1 package Text::Tradition::Collation;
4 use Encode qw( decode_utf8 );
8 use IPC::Run qw( run binary );
10 use Text::Tradition::Collation::Reading;
11 use Text::Tradition::Collation::RelationshipStore;
12 use Text::Tradition::Error;
13 use XML::Easy::Syntax qw( $xml10_namestartchar_rx $xml10_namechar_rx );
19 default => sub { Graph->new() },
27 isa => 'Text::Tradition::Collation::RelationshipStore',
29 relationships => 'relationships',
30 related_readings => 'related_readings',
31 get_relationship => 'get_relationship',
32 del_relationship => 'del_relationship',
33 equivalence => 'equivalence',
34 equivalence_graph => 'equivalence_graph',
36 writer => '_set_relations',
41 isa => 'Text::Tradition',
42 writer => '_set_tradition',
47 isa => 'HashRef[Text::Tradition::Collation::Reading]',
51 _add_reading => 'set',
52 del_reading => 'delete',
53 has_reading => 'exists',
54 # reading_keys => 'keys',
57 default => sub { {} },
60 has 'wit_list_separator' => (
69 default => 'base text',
92 isa => 'Text::Tradition::Collation::Reading',
93 writer => '_set_start',
99 isa => 'Text::Tradition::Collation::Reading',
100 writer => '_set_end',
104 has 'cached_table' => (
107 predicate => 'has_cached_table',
108 clearer => 'wipe_table',
111 has '_graphcalc_done' => (
119 Text::Tradition::Collation - a software model for a text collation
124 my $t = Text::Tradition->new(
125 'name' => 'this is a text',
127 'file' => '/path/to/tei_parallel_seg_file.xml' );
129 my $c = $t->collation;
130 my @readings = $c->readings;
131 my @paths = $c->paths;
132 my @relationships = $c->relationships;
134 my $svg_variant_graph = $t->collation->as_svg();
138 Text::Tradition is a library for representation and analysis of collated
139 texts, particularly medieval ones. The Collation is the central feature of
140 a Tradition, where the text, its sequence of readings, and its relationships
141 between readings are actually kept.
147 The constructor. Takes a hash or hashref of the following arguments:
151 =item * tradition - The Text::Tradition object to which the collation
154 =item * linear - Whether the collation should be linear; that is, whether
155 transposed readings should be treated as two linked readings rather than one,
156 and therefore whether the collation graph is acyclic. Defaults to true.
158 =item * baselabel - The default label for the path taken by a base text
159 (if any). Defaults to 'base text'.
161 =item * wit_list_separator - The string to join a list of witnesses for
162 purposes of making labels in display graphs. Defaults to ', '.
164 =item * ac_label - The extra label to tack onto a witness sigil when
165 representing another layer of path for the given witness - that is, when
166 a text has more than one possible reading due to scribal corrections or
167 the like. Defaults to ' (a.c.)'.
169 =item * wordsep - The string used to separate words in the original text.
180 =head2 wit_list_separator
188 Simple accessors for collation attributes.
192 The meta-reading at the start of every witness path.
196 The meta-reading at the end of every witness path.
200 Returns all Reading objects in the graph.
202 =head2 reading( $id )
204 Returns the Reading object corresponding to the given ID.
206 =head2 add_reading( $reading_args )
208 Adds a new reading object to the collation.
209 See L<Text::Tradition::Collation::Reading> for the available arguments.
211 =head2 del_reading( $object_or_id )
213 Removes the given reading from the collation, implicitly removing its
214 paths and relationships.
216 =head2 merge_readings( $main, $second, $concatenate, $with_str )
218 Merges the $second reading into the $main one. If $concatenate is true, then
219 the merged node will carry the text of both readings, concatenated with either
220 $with_str (if specified) or a sensible default (the empty string if the
221 appropriate 'join_*' flag is set on either reading, or else $self->wordsep.)
223 The first two arguments may be either readings or reading IDs.
225 =head2 has_reading( $id )
227 Predicate to see whether a given reading ID is in the graph.
229 =head2 reading_witnesses( $object_or_id )
231 Returns a list of sigils whose witnesses contain the reading.
235 Returns all reading paths within the document - that is, all edges in the
236 collation graph. Each path is an arrayref of [ $source, $target ] reading IDs.
238 =head2 add_path( $source, $target, $sigil )
240 Links the given readings in the collation in sequence, under the given witness
241 sigil. The readings may be specified by object or ID.
243 =head2 del_path( $source, $target, $sigil )
245 Links the given readings in the collation in sequence, under the given witness
246 sigil. The readings may be specified by object or ID.
248 =head2 has_path( $source, $target );
250 Returns true if the two readings are linked in sequence in any witness.
251 The readings may be specified by object or ID.
255 Returns all Relationship objects in the collation.
257 =head2 add_relationship( $reading, $other_reading, $options )
259 Adds a new relationship of the type given in $options between the two readings,
260 which may be specified by object or ID. Returns a value of ( $status, @vectors)
261 where $status is true on success, and @vectors is a list of relationship edges
262 that were ultimately added.
263 See L<Text::Tradition::Collation::Relationship> for the available options.
269 $self->_set_relations( Text::Tradition::Collation::RelationshipStore->new( 'collation' => $self ) );
270 $self->_set_start( $self->add_reading(
271 { 'collation' => $self, 'is_start' => 1, 'init' => 1 } ) );
272 $self->_set_end( $self->add_reading(
273 { 'collation' => $self, 'is_end' => 1, 'init' => 1 } ) );
276 ### Reading construct/destruct functions
279 my( $self, $reading ) = @_;
280 unless( ref( $reading ) eq 'Text::Tradition::Collation::Reading' ) {
281 my %args = %$reading;
282 if( $args{'init'} ) {
283 # If we are initializing an empty collation, don't assume that we
284 # have set a tradition.
285 delete $args{'init'};
286 } elsif( $self->tradition->has_language && !exists $args{'language'} ) {
287 $args{'language'} = $self->tradition->language;
289 $reading = Text::Tradition::Collation::Reading->new(
290 'collation' => $self,
293 # First check to see if a reading with this ID exists.
294 if( $self->reading( $reading->id ) ) {
295 throw( "Collation already has a reading with id " . $reading->id );
297 $self->_graphcalc_done(0);
298 $self->_add_reading( $reading->id => $reading );
299 # Once the reading has been added, put it in both graphs.
300 $self->sequence->add_vertex( $reading->id );
301 $self->relations->add_reading( $reading->id );
305 around del_reading => sub {
310 if( ref( $arg ) eq 'Text::Tradition::Collation::Reading' ) {
313 # Remove the reading from the graphs.
