so we need to
1636 # serialize them. Otherwise set nval to undef so that the
1637 # key is excluded from this reading.
1638 $nval = $nval ? $n->_serialize_lexemes : undef;
1639 } elsif( $d eq 'normal_form' && $n->normal_form eq $n->text ) {
1642 if( $rankoffset && $d eq 'rank' && $n ne $self->start ) {
1643 # Adjust the ranks within the subgraph.
1644 $nval = $n eq $self->end ? $end->rank - $rankoffset + 1
1645 : $nval - $rankoffset;
1647 _add_graphml_data( $node_el, $node_data_keys{$d}, $nval )
1652 # Add the path edges to the sequence graph
1654 foreach my $e ( sort { $a->[0] cmp $b->[0] } $self->sequence->edges() ) {
1655 # We add an edge in the graphml for every witness in $e.
1656 next unless( $use_readings{$e->[0]} || $use_readings{$e->[1]} );
1657 my @edge_wits = sort $self->path_witnesses( $e );
1658 $e->[0] = $self->start->id unless $use_readings{$e->[0]};
1659 $e->[1] = $self->end->id unless $use_readings{$e->[1]};
1660 # Skip any path from start to end; that witness is not in the subgraph.
1661 next if ( $e->[0] eq $self->start->id && $e->[1] eq $self->end->id );
1662 foreach my $wit ( @edge_wits ) {
1663 my( $id, $from, $to ) = ( 'e'.$edge_ctr++,
1664 $node_hash{ $e->[0] },
1665 $node_hash{ $e->[1] } );
1666 my $edge_el = $sgraph->addNewChild( $graphml_ns, 'edge' );
1667 $edge_el->setAttribute( 'source', $from );
1668 $edge_el->setAttribute( 'target', $to );
1669 $edge_el->setAttribute( 'id', $id );
1671 # It's a witness path, so add the witness
1673 my $key = $edge_data_keys{'witness'};
1674 # Is this an ante-corr witness?
1675 my $aclabel = $self->ac_label;
1676 if( $wit =~ /^(.*)\Q$aclabel\E$/ ) {
1677 # Keep the base witness
1679 # ...and record that this is an 'extra' reading path
1680 _add_graphml_data( $edge_el, $edge_data_keys{'extra'}, $aclabel );
1682 _add_graphml_data( $edge_el, $edge_data_keys{'witness'}, $base );
1686 # Report the actual number of nodes and edges that went in
1687 $sgraph->setAttribute( 'parse.edges', $edge_ctr );
1688 $sgraph->setAttribute( 'parse.nodes', $node_ctr );
1690 # Add the relationship graph to the XML
1691 map { delete $edge_data_keys{$_} } @path_attributes;
1692 $self->relations->_as_graphml( $graphml_ns, $root, \%node_hash,
1693 $node_data_keys{'id'}, \%edge_data_keys );
1695 # Save and return the thing
1696 my $result = decode_utf8( $graphml->toString(1) );
1700 sub _add_graphml_data {
1701 my( $el, $key, $value ) = @_;
1702 return unless defined $value;
1703 my $data_el = $el->addNewChild( $el->namespaceURI, 'data' );
1704 $data_el->setAttribute( 'key', $key );
1705 $data_el->appendText( $value );
1710 Returns a CSV alignment table representation of the collation graph, one
1711 row per witness (or witness uncorrected.)
1715 Returns a tab-separated alignment table representation of the collation graph,
1716 one row per witness (or witness uncorrected.)