314 $self->_graphcalc_done(0);
315 $self->_clear_cache; # Explicitly clear caches to GC the reading
316 $self->sequence->delete_vertex( $arg );
317 $self->relations->delete_reading( $arg );
320 $self->$orig( $arg );
327 my $cxfile = 't/data/Collatex-16.xml';
328 my $t = Text::Tradition->new(
330 'input' => 'CollateX',
333 my $c = $t->collation;
335 my $rno = scalar $c->readings;
336 # Split n21 for testing purposes
337 my $new_r = $c->add_reading( { 'id' => 'n21p0', 'text' => 'un', 'join_next' => 1 } );
338 my $old_r = $c->reading( 'n21' );
339 $old_r->alter_text( 'to' );
340 $c->del_path( 'n20', 'n21', 'A' );
341 $c->add_path( 'n20', 'n21p0', 'A' );
342 $c->add_path( 'n21p0', 'n21', 'A' );
344 ok( $c->reading( 'n21p0' ), "New reading exists" );
345 is( scalar $c->readings, $rno, "Reading add offset by flatten_ranks" );
347 # Combine n3 and n4 ( with his )
348 $c->merge_readings( 'n3', 'n4', 1 );
349 ok( !$c->reading('n4'), "Reading n4 is gone" );
350 is( $c->reading('n3')->text, 'with his', "Reading n3 has both words" );
352 # Collapse n9 and n10 ( rood / root )
353 $c->merge_readings( 'n9', 'n10' );
354 ok( !$c->reading('n10'), "Reading n10 is gone" );
355 is( $c->reading('n9')->text, 'rood', "Reading n9 has an unchanged word" );
357 # Combine n21 and n21p0
358 my $remaining = $c->reading('n21');
359 $remaining ||= $c->reading('n22'); # one of these should still exist
360 $c->merge_readings( 'n21p0', $remaining, 1 );
361 ok( !$c->reading('n21'), "Reading $remaining is gone" );
362 is( $c->reading('n21p0')->text, 'unto', "Reading n21p0 merged correctly" );
372 my( $kept_obj, $del_obj, $combine, $combine_char ) = $self->_objectify_args( @_ );
373 my $mergemeta = $kept_obj->is_meta;
374 throw( "Cannot merge meta and non-meta reading" )
375 unless ( $mergemeta && $del_obj->is_meta )
376 || ( !$mergemeta && !$del_obj->is_meta );
378 throw( "Cannot merge with start or end node" )
379 if( $kept_obj eq $self->start || $kept_obj eq $self->end
380 || $del_obj eq $self->start || $del_obj eq $self->end );
382 # We only need the IDs for adding paths to the graph, not the reading
383 # objects themselves.
384 my $kept = $kept_obj->id;
385 my $deleted = $del_obj->id;
386 $self->_graphcalc_done(0);
388 # The kept reading should inherit the paths and the relationships
389 # of the deleted reading.
390 foreach my $path ( $self->sequence->edges_at( $deleted ) ) {
391 my @vector = ( $kept );
392 push( @vector, $path->[1] ) if $path->[0] eq $deleted;
393 unshift( @vector, $path->[0] ) if $path->[1] eq $deleted;
394 next if $vector[0] eq $vector[1]; # Don't add a self loop
395 my %wits = %{$self->sequence->get_edge_attributes( @$path )};
396 $self->sequence->add_edge( @vector );
397 my $fwits = $self->sequence->get_edge_attributes( @vector );
398 @wits{keys %$fwits} = values %$fwits;
399 $self->sequence->set_edge_attributes( @vector, \%wits );
401 $self->relations->merge_readings( $kept, $deleted, $combine );
403 # Do the deletion deed.
405 # Combine the text of the readings
406 my $joinstr = $combine_char;
407 unless( defined $joinstr ) {
408 $joinstr = '' if $kept_obj->join_next || $del_obj->join_prior;
409 $joinstr = $self->wordsep unless defined $joinstr;
411 $kept_obj->alter_text( join( $joinstr, $kept_obj->text, $del_obj->text ) );
412 # Change this reading to a joining one if necessary
413 $kept_obj->_set_join_next( $del_obj->join_next );
414 $kept_obj->normal_form(
415 join( $joinstr, $kept_obj->normal_form, $del_obj->normal_form ) );
416 # Combine the lexemes present in the readings
417 if( $kept_obj->has_lexemes && $del_obj->has_lexemes ) {
418 $kept_obj->add_lexeme( $del_obj->lexemes );
421 $self->del_reading( $deleted );
424 =head2 compress_readings
426 Where possible in the graph, compresses plain sequences of readings into a
427 single reading. The sequences must consist of readings with no
428 relationships to other readings, with only a single witness path between
429 them and no other witness paths from either that would skip the other. The
430 readings must also not be marked as nonsense or bad grammar.
432 WARNING: This operation cannot be undone.
436 sub compress_readings {
438 # Anywhere in the graph that there is a reading that joins only to a single
439 # successor, and neither of these have any relationships, just join the two
442 foreach my $rdg ( sort { $a->rank <=> $b->rank } $self->readings ) {
443 # While we are here, get rid of any extra wordforms from a disambiguated
445 if( $rdg->disambiguated ) {
446 foreach my $lex ( $rdg->lexemes ) {
447 $lex->clear_matching_forms();
448 $lex->add_matching_form( $lex->form );
451 # Now look for readings that can be joined to their successors.
452 next if $rdg->is_meta;
453 next if $gobbled{$rdg->id};
454 next if $rdg->grammar_invalid || $rdg->is_nonsense;
455 next if $rdg->related_readings();
457 while( $self->sequence->successors( $rdg ) == 1 ) {
458 my( $next ) = $self->reading( $self->sequence->successors( $rdg ) );
459 throw( "Infinite loop" ) if $seen{$next->id};
460 $seen{$next->id} = 1;
461 last if $self->sequence->predecessors( $next ) > 1;
462 last if $next->is_meta;
463 last if $next->grammar_invalid || $next->is_nonsense;
464 last if $next->related_readings();
465 say "Joining readings $rdg and $next";
466 $self->merge_readings( $rdg, $next, 1 );
469 # Make sure we haven't screwed anything up
470 foreach my $wit ( $self->tradition->witnesses ) {
471 my $pathtext = $self->path_text( $wit->sigil );
472 my $origtext = join( ' ', @{$wit->text} );
473 throw( "Text differs for witness " . $wit->sigil )
474 unless $pathtext eq $origtext;
475 if( $wit->is_layered ) {
476 $pathtext = $self->path_text( $wit->sigil.$self->ac_label );
477 $origtext = join( ' ', @{$wit->layertext} );
478 throw( "Ante-corr text differs for witness " . $wit->sigil )
479 unless $pathtext eq $origtext;
483 $self->relations->rebuild_equivalence();
484 $self->calculate_ranks();
487 # Helper function for manipulating the graph.
488 sub _stringify_args {
489 my( $self, $first, $second, @args ) = @_;
491 if ref( $first ) eq 'Text::Tradition::Collation::Reading';
492 $second = $second->id
493 if ref( $second ) eq 'Text::Tradition::Collation::Reading';
494 return( $first, $second, @args );
497 # Helper function for manipulating the graph.
498 sub _objectify_args {
499 my( $self, $first, $second, $arg ) = @_;
500 $first = $self->reading( $first )
501 unless ref( $first ) eq 'Text::Tradition::Collation::Reading';
502 $second = $self->reading( $second )
503 unless ref( $second ) eq 'Text::Tradition::Collation::Reading';
504 return( $first, $second, $arg );
511 # We only need the IDs for adding paths to the graph, not the reading
512 # objects themselves.
513 my( $source, $target, $wit ) = $self->_stringify_args( @_ );
515 $self->_graphcalc_done(0);
516 # Connect the readings
517 unless( $self->sequence->has_edge( $source, $target ) ) {
518 $self->sequence->add_edge( $source, $target );
519 $self->relations->add_equivalence_edge( $source, $target );
521 # Note the witness in question
522 $self->sequence->set_edge_attribute( $source, $target, $wit, 1 );
528 if( ref( $_[0] ) eq 'ARRAY' ) {
535 # We only need the IDs for adding paths to the graph, not the reading
536 # objects themselves.
537 my( $source, $target, $wit ) = $self->_stringify_args( @args );
539 $self->_graphcalc_done(0);
540 if( $self->sequence->has_edge_attribute( $source, $target, $wit ) ) {
541 $self->sequence->delete_edge_attribute( $source, $target, $wit );
543 unless( keys %{$self->sequence->get_edge_attributes( $source, $target )} ) {
544 $self->sequence->delete_edge( $source, $target );
545 $self->relations->delete_equivalence_edge( $source, $target );
550 # Extra graph-alike utility
553 my( $source, $target, $wit ) = $self->_stringify_args( @_ );
554 return undef unless $self->sequence->has_edge( $source, $target );
555 return $self->sequence->has_edge_attribute( $source, $target, $wit );
558 =head2 clear_witness( @sigil_list )
560 Clear the given witnesses out of the collation entirely, removing references
561 to them in paths, and removing readings that belong only to them. Should only
562 be called via $tradition->del_witness.
567 my( $self, @sigils ) = @_;
569 $self->_graphcalc_done(0);
570 # Clear the witness(es) out of the paths
571 foreach my $e ( $self->paths ) {
572 foreach my $sig ( @sigils ) {
573 $self->del_path( $e, $sig );
577 # Clear out the newly unused readings
578 foreach my $r ( $self->readings ) {
579 unless( $self->reading_witnesses( $r ) ) {
580 $self->del_reading( $r );
585 sub add_relationship {
587 my( $source, $target, $opts ) = $self->_stringify_args( @_ );
588 my( @vectors ) = $self->relations->add_relationship( $source, $target, $opts );
589 $self->_graphcalc_done(0);
593 around qw/ get_relationship del_relationship / => sub {
597 if( @args == 1 && ref( $args[0] ) eq 'ARRAY' ) {
600 my( $source, $target ) = $self->_stringify_args( @args );
601 $self->$orig( $source, $target );
604 =head2 reading_witnesses( $reading )
606 Return a list of sigils corresponding to the witnesses in which the reading appears.
610 sub reading_witnesses {
611 my( $self, $reading ) = @_;
612 # We need only check either the incoming or the outgoing edges; I have
613 # arbitrarily chosen "incoming". Thus, special-case the start node.
614 if( $reading eq $self->start ) {
615 return map { $_->sigil } $self->tradition->witnesses;
618 foreach my $e ( $self->sequence->edges_to( $reading ) ) {
619 my $wits = $self->sequence->get_edge_attributes( @$e );
620 @all_witnesses{ keys %$wits } = 1;
622 my $acstr = $self->ac_label;
623 foreach my $acwit ( grep { $_ =~ s/^(.*)\Q$acstr\E$/$1/ } keys %all_witnesses ) {
624 delete $all_witnesses{$acwit.$acstr} if exists $all_witnesses{$acwit};
626 return keys %all_witnesses;
629 =head1 OUTPUT METHODS
631 =head2 as_svg( \%options )
633 Returns an SVG string that represents the graph, via as_dot and graphviz.
634 See as_dot for a list of options. Must have GraphViz (dot) installed to run.
639 my( $self, $opts ) = @_;
640 throw( "Need GraphViz installed to output SVG" )
641 unless File::Which::which( 'dot' );
642 my $want_subgraph = exists $opts->{'from'} || exists $opts->{'to'};
643 $self->calculate_ranks()
644 unless( $self->_graphcalc_done || $opts->{'nocalc'} || !$self->linear );
645 my @cmd = qw/dot -Tsvg/;
647 my $dotfile = File::Temp->new();
649 # $dotfile->unlink_on_destroy(0);
650 binmode $dotfile, ':utf8';
651 print $dotfile $self->as_dot( $opts );
652 push( @cmd, $dotfile->filename );
653 run( \@cmd, ">", binary(), \$svg );
654 $svg = decode_utf8( $svg );
659 =head2 as_dot( \%options )
661 Returns a string that is the collation graph expressed in dot
662 (i.e. GraphViz) format. Options include:
677 my( $self, $opts ) = @_;
678 my $startrank = $opts->{'from'} if $opts;
679 my $endrank = $opts->{'to'} if $opts;
680 my $color_common = $opts->{'color_common'} if $opts;
681 my $STRAIGHTENHACK = !$startrank && !$endrank && $self->end->rank
682 && $self->end->rank > 100;
683 $STRAIGHTENHACK = 1 if $opts->{'straight'}; # even for subgraphs or small graphs
685 # Check the arguments
687 return if $endrank && $startrank > $endrank;
688 return if $startrank > $self->end->rank;
690 if( defined $endrank ) {
691 return if $endrank < 0;
692 $endrank = undef if $endrank == $self->end->rank;
695 my $graph_name = $self->tradition->name;
696 $graph_name =~ s/[^\w\s]//g;
697 $graph_name = join( '_', split( /\s+/, $graph_name ) );
705 'fillcolor' => 'white',
710 'arrowhead' => 'open',
711 'color' => '#000000',
712 'fontcolor' => '#000000',
715 my $dot = sprintf( "digraph %s {\n", $graph_name );
716 $dot .= "\tgraph " . _dot_attr_string( \%graph_attrs ) . ";\n";
717 $dot .= "\tnode " . _dot_attr_string( \%node_attrs ) . ";\n";
719 # Output substitute start/end readings if necessary
721 $dot .= "\t\"__SUBSTART__\" [ label=\"...\",id=\"__START__\" ];\n";
724 $dot .= "\t\"__SUBEND__\" [ label=\"...\",id=\"__END__\" ];\n";
726 if( $STRAIGHTENHACK ) {
728 my $startlabel = $startrank ? '__SUBSTART__' : '__START__';
729 $dot .= "\tsubgraph { rank=same \"$startlabel\" \"#SILENT#\" }\n";
730 $dot .= "\t\"#SILENT#\" [ shape=diamond,color=white,penwidth=0,label=\"\" ];"
732 my %used; # Keep track of the readings that actually appear in the graph
733 # Sort the readings by rank if we have ranks; this speeds layout.
734 my @all_readings = $self->end->has_rank
735 ? sort { $a->rank <=> $b->rank } $self->readings
737 # TODO Refrain from outputting lacuna nodes - just grey out the edges.
738 foreach my $reading ( @all_readings ) {
739 # Only output readings within our rank range.
740 next if $startrank && $reading->rank < $startrank;
741 next if $endrank && $reading->rank > $endrank;
742 $used{$reading->id} = 1;
743 # Need not output nodes without separate labels
744 next if $reading->id eq $reading->text;
746 my $label = $reading->text;
747 $label .= '-' if $reading->join_next;
748 $label = "-$label" if $reading->join_prior;
749 $label =~ s/\"/\\\"/g;
750 $rattrs->{'label'} = $label;
751 $rattrs->{'id'} = $reading->id;
752 $rattrs->{'fillcolor'} = '#b3f36d' if $reading->is_common && $color_common;
753 $dot .= sprintf( "\t\"%s\" %s;\n", $reading->id, _dot_attr_string( $rattrs ) );
756 # Add the real edges. Need to weight one edge per rank jump, in a
758 # my $weighted = $self->_add_edge_weights;
759 my @edges = $self->paths;
760 my( %substart, %subend );
761 foreach my $edge ( @edges ) {
762 # Do we need to output this edge?