1720 use Text::Tradition;
1728 my $datafile = 't/data/florilegium_tei_ps.xml';
1729 my $tradition = Text::Tradition->new( 'input' => 'TEI',
1731 'file' => $datafile,
1734 my $c = $tradition->collation;
1735 # Export the thing to CSV
1736 my $csvstr = $c->as_csv();
1738 my $csv = Text::CSV->new({ sep_char => ',', binary => 1 });
1739 my @lines = split(/\n/, $csvstr );
1740 ok( $csv->parse( $lines[0] ), "Successfully parsed first line of CSV" );
1741 is( scalar( $csv->fields ), $WITS + $WITAC, "CSV has correct number of witness columns" );
1742 my @q_ac = grep { $_ eq 'Q'.$c->ac_label } $csv->fields;
1743 ok( @q_ac, "Found a layered witness" );
1745 my $t2 = Text::Tradition->new( input => 'Tabular',
1749 is( scalar $t2->collation->readings, $READINGS, "Reparsed CSV collation has all readings" );
1750 is( scalar $t2->collation->paths, $PATHS, "Reparsed CSV collation has all paths" );
1752 # Now do it with TSV
1753 my $tsvstr = $c->as_tsv();
1754 my $t3 = Text::Tradition->new( input => 'Tabular',
1758 is( scalar $t3->collation->readings, $READINGS, "Reparsed TSV collation has all readings" );
1759 is( scalar $t3->collation->paths, $PATHS, "Reparsed TSV collation has all paths" );
1761 my $table = $c->alignment_table;
1762 my $noaccsv = $c->as_csv({ noac => 1 });
1763 my @noaclines = split(/\n/, $noaccsv );
1764 ok( $csv->parse( $noaclines[0] ), "Successfully parsed first line of no-ac CSV" );
1765 is( scalar( $csv->fields ), $WITS, "CSV has correct number of witness columns" );
1766 is( $c->alignment_table, $table, "Request for CSV did not alter the alignment table" );
1768 my $safecsv = $c->as_csv({ safe_ac => 1});
1769 my @safelines = split(/\n/, $safecsv );
1770 ok( $csv->parse( $safelines[0] ), "Successfully parsed first line of safe CSV" );
1771 is( scalar( $csv->fields ), $WITS + $WITAC, "CSV has correct number of witness columns" );
1772 @q_ac = grep { $_ eq 'Q__L' } $csv->fields;
1773 ok( @q_ac, "Found a sanitized layered witness" );
1774 is( $c->alignment_table, $table, "Request for CSV did not alter the alignment table" );
1776 # Test relationship collapse
1777 $c->add_relationship( $c->readings_at_rank( 37 ), { type => 'spelling' } );
1778 $c->add_relationship( $c->readings_at_rank( 60 ), { type => 'spelling' } );
1780 my $mergedtsv = $c->as_tsv({mergetypes => [ 'spelling', 'orthographic' ] });
1781 my $t4 = Text::Tradition->new( input => 'Tabular',
1783 string => $mergedtsv,
1785 is( scalar $t4->collation->readings, $READINGS - 2, "Reparsed TSV merge collation has fewer readings" );
1786 is( scalar $t4->collation->paths, $PATHS - 4, "Reparsed TSV merge collation has fewer paths" );
1788 # Test non-ASCII sigla
1789 my $t5 = Text::Tradition->new( input => 'Tabular',
1791 file => 't/data/armexample.xlsx',
1793 my $awittsv = $t5->collation->as_tsv({ noac => 1, ascii => 1 });
1794 my @awitlines = split( /\n/, $awittsv );
1795 like( $awitlines[0], qr/_A_5315622/, "Found ASCII sigil variant in TSV" );
1802 my( $self, $opts ) = @_;
1803 my $table = $self->alignment_table( $opts );
1804 my $csv_options = { binary => 1, quote_null => 0 };
1805 $csv_options->{'sep_char'} = $opts->{fieldsep};
1806 if( $opts->{fieldsep} eq "\t" ) {
1807 # If it is really tab separated, nothing is an escape char.