763 if( $used{$edge->[0]} && $used{$edge->[1]} ) {
764 my $label = $self->_path_display_label( $self->path_witnesses( $edge ) );
765 my $variables = { %edge_attrs, 'label' => $label };
767 # Account for the rank gap if necessary
768 my $rank0 = $self->reading( $edge->[0] )->rank
769 if $self->reading( $edge->[0] )->has_rank;
770 my $rank1 = $self->reading( $edge->[1] )->rank
771 if $self->reading( $edge->[1] )->has_rank;
772 if( defined $rank0 && defined $rank1 && $rank1 - $rank0 > 1 ) {
773 $variables->{'minlen'} = $rank1 - $rank0;
776 # Add the calculated edge weights
777 # if( exists $weighted->{$edge->[0]}
778 # && $weighted->{$edge->[0]} eq $edge->[1] ) {
779 # # $variables->{'color'} = 'red';
780 # $variables->{'weight'} = 3.0;
783 # EXPERIMENTAL: make edge width reflect no. of witnesses
784 my $extrawidth = scalar( $self->path_witnesses( $edge ) ) * 0.2;
785 $variables->{'penwidth'} = $extrawidth + 0.8; # gives 1 for a single wit
787 my $varopts = _dot_attr_string( $variables );
788 $dot .= sprintf( "\t\"%s\" -> \"%s\" %s;\n",
789 $edge->[0], $edge->[1], $varopts );
790 } elsif( $used{$edge->[0]} ) {
791 $subend{$edge->[0]} = $edge->[1];
792 } elsif( $used{$edge->[1]} ) {
793 $substart{$edge->[1]} = $edge->[0];
796 # Add substitute start and end edges if necessary
797 foreach my $node ( keys %substart ) {
798 my $witstr = $self->_path_display_label ( $self->path_witnesses( $substart{$node}, $node ) );
799 my $variables = { %edge_attrs, 'label' => $witstr };
800 my $nrdg = $self->reading( $node );
801 if( $nrdg->has_rank && $nrdg->rank > $startrank ) {
802 # Substart is actually one lower than $startrank
803 $variables->{'minlen'} = $nrdg->rank - ( $startrank - 1 );
805 my $varopts = _dot_attr_string( $variables );
806 $dot .= "\t\"__SUBSTART__\" -> \"$node\" $varopts;\n";
808 foreach my $node ( keys %subend ) {
809 my $witstr = $self->_path_display_label ( $self->path_witnesses( $node, $subend{$node} ) );
810 my $variables = { %edge_attrs, 'label' => $witstr };
811 my $varopts = _dot_attr_string( $variables );
812 $dot .= "\t\"$node\" -> \"__SUBEND__\" $varopts;\n";
815 if( $STRAIGHTENHACK ) {
816 my $endlabel = $endrank ? '__SUBEND__' : '__END__';
817 $dot .= "\t\"$endlabel\" -> \"#SILENT#\" [ color=white,penwidth=0 ];\n";
824 sub _dot_attr_string {
827 foreach my $k ( sort keys %$hash ) {
829 push( @attrs, $k.'="'.$v.'"' );
831 return( '[ ' . join( ', ', @attrs ) . ' ]' );
834 sub _add_edge_weights {
836 # Walk the graph from START to END, choosing the successor node with
837 # the largest number of witness paths each time.
839 my $curr = $self->start->id;
840 my $ranked = $self->end->has_rank;
841 while( $curr ne $self->end->id ) {
842 my $rank = $ranked ? $self->reading( $curr )->rank : 0;
843 my @succ = sort { $self->path_witnesses( $curr, $a )
844 <=> $self->path_witnesses( $curr, $b ) }
845 $self->sequence->successors( $curr );
846 my $next = pop @succ;
847 my $nextrank = $ranked ? $self->reading( $next )->rank : 0;
848 # Try to avoid lacunae in the weighted path.
850 ( $self->reading( $next )->is_lacuna ||
851 $nextrank - $rank > 1 ) ){
854 $weighted->{$curr} = $next;
860 =head2 path_witnesses( $edge )
862 Returns the list of sigils whose witnesses are associated with the given edge.
863 The edge can be passed as either an array or an arrayref of ( $source, $target ).
868 my( $self, @edge ) = @_;
869 # If edge is an arrayref, cope.
870 if( @edge == 1 && ref( $edge[0] ) eq 'ARRAY' ) {
874 my @wits = keys %{$self->sequence->get_edge_attributes( @edge )};
878 # Helper function. Make a display label for the given witnesses, showing a.c.
879 # witnesses only where the main witness is not also in the list.
880 sub _path_display_label {
883 map { $wits{$_} = 1 } @_;
885 # If an a.c. wit is listed, remove it if the main wit is also listed.
886 # Otherwise keep it for explicit listing.
887 my $aclabel = $self->ac_label;
889 foreach my $w ( sort keys %wits ) {
890 if( $w =~ /^(.*)\Q$aclabel\E$/ ) {
891 if( exists $wits{$1} ) {
894 push( @disp_ac, $w );
899 # See if we are in a majority situation.
900 my $maj = scalar( $self->tradition->witnesses ) * 0.6;
901 $maj = $maj > 5 ? $maj : 5;
902 if( scalar keys %wits > $maj ) {
903 unshift( @disp_ac, 'majority' );
904 return join( ', ', @disp_ac );
906 return join( ', ', sort keys %wits );
910 =head2 readings_at_rank( $rank )
912 Returns a list of readings at a given rank, taken from the alignment table.
916 sub readings_at_rank {
917 my( $self, $rank ) = @_;
918 my $table = $self->alignment_table;
919 # Table rank is real rank - 1.
920 my @elements = map { $_->{'tokens'}->[$rank-1] } @{$table->{'alignment'}};
922 foreach my $e ( @elements ) {
923 next unless ref( $e ) eq 'HASH';
924 next unless exists $e->{'t'};
925 $readings{$e->{'t'}->id} = $e->{'t'};
927 return values %readings;
932 Returns a GraphML representation of the collation. The GraphML will contain
933 two graphs. The first expresses the attributes of the readings and the witness
934 paths that link them; the second expresses the relationships that link the
935 readings. This is the native transfer format for a tradition.
944 my $datafile = 't/data/florilegium_tei_ps.xml';
945 my $tradition = Text::Tradition->new( 'input' => 'TEI',
950 ok( $tradition, "Got a tradition object" );
951 is( scalar $tradition->witnesses, 13, "Found all witnesses" );
952 ok( $tradition->collation, "Tradition has a collation" );
954 my $c = $tradition->collation;
955 is( scalar $c->readings, $READINGS, "Collation has all readings" );
956 is( scalar $c->paths, $PATHS, "Collation has all paths" );
957 is( scalar $c->relationships, 0, "Collation has all relationships" );
959 # Add a few relationships
960 $c->add_relationship( 'w123', 'w125', { 'type' => 'collated' } );
961 $c->add_relationship( 'w193', 'w196', { 'type' => 'collated' } );
962 $c->add_relationship( 'w257', 'w262', { 'type' => 'transposition' } );
964 # Now write it to GraphML and parse it again.
966 my $graphml = $c->as_graphml;
967 my $st = Text::Tradition->new( 'input' => 'Self', 'string' => $graphml );
968 is( scalar $st->collation->readings, $READINGS, "Reparsed collation has all readings" );
969 is( scalar $st->collation->paths, $PATHS, "Reparsed collation has all paths" );
970 is( scalar $st->collation->relationships, 3, "Reparsed collation has new relationships" );
972 # Now add a stemma, write to GraphML, and look at the output.
973 my $stemma = $tradition->add_stemma( 'dotfile' => 't/data/florilegium.dot' );
974 is( ref( $stemma ), 'Text::Tradition::Stemma', "Parsed dotfile into stemma" );
975 is( $tradition->stemmata, 1, "Tradition now has the stemma" );
976 $graphml = $c->as_graphml;
977 like( $graphml, qr/digraph/, "Digraph declaration exists in GraphML" );
979 # Now add a user, write to GraphML, and look at the output.
980 unlike( $graphml, qr/testuser/, "Test user name does not exist in GraphML yet" );
981 my $testuser = Text::Tradition::User->new(
982 id => 'testuser', password => 'testpass' );
983 is( ref( $testuser ), 'Text::Tradition::User', "Created test user object" );
984 $testuser->add_tradition( $tradition );
985 is( $tradition->user->id, $testuser->id, "Tradition assigned to test user" );
986 $graphml = $c->as_graphml;
987 like( $graphml, qr/testuser/, "Test user name now exists in GraphML" );
994 my( $self, $options ) = @_;
995 $self->calculate_ranks unless $self->_graphcalc_done;
997 my $start = $options->{'from'}
998 ? $self->reading( $options->{'from'} ) : $self->start;
999 my $end = $options->{'to'}
1000 ? $self->reading( $options->{'to'} ) : $self->end;
1001 if( $start->has_rank && $end->has_rank && $end->rank < $start->rank ) {
1002 throw( 'Start node must be before end node' );
1004 # The readings need to be ranked for this to work.
1005 $start = $self->start unless $start->has_rank;
1006 $end = $self->end unless $end->has_rank;
1008 unless( $start eq $self->start ) {
1009 $rankoffset = $start->rank - 1;
1014 my $graphml_ns = 'http://graphml.graphdrawing.org/xmlns';
1015 my $xsi_ns = 'http://www.w3.org/2001/XMLSchema-instance';
1016 my $graphml_schema = 'http://graphml.graphdrawing.org/xmlns ' .
1017 'http://graphml.graphdrawing.org/xmlns/1.0/graphml.xsd';
1019 # Create the document and root node
1020 require XML::LibXML;
1021 my $graphml = XML::LibXML->createDocument( "1.0", "UTF-8" );
1022 my $root = $graphml->createElementNS( $graphml_ns, 'graphml' );
1023 $graphml->setDocumentElement( $root );
1024 $root->setNamespace( $xsi_ns, 'xsi', 0 );
1025 $root->setAttributeNS( $xsi_ns, 'schemaLocation', $graphml_schema );
1027 # List of attribute types to save on our objects and their corresponding
1032 'Bool' => 'boolean',
1033 'ReadingID' => 'string',
1034 'RelationshipType' => 'string',
1035 'RelationshipScope' => 'string',
1038 # Add the data keys for the graph. Include an extra key 'version' for the
1039 # GraphML output version.
1040 my %graph_data_keys;
1042 my %graph_attributes = ( 'version' => 'string' );
1043 # Graph attributes include those of Tradition and those of Collation.
1045 my $tmeta = $self->tradition->meta;
1046 my $cmeta = $self->meta;
1047 map { $gattr_from{$_->name} = 'Tradition' } $tmeta->get_all_attributes;
1048 map { $gattr_from{$_->name} = 'Collation' } $cmeta->get_all_attributes;
1049 foreach my $attr ( ( $tmeta->get_all_attributes, $cmeta->get_all_attributes ) ) {
1050 next if $attr->name =~ /^_/;
1051 next unless $save_types{$attr->type_constraint->name};
1052 $graph_attributes{$attr->name} = $save_types{$attr->type_constraint->name};
1054 # Extra custom keys for complex objects that should be saved in some form.
1055 # The subroutine should return a string, or undef/empty.
1056 $graph_attributes{'stemmata'} = sub {
1058 map { push( @stemstrs, $_->editable( {linesep => ''} ) ) }
1059 $self->tradition->stemmata;
1060 join( "\n", @stemstrs );
1062 $graph_attributes{'user'} = sub {
1063 $self->tradition->user ? $self->tradition->user->id : undef
1066 foreach my $datum ( sort keys %graph_attributes ) {
1067 $graph_data_keys{$datum} = 'dg'.$gdi++;
1068 my $key = $root->addNewChild( $graphml_ns, 'key' );
1069 my $dtype = ref( $graph_attributes{$datum} ) ? 'string'
1070 : $graph_attributes{$datum};
1071 $key->setAttribute( 'attr.name', $datum );
1072 $key->setAttribute( 'attr.type', $dtype );
1073 $key->setAttribute( 'for', 'graph' );
1074 $key->setAttribute( 'id', $graph_data_keys{$datum} );
1077 # Add the data keys for reading nodes
1078 my %reading_attributes;
1079 my $rmeta = Text::Tradition::Collation::Reading->meta;
1080 foreach my $attr( $rmeta->get_all_attributes ) {
1081 next if $attr->name =~ /^_/;
1082 next unless $save_types{$attr->type_constraint->name};
1083 $reading_attributes{$attr->name} = $save_types{$attr->type_constraint->name};
1085 # Extra custom key for the reading morphology
1086 $reading_attributes{'lexemes'} = 'string';
1090 foreach my $datum ( sort keys %reading_attributes ) {
1091 $node_data_keys{$datum} = 'dn'.$ndi++;
1092 my $key = $root->addNewChild( $graphml_ns, 'key' );
1093 $key->setAttribute( 'attr.name', $datum );
1094 $key->setAttribute( 'attr.type', $reading_attributes{$datum} );
1095 $key->setAttribute( 'for', 'node' );
1096 $key->setAttribute( 'id', $node_data_keys{$datum} );
1099 # Add the data keys for edges, that is, paths and relationships. Path
1100 # data does not come from a Moose class so is here manually.
1103 my %edge_attributes = (
1104 witness => 'string', # ID/label for a path
1105 extra => 'boolean', # Path key
1107 my @path_attributes = keys %edge_attributes; # track our manual additions
1108 my $pmeta = Text::Tradition::Collation::Relationship->meta;
1109 foreach my $attr( $pmeta->get_all_attributes ) {
1110 next if $attr->name =~ /^_/;
1111 next unless $save_types{$attr->type_constraint->name};
1112 $edge_attributes{$attr->name} = $save_types{$attr->type_constraint->name};
1114 foreach my $datum ( sort keys %edge_attributes ) {
1115 $edge_data_keys{$datum} = 'de'.$edi++;
1116 my $key = $root->addNewChild( $graphml_ns, 'key' );
1117 $key->setAttribute( 'attr.name', $datum );
1118 $key->setAttribute( 'attr.type', $edge_attributes{$datum} );
1119 $key->setAttribute( 'for', 'edge' );
1120 $key->setAttribute( 'id', $edge_data_keys{$datum} );
1123 # Add the collation graph itself. First, sanitize the name to a valid XML ID.
1124 my $xmlidname = $self->tradition->name;
1125 $xmlidname =~ s/(?!$xml10_namechar_rx)./_/g;
1126 if( $xmlidname !~ /^$xml10_namestartchar_rx/ ) {
1127 $xmlidname = '_'.$xmlidname;
1129 my $sgraph = $root->addNewChild( $graphml_ns, 'graph' );
1130 $sgraph->setAttribute( 'edgedefault', 'directed' );
1131 $sgraph->setAttribute( 'id', $xmlidname );
1132 $sgraph->setAttribute( 'parse.edgeids', 'canonical' );
1133 $sgraph->setAttribute( 'parse.edges', 0 ); # fill in later
1134 $sgraph->setAttribute( 'parse.nodeids', 'canonical' );
1135 $sgraph->setAttribute( 'parse.nodes', 0 ); # fill in later
1136 $sgraph->setAttribute( 'parse.order', 'nodesfirst' );
1138 # Tradition/collation attribute data
1139 foreach my $datum ( keys %graph_attributes ) {
1141 if( $datum eq 'version' ) {
1143 } elsif( ref( $graph_attributes{$datum} ) ) {
1144 my $sub = $graph_attributes{$datum};
1146 } elsif( $gattr_from{$datum} eq 'Tradition' ) {
1147 $value = $self->tradition->$datum;
1149 $value = $self->$datum;
1151 _add_graphml_data( $sgraph, $graph_data_keys{$datum}, $value );
1156 # Add our readings to the graph
1157 foreach my $n ( sort { $a->id cmp $b->id } $self->readings ) {
1158 next if $n->has_rank && $n ne $self->start && $n ne $self->end &&
1159 ( $n->rank < $start->rank || $n->rank > $end->rank );
1160 $use_readings{$n->id} = 1;
1161 # Add to the main graph
1162 my $node_el = $sgraph->addNewChild( $graphml_ns, 'node' );
1163 my $node_xmlid = 'n' . $node_ctr++;
1164 $node_hash{ $n->id } = $node_xmlid;
1165 $node_el->setAttribute( 'id', $node_xmlid );
1166 foreach my $d ( keys %reading_attributes ) {
1168 # Custom serialization
1169 if( $d eq 'lexemes' ) {
1170 # If nval is a true value, we have lexemes so we need to
1171 # serialize them. Otherwise set nval to undef so that the
1172 # key is excluded from this reading.
1173 $nval = $nval ? $n->_serialize_lexemes : undef;
1174 } elsif( $d eq 'normal_form' && $n->normal_form eq $n->text ) {
1177 if( $rankoffset && $d eq 'rank' && $n ne $self->start ) {
1178 # Adjust the ranks within the subgraph.