1808 $csv_options->{'quote_char'} = undef;
1809 $csv_options->{'escape_char'} = '';
1811 my $csv = Text::CSV->new( $csv_options );
1814 # Make the header row
1815 my @witnesses = map { $_->{'witness'} } @{$table->{'alignment'}};
1816 if( $opts->{ascii} ) {
1817 # TODO think of a fix for this
1818 throw( "Cannot currently produce ASCII sigla with witness layers" )
1819 unless $opts->{noac};
1820 my @awits = map { $self->tradition->witness( $_ )->ascii_sigil } @witnesses;
1821 @witnesses = @awits;
1823 $csv->combine( @witnesses );
1824 push( @result, $csv->string );
1826 # Make the rest of the rows
1827 foreach my $idx ( 0 .. $table->{'length'} - 1 ) {
1828 my @rowobjs = map { $_->{'tokens'}->[$idx] } @{$table->{'alignment'}};
1829 my @row = map { $_ ? $_->{'t'}->text : $_ } @rowobjs;
1830 # Quick and dirty collapse of requested relationship types
1831 if( ref( $opts->{mergetypes} ) eq 'ARRAY' ) {
1832 # Now substitute the reading in the relevant index of @row
1833 # for its merge-related reading
1836 my $thisr = shift @rowobjs;
1838 next if exists $substitutes{$thisr->{t}->text};
1839 # Make sure we don't have A <-> B substitutions.
1840 $substitutes{$thisr->{t}->text} = $thisr->{t}->text;
1841 foreach my $thatr ( @rowobjs ) {
1843 next if exists $substitutes{$thatr->{t}->text};
1844 my $ttrel = $self->get_relationship( $thisr->{t}, $thatr->{t} );
1846 next unless grep { $ttrel->type eq $_ } @{$opts->{mergetypes}};
1847 # If we have got this far then we need to merge them.
1848 $substitutes{$thatr->{t}->text} = $thisr->{t}->text;
1851 @row = map { $_ && exists $substitutes{$_} ? $substitutes{$_} : $_ } @row;
1853 $csv->combine( @row );
1854 push( @result, $csv->string );
1856 return join( "\n", @result );
1861 my $opts = shift || {};
1862 $opts->{fieldsep} = ',';
1863 return $self->_tabular( $opts );
1868 my $opts = shift || {};
1869 $opts->{fieldsep} = "\t";
1870 return $self->_tabular( $opts );
1873 =head2 alignment_table
1875 Return a reference to an alignment table, in a slightly enhanced CollateX
1876 format which looks like this:
1878 $table = { alignment => [ { witness => "SIGIL",
1879 tokens => [ { t => "TEXT" }, ... ] },
1880 { witness => "SIG2",
1881 tokens => [ { t => "TEXT" }, ... ] },
1883 length => TEXTLEN };
1887 sub alignment_table {
1888 my( $self, $opts ) = @_;
1889 if( $self->has_cached_table ) {
1890 return $self->cached_table
1891 unless $opts->{noac} || $opts->{safe_ac};
1894 # Make sure we can do this
1895 throw( "Need a linear graph in order to make an alignment table" )
1896 unless $self->linear;
1897 $self->calculate_ranks()
1898 unless $self->_graphcalc_done && $self->end->has_rank;
1900 my $table = { 'alignment' => [], 'length' => $self->end->rank - 1 };
1901 my @all_pos = ( 1 .. $self->end->rank - 1 );
1902 foreach my $wit ( sort { $a->sigil cmp $b->sigil } $self->tradition->witnesses ) {
1903 # say STDERR "Making witness row(s) for " . $wit->sigil;
1904 my @wit_path = $self->reading_sequence( $self->start, $self->end, $wit->sigil );
1905 my @row = _make_witness_row( \@wit_path, \@all_pos );
1906 my $witobj = { 'witness' => $wit->sigil, 'tokens' => \@row };
1907 $witobj->{'identifier'} = $wit->identifier if $wit->identifier;
1908 push( @{$table->{'alignment'}}, $witobj );
1909 if( $wit->is_layered && !$opts->{noac} ) {
1910 my @wit_ac_path = $self->reading_sequence( $self->start, $self->end,
1911 $wit->sigil.$self->ac_label );
1912 my @ac_row = _make_witness_row( \@wit_ac_path, \@all_pos );
1913 my $witlabel = $opts->{safe_ac}
1914 ? $wit->sigil . '__L' : $wit->sigil.$self->ac_label;
1915 my $witacobj = { 'witness' => $witlabel,
1916 'tokens' => \@ac_row };
1917 $witacobj->{'identifier'} = $wit->identifier if $wit->identifier;
1918 push( @{$table->{'alignment'}}, $witacobj );
1921 unless( $opts->{noac} || $opts->{safe_ac} ) {
1922 $self->cached_table( $table );
1927 sub _make_witness_row {
1928 my( $path, $positions ) = @_;
1930 map { $char_hash{$_} = undef } @$positions;
1932 foreach my $rdg ( @$path ) {
1933 say STDERR "rank " . $rdg->rank if $debug;
1934 # say STDERR "No rank for " . $rdg->id unless defined $rdg->rank;
1935 $char_hash{$rdg->rank} = { 't' => $rdg };
1937 my @row = map { $char_hash{$_} } @$positions;
1938 # Fill in lacuna markers for undef spots in the row
1939 my $last_el = shift @row;
1940 my @filled_row = ( $last_el );
1941 foreach my $el ( @row ) {
1942 # If we are using node reference, make the lacuna node appear many times
1943 # in the table. If not, use the lacuna tag.