1179 $nval = $n eq $self->end ? $end->rank - $rankoffset + 1
1180 : $nval - $rankoffset;
1182 _add_graphml_data( $node_el, $node_data_keys{$d}, $nval )
1187 # Add the path edges to the sequence graph
1189 foreach my $e ( sort { $a->[0] cmp $b->[0] } $self->sequence->edges() ) {
1190 # We add an edge in the graphml for every witness in $e.
1191 next unless( $use_readings{$e->[0]} || $use_readings{$e->[1]} );
1192 my @edge_wits = sort $self->path_witnesses( $e );
1193 $e->[0] = $self->start->id unless $use_readings{$e->[0]};
1194 $e->[1] = $self->end->id unless $use_readings{$e->[1]};
1195 # Skip any path from start to end; that witness is not in the subgraph.
1196 next if ( $e->[0] eq $self->start->id && $e->[1] eq $self->end->id );
1197 foreach my $wit ( @edge_wits ) {
1198 my( $id, $from, $to ) = ( 'e'.$edge_ctr++,
1199 $node_hash{ $e->[0] },
1200 $node_hash{ $e->[1] } );
1201 my $edge_el = $sgraph->addNewChild( $graphml_ns, 'edge' );
1202 $edge_el->setAttribute( 'source', $from );
1203 $edge_el->setAttribute( 'target', $to );
1204 $edge_el->setAttribute( 'id', $id );
1206 # It's a witness path, so add the witness
1208 my $key = $edge_data_keys{'witness'};
1209 # Is this an ante-corr witness?
1210 my $aclabel = $self->ac_label;
1211 if( $wit =~ /^(.*)\Q$aclabel\E$/ ) {
1212 # Keep the base witness
1214 # ...and record that this is an 'extra' reading path
1215 _add_graphml_data( $edge_el, $edge_data_keys{'extra'}, $aclabel );
1217 _add_graphml_data( $edge_el, $edge_data_keys{'witness'}, $base );
1221 # Report the actual number of nodes and edges that went in
1222 $sgraph->setAttribute( 'parse.edges', $edge_ctr );
1223 $sgraph->setAttribute( 'parse.nodes', $node_ctr );
1225 # Add the relationship graph to the XML
1226 map { delete $edge_data_keys{$_} } @path_attributes;
1227 $self->relations->_as_graphml( $graphml_ns, $root, \%node_hash,
1228 $node_data_keys{'id'}, \%edge_data_keys );
1230 # Save and return the thing
1231 my $result = decode_utf8( $graphml->toString(1) );
1235 sub _add_graphml_data {
1236 my( $el, $key, $value ) = @_;
1237 return unless defined $value;
1238 my $data_el = $el->addNewChild( $el->namespaceURI, 'data' );
1239 $data_el->setAttribute( 'key', $key );
1240 $data_el->appendText( $value );
1245 Returns a CSV alignment table representation of the collation graph, one
1246 row per witness (or witness uncorrected.)
1252 my $table = $self->alignment_table;
1253 my $csv = Text::CSV->new( { binary => 1, quote_null => 0 } );
1255 # Make the header row
1256 $csv->combine( map { $_->{'witness'} } @{$table->{'alignment'}} );
1257 push( @result, decode_utf8( $csv->string ) );
1258 # Make the rest of the rows
1259 foreach my $idx ( 0 .. $table->{'length'} - 1 ) {
1260 my @rowobjs = map { $_->{'tokens'}->[$idx] } @{$table->{'alignment'}};
1261 my @row = map { $_ ? $_->{'t'}->text : $_ } @rowobjs;
1262 $csv->combine( @row );
1263 push( @result, decode_utf8( $csv->string ) );
1265 return join( "\n", @result );
1268 =head2 alignment_table( $use_refs, $include_witnesses )
1270 Return a reference to an alignment table, in a slightly enhanced CollateX
1271 format which looks like this:
1273 $table = { alignment => [ { witness => "SIGIL",
1274 tokens => [ { t => "TEXT" }, ... ] },
1275 { witness => "SIG2",
1276 tokens => [ { t => "TEXT" }, ... ] },
1278 length => TEXTLEN };
1280 If $use_refs is set to 1, the reading object is returned in the table
1281 instead of READINGTEXT; if not, the text of the reading is returned.
1283 If $include_witnesses is set to a hashref, only the witnesses whose sigil
1284 keys have a true hash value will be included.
1288 sub alignment_table {
1290 $self->calculate_ranks() unless $self->_graphcalc_done;
1291 return $self->cached_table if $self->has_cached_table;
1293 # Make sure we can do this
1294 throw( "Need a linear graph in order to make an alignment table" )
1295 unless $self->linear;
1296 $self->calculate_ranks unless $self->end->has_rank;
1298 my $table = { 'alignment' => [], 'length' => $self->end->rank - 1 };
1299 my @all_pos = ( 1 .. $self->end->rank - 1 );
1300 foreach my $wit ( sort { $a->sigil cmp $b->sigil } $self->tradition->witnesses ) {
1301 # say STDERR "Making witness row(s) for " . $wit->sigil;
1302 my @wit_path = $self->reading_sequence( $self->start, $self->end, $wit->sigil );
1303 my @row = _make_witness_row( \@wit_path, \@all_pos );
1304 push( @{$table->{'alignment'}},
1305 { 'witness' => $wit->sigil, 'tokens' => \@row } );
1306 if( $wit->is_layered ) {
1307 my @wit_ac_path = $self->reading_sequence( $self->start, $self->end,
1308 $wit->sigil.$self->ac_label );
1309 my @ac_row = _make_witness_row( \@wit_ac_path, \@all_pos );
1310 push( @{$table->{'alignment'}},
1311 { 'witness' => $wit->sigil.$self->ac_label, 'tokens' => \@ac_row } );
1314 $self->cached_table( $table );
1318 sub _make_witness_row {
1319 my( $path, $positions ) = @_;
1321 map { $char_hash{$_} = undef } @$positions;
1323 foreach my $rdg ( @$path ) {
1324 my $rtext = $rdg->text;
1325 $rtext = '#LACUNA#' if $rdg->is_lacuna;
1326 say STDERR "rank " . $rdg->rank if $debug;
1327 # say STDERR "No rank for " . $rdg->id unless defined $rdg->rank;
1328 $char_hash{$rdg->rank} = { 't' => $rdg };
1330 my @row = map { $char_hash{$_} } @$positions;
1331 # Fill in lacuna markers for undef spots in the row
1332 my $last_el = shift @row;
1333 my @filled_row = ( $last_el );
1334 foreach my $el ( @row ) {
1335 # If we are using node reference, make the lacuna node appear many times
1336 # in the table. If not, use the lacuna tag.
1337 if( $last_el && $last_el->{'t'}->is_lacuna && !defined $el ) {
1340 push( @filled_row, $el );
1346 =head1 NAVIGATION METHODS
1348 =head2 reading_sequence( $first, $last, $sigil, $backup )
1350 Returns the ordered list of readings, starting with $first and ending
1351 with $last, for the witness given in $sigil. If a $backup sigil is
1352 specified (e.g. when walking a layered witness), it will be used wherever
1353 no $sigil path exists. If there is a base text reading, that will be
1354 used wherever no path exists for $sigil or $backup.
1358 # TODO Think about returning some lazy-eval iterator.
1359 # TODO Get rid of backup; we should know from what witness is whether we need it.
1361 sub reading_sequence {
1362 my( $self, $start, $end, $witness ) = @_;
1364 $witness = $self->baselabel unless $witness;
1365 my @readings = ( $start );
1368 while( $n && $n->id ne $end->id ) {
1369 if( exists( $seen{$n->id} ) ) {
1370 throw( "Detected loop for $witness at " . $n->id );
1374 my $next = $self->next_reading( $n, $witness );
1376 throw( "Did not find any path for $witness from reading " . $n->id );
1378 push( @readings, $next );
1381 # Check that the last reading is our end reading.
1382 my $last = $readings[$#readings];
1383 throw( "Last reading found from " . $start->text .
1384 " for witness $witness is not the end!" ) # TODO do we get this far?
1385 unless $last->id eq $end->id;
1390 =head2 next_reading( $reading, $sigil );
1392 Returns the reading that follows the given reading along the given witness
1398 # Return the successor via the corresponding path.
1400 my $answer = $self->_find_linked_reading( 'next', @_ );
1401 return undef unless $answer;
1402 return $self->reading( $answer );
1405 =head2 prior_reading( $reading, $sigil )
1407 Returns the reading that precedes the given reading along the given witness
1413 # Return the predecessor via the corresponding path.
1415 my $answer = $self->_find_linked_reading( 'prior', @_ );
1416 return $self->reading( $answer );
1419 sub _find_linked_reading {
1420 my( $self, $direction, $node, $path ) = @_;
1422 # Get a backup if we are dealing with a layered witness
1424 my $aclabel = $self->ac_label;
1425 if( $path && $path =~ /^(.*)\Q$aclabel\E$/ ) {
1429 my @linked_paths = $direction eq 'next'
1430 ? $self->sequence->edges_from( $node )
1431 : $self->sequence->edges_to( $node );
1432 return undef unless scalar( @linked_paths );
1434 # We have to find the linked path that contains all of the
1435 # witnesses supplied in $path.
1436 my( @path_wits, @alt_path_wits );
1437 @path_wits = sort( $self->_witnesses_of_label( $path ) ) if $path;
1438 @alt_path_wits = sort( $self->_witnesses_of_label( $alt_path ) ) if $alt_path;
1441 foreach my $le ( @linked_paths ) {
1442 if( $self->sequence->has_edge_attribute( @$le, $self->baselabel ) ) {
1445 my @le_wits = sort $self->path_witnesses( $le );
1446 if( _is_within( \@path_wits, \@le_wits ) ) {
1447 # This is the right path.
1448 return $direction eq 'next' ? $le->[1] : $le->[0];
1449 } elsif( _is_within( \@alt_path_wits, \@le_wits ) ) {
1453 # Got this far? Return the alternate path if it exists.
1454 return $direction eq 'next' ? $alt_le->[1] : $alt_le->[0]
1457 # Got this far? Return the base path if it exists.
1458 return $direction eq 'next' ? $base_le->[1] : $base_le->[0]
1461 # Got this far? We have no appropriate path.
1462 warn "Could not find $direction node from " . $node->id
1463 . " along path $path";
1469 my( $set1, $set2 ) = @_;
1470 my $ret = @$set1; # will be 0, i.e. false, if set1 is empty
1471 foreach my $el ( @$set1 ) {
1472 $ret = 0 unless grep { /^\Q$el\E$/ } @$set2;
1477 # Return the string that joins together a list of witnesses for
1478 # display on a single path.
1479 sub _witnesses_of_label {
1480 my( $self, $label ) = @_;
1481 my $regex = $self->wit_list_separator;
1482 my @answer = split( /\Q$regex\E/, $label );
1486 =head2 common_readings
1488 Returns the list of common readings in the graph (i.e. those readings that are
1489 shared by all non-lacunose witnesses.)
1493 sub common_readings {
1495 my @common = grep { $_->is_common } $self->readings;
1499 =head2 path_text( $sigil, [, $start, $end ] )
1501 Returns the text of a witness (plus its backup, if we are using a layer)
1502 as stored in the collation. The text is returned as a string, where the
1503 individual readings are joined with spaces and the meta-readings (e.g.
1504 lacunae) are omitted. Optional specification of $start and $end allows
1505 the generation of a subset of the witness text.
1510 my( $self, $wit, $start, $end ) = @_;
1511 $start = $self->start unless $start;
1512 $end = $self->end unless $end;
1513 my @path = grep { !$_->is_meta } $self->reading_sequence( $start, $end, $wit );
1516 foreach my $r ( @path ) {
1517 unless ( $r->join_prior || !$last || $last->join_next ) {
1520 $pathtext .= $r->text;
1526 =head1 INITIALIZATION METHODS
1528 These are mostly for use by parsers.
1530 =head2 make_witness_path( $witness )
1532 Link the array of readings contained in $witness->path (and in
1533 $witness->uncorrected_path if it exists) into collation paths.
1534 Clear out the arrays when finished.
1536 =head2 make_witness_paths
1538 Call make_witness_path for all witnesses in the tradition.
1542 # For use when a collation is constructed from a base text and an apparatus.
1543 # We have the sequences of readings and just need to add path edges.
1544 # When we are done, clear out the witness path attributes, as they are no
1546 # TODO Find a way to replace the witness path attributes with encapsulated functions?
1548 sub make_witness_paths {
1550 foreach my $wit ( $self->tradition->witnesses ) {
1551 # say STDERR "Making path for " . $wit->sigil;
1552 $self->make_witness_path( $wit );
1556 sub make_witness_path {
1557 my( $self, $wit ) = @_;
1558 my @chain = @{$wit->path};
1559 my $sig = $wit->sigil;
1560 # Add start and end if necessary
1561 unshift( @chain, $self->start ) unless $chain[0] eq $self->start;
1562 push( @chain, $self->end ) unless $chain[-1] eq $self->end;
1563 foreach my $idx ( 0 .. $#chain-1 ) {
1564 $self->add_path( $chain[$idx], $chain[$idx+1], $sig );
1566 if( $wit->is_layered ) {
1567 @chain = @{$wit->uncorrected_path};
1568 unshift( @chain, $self->start ) unless $chain[0] eq $self->start;
1569 push( @chain, $self->end ) unless $chain[-1] eq $self->end;
1570 foreach my $idx( 0 .. $#chain-1 ) {
1571 my $source = $chain[$idx];
1572 my $target = $chain[$idx+1];
1573 $self->add_path( $source, $target, $sig.$self->ac_label )
1574 unless $self->has_path( $source, $target, $sig );
1578 $wit->clear_uncorrected_path;
1581 =head2 calculate_ranks
1583 Calculate the reading ranks (that is, their aligned positions relative
1584 to each other) for the graph. This can only be called on linear collations.
1588 use Text::Tradition;
1590 my $cxfile = 't/data/Collatex-16.xml';
1591 my $t = Text::Tradition->new(
1593 'input' => 'CollateX',
1596 my $c = $t->collation;
1599 my $table = $c->alignment_table;
1600 ok( $c->has_cached_table, "Alignment table was cached" );
1601 is( $c->alignment_table, $table, "Cached table returned upon second call" );
1602 $c->calculate_ranks;
1603 is( $c->alignment_table, $table, "Cached table retained with no rank change" );
1604 $c->add_relationship( 'n24', 'n23', { 'type' => 'spelling' } );
1605 isnt( $c->alignment_table, $table, "Alignment table changed after relationship add" );
1611 sub calculate_ranks {
1613 # Save the existing ranks, in case we need to invalidate the cached SVG.
1615 map { $existing_ranks{$_} = $_->rank } $self->readings;
1617 # Do the rankings based on the relationship equivalence graph, starting
1618 # with the start node.
1619 my ( $node_ranks, $rank_nodes ) = $self->relations->equivalence_ranks();
1621 # Transfer our rankings from the topological graph to the real one.
1622 foreach my $r ( $self->readings ) {
1623 if( defined $node_ranks->{$self->equivalence( $r->id )} ) {
1624 $r->rank( $node_ranks->{$self->equivalence( $r->id )} );
1626 # Die. Find the last rank we calculated.
1627 my @all_defined = sort { ( $node_ranks->{$self->equivalence( $a->id )}||-1 )
1628 <=> ( $node_ranks->{$self->equivalence( $b->id )}||-1 ) }
1630 my $last = pop @all_defined;
1631 throw( "Ranks not calculated after $last - do you have a cycle in the graph?" );
1634 # Do we need to invalidate the cached data?