1944 if( $last_el && $last_el->{'t'}->is_lacuna && !defined $el ) {
1947 push( @filled_row, $el );
1954 =head1 NAVIGATION METHODS
1956 =head2 reading_sequence( $first, $last, $sigil, $backup )
1958 Returns the ordered list of readings, starting with $first and ending
1959 with $last, for the witness given in $sigil. If a $backup sigil is
1960 specified (e.g. when walking a layered witness), it will be used wherever
1961 no $sigil path exists. If there is a base text reading, that will be
1962 used wherever no path exists for $sigil or $backup.
1966 # TODO Think about returning some lazy-eval iterator.
1967 # TODO Get rid of backup; we should know from what witness is whether we need it.
1969 sub reading_sequence {
1970 my( $self, $start, $end, $witness ) = @_;
1972 $witness = $self->baselabel unless $witness;
1973 my @readings = ( $start );
1976 while( $n && $n->id ne $end->id ) {
1977 if( exists( $seen{$n->id} ) ) {
1978 throw( "Detected loop for $witness at " . $n->id );
1982 my $next = $self->next_reading( $n, $witness );
1984 throw( "Did not find any path for $witness from reading " . $n->id );
1986 push( @readings, $next );
1989 # Check that the last reading is our end reading.
1990 my $last = $readings[$#readings];
1991 throw( "Last reading found from " . $start->text .
1992 " for witness $witness is not the end!" ) # TODO do we get this far?
1993 unless $last->id eq $end->id;
1998 =head2 readings_at_rank( $rank )
2000 Returns a list of readings at a given rank, taken from the alignment table.
2004 sub readings_at_rank {
2005 my( $self, $rank, $nolacuna ) = @_;
2006 my $table = $self->alignment_table;
2007 # Table rank is real rank - 1.
2008 my @elements = map { $_->{'tokens'}->[$rank-1] } @{$table->{'alignment'}};
2010 foreach my $e ( @elements ) {
2011 next unless ref( $e ) eq 'HASH';
2012 next unless exists $e->{'t'};
2013 my $rdg = $e->{'t'};
2014 next if $nolacuna && $rdg->is_lacuna && $rdg->rank ne $rank;
2015 $readings{$e->{'t'}->id} = $e->{'t'};
2017 return values %readings;
2020 =head2 next_reading( $reading, $sigil );
2022 Returns the reading that follows the given reading along the given witness
2028 # Return the successor via the corresponding path.
2030 my $answer = $self->_find_linked_reading( 'next', @_ );
2031 return undef unless $answer;
2032 return $self->reading( $answer );
2035 =head2 prior_reading( $reading, $sigil )
2037 Returns the reading that precedes the given reading along the given witness
2043 # Return the predecessor via the corresponding path.
2045 my $answer = $self->_find_linked_reading( 'prior', @_ );
2046 return $self->reading( $answer );
2049 sub _find_linked_reading {
2050 my( $self, $direction, $node, $path ) = @_;
2052 # Get a backup if we are dealing with a layered witness
2054 my $aclabel = $self->ac_label;
2055 if( $path && $path =~ /^(.*)\Q$aclabel\E$/ ) {
2059 my @linked_paths = $direction eq 'next'
2060 ? $self->sequence->edges_from( $node )
2061 : $self->sequence->edges_to( $node );
2062 return undef unless scalar( @linked_paths );
2064 # We have to find the linked path that contains all of the
2065 # witnesses supplied in $path.
2066 my( @path_wits, @alt_path_wits );
2067 @path_wits = sort( $self->_witnesses_of_label( $path ) ) if $path;
2068 @alt_path_wits = sort( $self->_witnesses_of_label( $alt_path ) ) if $alt_path;
2071 foreach my $le ( @linked_paths ) {
2072 if( $self->sequence->has_edge_attribute( @$le, $self->baselabel ) ) {
2075 my @le_wits = sort $self->path_witnesses( $le );
2076 if( _is_within( \@path_wits, \@le_wits ) ) {
2077 # This is the right path.
2078 return $direction eq 'next' ? $le->[1] : $le->[0];
2079 } elsif( _is_within( \@alt_path_wits, \@le_wits ) ) {
2083 # Got this far? Return the alternate path if it exists.
2084 return $direction eq 'next' ? $alt_le->[1] : $alt_le->[0]
2087 # Got this far? Return the base path if it exists.
2088 return $direction eq 'next' ? $base_le->[1] : $base_le->[0]
2091 # Got this far? We have no appropriate path.
2092 warn "Could not find $direction node from " . $node->id
2093 . " along path $path";
2099 my( $set1, $set2 ) = @_;
2100 my $ret = @$set1; # will be 0, i.e. false, if set1 is empty
2101 foreach my $el ( @$set1 ) {
2102 $ret = 0 unless grep { /^\Q$el\E$/ } @$set2;
2107 # Return the string that joins together a list of witnesses for
2108 # display on a single path.
2109 sub _witnesses_of_label {
2110 my( $self, $label ) = @_;
2111 my $regex = $self->wit_list_separator;
2112 my @answer = split( /\Q$regex\E/, $label );
2116 =head2 common_readings
2118 Returns the list of common readings in the graph (i.e. those readings that are
2119 shared by all non-lacunose witnesses.)
2123 sub common_readings {
2125 my @common = grep { $_->is_common } $self->readings;
2129 =head2 path_text( $sigil, [, $start, $end ] )
2131 Returns the text of a witness (plus its backup, if we are using a layer)
2132 as stored in the collation. The text is returned as a string, where the
2133 individual readings are joined with spaces and the meta-readings (e.g.
2134 lacunae) are omitted. Optional specification of $start and $end allows
2135 the generation of a subset of the witness text.
2140 my( $self, $wit, $start, $end ) = @_;
2141 $start = $self->start unless $start;
2142 $end = $self->end unless $end;
2143 my @path = grep { !$_->is_meta } $self->reading_sequence( $start, $end, $wit );
2146 foreach my $r ( @path ) {
2147 unless ( $r->join_prior || !$last || $last->join_next ) {
2150 $pathtext .= $r->text;
2156 =head1 INITIALIZATION METHODS
2158 These are mostly for use by parsers.
2160 =head2 make_witness_path( $witness )
2162 Link the array of readings contained in $witness->path (and in
2163 $witness->uncorrected_path if it exists) into collation paths.
2164 Clear out the arrays when finished.
2166 =head2 make_witness_paths
2168 Call make_witness_path for all witnesses in the tradition.
2172 # For use when a collation is constructed from a base text and an apparatus.
2173 # We have the sequences of readings and just need to add path edges.
2174 # When we are done, clear out the witness path attributes, as they are no
2176 # TODO Find a way to replace the witness path attributes with encapsulated functions?