1635 if( $self->has_cached_table ) {
1636 foreach my $r ( $self->readings ) {
1637 next if defined( $existing_ranks{$r} )
1638 && $existing_ranks{$r} == $r->rank;
1639 # Something has changed, so clear the cache
1640 $self->_clear_cache;
1641 # ...and recalculate the common readings.
1642 $self->calculate_common_readings();
1646 # The graph calculation information is now up to date.
1647 $self->_graphcalc_done(1);
1652 $self->wipe_table if $self->has_cached_table;
1656 =head2 flatten_ranks
1658 A convenience method for parsing collation data. Searches the graph for readings
1659 with the same text at the same rank, and merges any that are found.
1665 my %unique_rank_rdg;
1667 foreach my $rdg ( $self->readings ) {
1668 next unless $rdg->has_rank;
1669 my $key = $rdg->rank . "||" . $rdg->text;
1670 if( exists $unique_rank_rdg{$key} ) {
1671 # Make sure they don't have different grammatical forms
1672 my $ur = $unique_rank_rdg{$key};
1673 if( $rdg->disambiguated && $ur->disambiguated ) {
1674 my $rform = join( '//', map { $_->form->to_string } $rdg->lexemes );
1675 my $uform = join( '//', map { $_->form->to_string } $ur->lexemes );
1676 next unless $rform eq $uform;
1677 } elsif( $rdg->disambiguated xor $ur->disambiguated ) {
1681 #say STDERR "Combining readings at same rank: $key";
1683 $self->merge_readings( $unique_rank_rdg{$key}, $rdg );
1684 # TODO see if this now makes a common point.
1686 $unique_rank_rdg{$key} = $rdg;
1689 # If we merged readings, the ranks are still fine but the alignment
1690 # table is wrong. Wipe it.
1691 $self->wipe_table() if $changed;
1695 =head2 calculate_common_readings
1697 Goes through the graph identifying the readings that appear in every witness
1698 (apart from those with lacunae at that spot.) Marks them as common and returns
1703 use Text::Tradition;
1705 my $cxfile = 't/data/Collatex-16.xml';
1706 my $t = Text::Tradition->new(
1708 'input' => 'CollateX',
1711 my $c = $t->collation;
1713 my @common = $c->calculate_common_readings();
1714 is( scalar @common, 8, "Found correct number of common readings" );
1715 my @marked = sort $c->common_readings();
1716 is( scalar @common, 8, "All common readings got marked as such" );
1717 my @expected = qw/ n1 n11 n16 n19 n20 n5 n6 n7 /;
1718 is_deeply( \@marked, \@expected, "Found correct list of common readings" );
1724 sub calculate_common_readings {
1727 map { $_->is_common( 0 ) } $self->readings;
1728 # Implicitly calls calculate_ranks
1729 my $table = $self->alignment_table;
1730 foreach my $idx ( 0 .. $table->{'length'} - 1 ) {
1731 my @row = map { $_->{'tokens'}->[$idx]
1732 ? $_->{'tokens'}->[$idx]->{'t'} : '' }
1733 @{$table->{'alignment'}};
1735 foreach my $r ( @row ) {
1737 $hash{$r->id} = $r unless $r->is_meta;
1739 $hash{'UNDEF'} = $r;
1742 if( keys %hash == 1 && !exists $hash{'UNDEF'} ) {
1743 my( $r ) = values %hash;
1745 push( @common, $r );
1751 =head2 text_from_paths
1753 Calculate the text array for all witnesses from the path, for later consistency
1754 checking. Only to be used if there is no non-graph-based way to know the
1759 sub text_from_paths {
1761 foreach my $wit ( $self->tradition->witnesses ) {
1762 my @readings = $self->reading_sequence( $self->start, $self->end, $wit->sigil );
1764 foreach my $r ( @readings ) {
1765 next if $r->is_meta;
1766 push( @text, $r->text );
1768 $wit->text( \@text );
1769 if( $wit->is_layered ) {
1770 my @ucrdgs = $self->reading_sequence( $self->start, $self->end,
1771 $wit->sigil.$self->ac_label );
1773 foreach my $r ( @ucrdgs ) {
1774 next if $r->is_meta;
1775 push( @uctext, $r->text );
1777 $wit->layertext( \@uctext );
1782 =head1 UTILITY FUNCTIONS
1784 =head2 common_predecessor( $reading_a, $reading_b )
1786 Find the last reading that occurs in sequence before both the given readings.
1787 At the very least this should be $self->start.
1789 =head2 common_successor( $reading_a, $reading_b )
1791 Find the first reading that occurs in sequence after both the given readings.
1792 At the very least this should be $self->end.
1796 use Text::Tradition;
1798 my $cxfile = 't/data/Collatex-16.xml';
1799 my $t = Text::Tradition->new(
1801 'input' => 'CollateX',
1804 my $c = $t->collation;
1806 is( $c->common_predecessor( 'n24', 'n23' )->id,
1807 'n20', "Found correct common predecessor" );
1808 is( $c->common_successor( 'n24', 'n23' )->id,
1809 '__END__', "Found correct common successor" );
1811 is( $c->common_predecessor( 'n19', 'n17' )->id,
1812 'n16', "Found correct common predecessor for readings on same path" );
1813 is( $c->common_successor( 'n21', 'n10' )->id,
1814 '__END__', "Found correct common successor for readings on same path" );
1820 ## Return the closest reading that is a predecessor of both the given readings.
1821 sub common_predecessor {
1823 my( $r1, $r2 ) = $self->_objectify_args( @_ );
1824 return $self->_common_in_path( $r1, $r2, 'predecessors' );
1827 sub common_successor {
1829 my( $r1, $r2 ) = $self->_objectify_args( @_ );
1830 return $self->_common_in_path( $r1, $r2, 'successors' );
1834 # TODO think about how to do this without ranks...
1835 sub _common_in_path {
1836 my( $self, $r1, $r2, $dir ) = @_;
1837 my $iter = $self->end->rank;
1839 my @last_r1 = ( $r1 );
1840 my @last_r2 = ( $r2 );
1841 # my %all_seen = ( $r1 => 'r1', $r2 => 'r2' );
1843 # say STDERR "Finding common $dir for $r1, $r2";
1844 while( !@candidates ) {
1845 last unless $iter--; # Avoid looping infinitely
1846 # Iterate separately down the graph from r1 and r2
1847 my( @new_lc1, @new_lc2 );
1848 foreach my $lc ( @last_r1 ) {
1849 foreach my $p ( $lc->$dir ) {
1850 if( $all_seen{$p->id} && $all_seen{$p->id} ne 'r1' ) {
1851 # say STDERR "Path candidate $p from $lc";
1852 push( @candidates, $p );
1853 } elsif( !$all_seen{$p->id} ) {
1854 $all_seen{$p->id} = 'r1';
1855 push( @new_lc1, $p );
1859 foreach my $lc ( @last_r2 ) {
1860 foreach my $p ( $lc->$dir ) {
1861 if( $all_seen{$p->id} && $all_seen{$p->id} ne 'r2' ) {
1862 # say STDERR "Path candidate $p from $lc";
1863 push( @candidates, $p );
1864 } elsif( !$all_seen{$p->id} ) {
1865 $all_seen{$p->id} = 'r2';
1866 push( @new_lc2, $p );
1870 @last_r1 = @new_lc1;
1871 @last_r2 = @new_lc2;
1873 my @answer = sort { $a->rank <=> $b->rank } @candidates;
1874 return $dir eq 'predecessors' ? pop( @answer ) : shift ( @answer );
1878 Text::Tradition::Error->throw(
1879 'ident' => 'Collation error',
1885 __PACKAGE__->meta->make_immutable;
1889 This package is free software and is provided "as is" without express
1890 or implied warranty. You can redistribute it and/or modify it under
1891 the same terms as Perl itself.
1895 Tara L Andrews E<lt>aurum@cpan.orgE<gt>