2178 sub make_witness_paths {
2180 foreach my $wit ( $self->tradition->witnesses ) {
2181 # say STDERR "Making path for " . $wit->sigil;
2182 $self->make_witness_path( $wit );
2186 sub make_witness_path {
2187 my( $self, $wit ) = @_;
2188 my @chain = @{$wit->path};
2189 my $sig = $wit->sigil;
2190 # Add start and end if necessary
2191 unshift( @chain, $self->start ) unless $chain[0] eq $self->start;
2192 push( @chain, $self->end ) unless $chain[-1] eq $self->end;
2193 foreach my $idx ( 0 .. $#chain-1 ) {
2194 $self->add_path( $chain[$idx], $chain[$idx+1], $sig );
2196 if( $wit->is_layered ) {
2197 @chain = @{$wit->uncorrected_path};
2198 unshift( @chain, $self->start ) unless $chain[0] eq $self->start;
2199 push( @chain, $self->end ) unless $chain[-1] eq $self->end;
2200 foreach my $idx( 0 .. $#chain-1 ) {
2201 my $source = $chain[$idx];
2202 my $target = $chain[$idx+1];
2203 $self->add_path( $source, $target, $sig.$self->ac_label )
2204 unless $self->has_path( $source, $target, $sig );
2208 $wit->clear_uncorrected_path;
2211 =head2 calculate_ranks
2213 Calculate the reading ranks (that is, their aligned positions relative
2214 to each other) for the graph. This can only be called on linear collations.
2218 use Text::Tradition;
2220 my $cxfile = 't/data/Collatex-16.xml';
2221 my $t = Text::Tradition->new(
2223 'input' => 'CollateX',
2226 my $c = $t->collation;
2229 my $table = $c->alignment_table;
2230 ok( $c->has_cached_table, "Alignment table was cached" );
2231 is( $c->alignment_table, $table, "Cached table returned upon second call" );
2232 $c->calculate_ranks;
2233 is( $c->alignment_table, $table, "Cached table retained with no rank change" );
2234 $c->add_relationship( 'n13', 'n23', { type => 'repetition' } );
2235 is( $c->alignment_table, $table, "Alignment table unchanged after non-colo relationship add" );
2236 $c->add_relationship( 'n24', 'n23', { type => 'spelling' } );
2237 isnt( $c->alignment_table, $table, "Alignment table changed after colo relationship add" );
2243 sub calculate_ranks {
2245 # Save the existing ranks, in case we need to invalidate the cached SVG.
2246 throw( "Cannot calculate ranks on a non-linear graph" )
2247 unless $self->linear;
2249 map { $existing_ranks{$_} = $_->rank } $self->readings;
2251 # Do the rankings based on the relationship equivalence graph, starting
2252 # with the start node.
2253 my ( $node_ranks, $rank_nodes ) = $self->relations->equivalence_ranks();
2255 # Transfer our rankings from the topological graph to the real one.
2256 foreach my $r ( $self->readings ) {
2257 if( defined $node_ranks->{$self->equivalence( $r->id )} ) {
2258 $r->rank( $node_ranks->{$self->equivalence( $r->id )} );
2260 # Die. Find the last rank we calculated.
2261 my @all_defined = sort { ( $node_ranks->{$self->equivalence( $a->id )}||-1 )
2262 <=> ( $node_ranks->{$self->equivalence( $b->id )}||-1 ) }
2264 my $last = pop @all_defined;
2265 throw( "Ranks not calculated after $last - do you have a cycle in the graph?" );
2268 # Do we need to invalidate the cached data?
2269 if( $self->has_cached_table ) {
2270 foreach my $r ( $self->readings ) {
2271 next if defined( $existing_ranks{$r} )
2272 && $existing_ranks{$r} == $r->rank;
2273 # Something has changed, so clear the cache
2274 $self->_clear_cache;
2275 # ...and recalculate the common readings.
2276 $self->calculate_common_readings();
2280 # The graph calculation information is now up to date.
2281 $self->_graphcalc_done(1);
2286 $self->wipe_table if $self->has_cached_table;
2290 =head2 flatten_ranks
2292 A convenience method for parsing collation data. Searches the graph for readings
2293 with the same text at the same rank, and merges any that are found.
2298 my ( $self, %args ) = shift;
2299 my %unique_rank_rdg;
2301 foreach my $p ( $self->identical_readings( %args ) ) {
2302 # say STDERR "Combining readings at same rank: @$p";
2304 $self->merge_readings( @$p );
2305 # TODO see if this now makes a common point.
2307 # If we merged readings, the ranks are still fine but the alignment
2308 # table is wrong. Wipe it.
2309 $self->wipe_table() if $changed;
2312 =head2 identical_readings
2313 =head2 identical_readings( start => $startnode, end => $endnode )
2314 =head2 identical_readings( startrank => $startrank, endrank => $endrank )
2316 Goes through the graph identifying all pairs of readings that appear to be
2317 identical, and therefore able to be merged into a single reading. Returns the
2318 relevant identical pairs. Can be restricted to run over only a part of the
2319 graph, specified either by node or by rank.
2323 sub identical_readings {
2324 my ( $self, %args ) = @_;
2325 # Find where we should start and end.
2326 my $startrank = $args{startrank} || 0;
2327 if( $args{start} ) {
2328 throw( "Starting reading has no rank" ) unless $self->reading( $args{start} )
2329 && $self->reading( $args{start} )->has_rank;
2330 $startrank = $self->reading( $args{start} )->rank;
2332 my $endrank = $args{endrank} || $self->end->rank;
2334 throw( "Ending reading has no rank" ) unless $self->reading( $args{end} )
2335 && $self->reading( $args{end} )->has_rank;
2336 $endrank = $self->reading( $args{end} )->rank;
2339 # Make sure the ranks are correct.
2340 unless( $self->_graphcalc_done ) {
2341 $self->calculate_ranks;
2343 # Go through the readings looking for duplicates.
2344 my %unique_rank_rdg;
2346 foreach my $rdg ( $self->readings ) {
2347 next unless $rdg->has_rank;
2348 my $rk = $rdg->rank;
2349 next if $rk > $endrank || $rk < $startrank;
2350 my $key = $rk . "||" . $rdg->text;
2351 if( exists $unique_rank_rdg{$key} ) {
2352 # Make sure they don't have different grammatical forms
2353 my $ur = $unique_rank_rdg{$key};
2354 if( $rdg->is_identical( $ur ) ) {
2355 push( @pairs, [ $ur, $rdg ] );
2358 $unique_rank_rdg{$key} = $rdg;
2366 =head2 calculate_common_readings
2368 Goes through the graph identifying the readings that appear in every witness
2369 (apart from those with lacunae at that spot.) Marks them as common and returns
2374 use Text::Tradition;
2376 my $cxfile = 't/data/Collatex-16.xml';
2377 my $t = Text::Tradition->new(
2379 'input' => 'CollateX',
2382 my $c = $t->collation;
2384 my @common = $c->calculate_common_readings();
2385 is( scalar @common, 8, "Found correct number of common readings" );
2386 my @marked = sort $c->common_readings();
2387 is( scalar @common, 8, "All common readings got marked as such" );
2388 my @expected = qw/ n1 n11 n16 n19 n20 n5 n6 n7 /;
2389 is_deeply( \@marked, \@expected, "Found correct list of common readings" );
2395 sub calculate_common_readings {
2398 map { $_->is_common( 0 ) } $self->readings;
2399 # Implicitly calls calculate_ranks
2400 my $table = $self->alignment_table;
2401 foreach my $idx ( 0 .. $table->{'length'} - 1 ) {
2402 my @row = map { $_->{'tokens'}->[$idx]
2403 ? $_->{'tokens'}->[$idx]->{'t'} : '' }
2404 @{$table->{'alignment'}};
2406 foreach my $r ( @row ) {
2408 $hash{$r->id} = $r unless $r->is_meta;
2410 $hash{'UNDEF'} = $r;
2413 if( keys %hash == 1 && !exists $hash{'UNDEF'} ) {
2414 my( $r ) = values %hash;
2416 push( @common, $r );
2422 =head2 text_from_paths
2424 Calculate the text array for all witnesses from the path, for later consistency
2425 checking. Only to be used if there is no non-graph-based way to know the
2430 sub text_from_paths {
2432 foreach my $wit ( $self->tradition->witnesses ) {
2433 my @readings = $self->reading_sequence( $self->start, $self->end, $wit->sigil );
2435 foreach my $r ( @readings ) {
2436 next if $r->is_meta;
2437 push( @text, $r->text );
2439 $wit->text( \@text );
2440 if( $wit->is_layered ) {
2441 my @ucrdgs = $self->reading_sequence( $self->start, $self->end,
2442 $wit->sigil.$self->ac_label );
2444 foreach my $r ( @ucrdgs ) {
2445 next if $r->is_meta;
2446 push( @uctext, $r->text );
2448 $wit->layertext( \@uctext );
2453 =head1 UTILITY FUNCTIONS
2455 =head2 common_predecessor( $reading_a, $reading_b )
2457 Find the last reading that occurs in sequence before both the given readings.
2458 At the very least this should be $self->start.
2460 =head2 common_successor( $reading_a, $reading_b )
2462 Find the first reading that occurs in sequence after both the given readings.
2463 At the very least this should be $self->end.
2467 use Text::Tradition;
2469 my $cxfile = 't/data/Collatex-16.xml';
2470 my $t = Text::Tradition->new(
2472 'input' => 'CollateX',
2475 my $c = $t->collation;
2477 is( $c->common_predecessor( 'n24', 'n23' )->id,
2478 'n20', "Found correct common predecessor" );
2479 is( $c->common_successor( 'n24', 'n23' )->id,
2480 '__END__', "Found correct common successor" );
2482 is( $c->common_predecessor( 'n19', 'n17' )->id,
2483 'n16', "Found correct common predecessor for readings on same path" );
2484 is( $c->common_successor( 'n21', 'n10' )->id,
2485 '__END__', "Found correct common successor for readings on same path" );
2491 ## Return the closest reading that is a predecessor of both the given readings.
2492 sub common_predecessor {
2494 my( $r1, $r2 ) = $self->_objectify_args( @_ );
2495 return $self->_common_in_path( $r1, $r2, 'predecessors' );
2498 sub common_successor {
2500 my( $r1, $r2 ) = $self->_objectify_args( @_ );
2501 return $self->_common_in_path( $r1, $r2, 'successors' );
2505 # TODO think about how to do this without ranks...
2506 sub _common_in_path {
2507 my( $self, $r1, $r2, $dir ) = @_;
2508 my $iter = $self->end->rank;
2510 my @last_r1 = ( $r1 );
2511 my @last_r2 = ( $r2 );
2512 # my %all_seen = ( $r1 => 'r1', $r2 => 'r2' );
2514 # say STDERR "Finding common $dir for $r1, $r2";
2515 while( !@candidates ) {
2516 last unless $iter--; # Avoid looping infinitely
2517 # Iterate separately down the graph from r1 and r2
2518 my( @new_lc1, @new_lc2 );
2519 foreach my $lc ( @last_r1 ) {
2520 foreach my $p ( $lc->$dir ) {
2521 if( $all_seen{$p->id} && $all_seen{$p->id} ne 'r1' ) {
2522 # say STDERR "Path candidate $p from $lc";
2523 push( @candidates, $p );
2524 } elsif( !$all_seen{$p->id} ) {
2525 $all_seen{$p->id} = 'r1';
2526 push( @new_lc1, $p );
2530 foreach my $lc ( @last_r2 ) {
2531 foreach my $p ( $lc->$dir ) {
2532 if( $all_seen{$p->id} && $all_seen{$p->id} ne 'r2' ) {
2533 # say STDERR "Path candidate $p from $lc";
2534 push( @candidates, $p );
2535 } elsif( !$all_seen{$p->id} ) {
2536 $all_seen{$p->id} = 'r2';
2537 push( @new_lc2, $p );
2541 @last_r1 = @new_lc1;
2542 @last_r2 = @new_lc2;
2544 my @answer = sort { $a->rank <=> $b->rank } @candidates;
2545 return $dir eq 'predecessors' ? pop( @answer ) : shift ( @answer );
2549 Text::Tradition::Error->throw(
2550 'ident' => 'Collation error',
2556 __PACKAGE__->meta->make_immutable;
2562 =item * Rework XML serialization in a more modular way
2568 This package is free software and is provided "as is" without express
2569 or implied warranty. You can redistribute it and/or modify it under
2570 the same terms as Perl itself.
2574 Tara L Andrews E<lt>aurum@cpan.orgE<gt>