1 package Text::Tradition::Collation;
3 use Encode qw( decode_utf8 );
6 use IPC::Run qw( run binary );
8 use Text::Tradition::Collation::Reading;
9 use Text::Tradition::Collation::RelationshipStore;
10 use Text::Tradition::Error;
17 default => sub { Graph->new() },
25 isa => 'Text::Tradition::Collation::RelationshipStore',
27 relationships => 'relationships',
28 related_readings => 'related_readings',
30 writer => '_set_relations',
35 isa => 'Text::Tradition',
40 isa => 'HashRef[Text::Tradition::Collation::Reading]',
44 _add_reading => 'set',
45 del_reading => 'delete',
46 has_reading => 'exists',
49 default => sub { {} },
52 has 'wit_list_separator' => (
61 default => 'base text',
78 isa => 'Text::Tradition::Collation::Reading',
79 writer => '_set_start',
85 isa => 'Text::Tradition::Collation::Reading',
92 Text::Tradition::Collation - a software model for a text collation
97 my $t = Text::Tradition->new(
98 'name' => 'this is a text',
100 'file' => '/path/to/tei_parallel_seg_file.xml' );
102 my $c = $t->collation;
103 my @readings = $c->readings;
104 my @paths = $c->paths;
105 my @relationships = $c->relationships;
107 my $svg_variant_graph = $t->collation->as_svg();
111 Text::Tradition is a library for representation and analysis of collated
112 texts, particularly medieval ones. The Collation is the central feature of
113 a Tradition, where the text, its sequence of readings, and its relationships
114 between readings are actually kept.
120 The constructor. Takes a hash or hashref of the following arguments:
124 =item * tradition - The Text::Tradition object to which the collation
127 =item * linear - Whether the collation should be linear; that is, whether
128 transposed readings should be treated as two linked readings rather than one,
129 and therefore whether the collation graph is acyclic. Defaults to true.
131 =item * baselabel - The default label for the path taken by a base text
132 (if any). Defaults to 'base text'.
134 =item * wit_list_separator - The string to join a list of witnesses for
135 purposes of making labels in display graphs. Defaults to ', '.
137 =item * ac_label - The extra label to tack onto a witness sigil when
138 representing another layer of path for the given witness - that is, when
139 a text has more than one possible reading due to scribal corrections or
140 the like. Defaults to ' (a.c.)'.
150 =head2 wit_list_separator
156 Simple accessors for collation attributes.
160 The meta-reading at the start of every witness path.
164 The meta-reading at the end of every witness path.
168 Returns all Reading objects in the graph.
170 =head2 reading( $id )
172 Returns the Reading object corresponding to the given ID.
174 =head2 add_reading( $reading_args )
176 Adds a new reading object to the collation.
177 See L<Text::Tradition::Collation::Reading> for the available arguments.
179 =head2 del_reading( $object_or_id )
181 Removes the given reading from the collation, implicitly removing its
182 paths and relationships.
184 =head2 merge_readings( $main, $second )
186 Merges the $second reading into the $main one.
187 The arguments may be either readings or reading IDs.
189 =head2 has_reading( $id )
191 Predicate to see whether a given reading ID is in the graph.
193 =head2 reading_witnesses( $object_or_id )
195 Returns a list of sigils whose witnesses contain the reading.
199 Returns all reading paths within the document - that is, all edges in the
200 collation graph. Each path is an arrayref of [ $source, $target ] reading IDs.
202 =head2 add_path( $source, $target, $sigil )
204 Links the given readings in the collation in sequence, under the given witness
205 sigil. The readings may be specified by object or ID.
207 =head2 del_path( $source, $target, $sigil )
209 Links the given readings in the collation in sequence, under the given witness
210 sigil. The readings may be specified by object or ID.
212 =head2 has_path( $source, $target );
214 Returns true if the two readings are linked in sequence in any witness.
215 The readings may be specified by object or ID.
219 Returns all Relationship objects in the collation.
221 =head2 add_relationship( $reading, $other_reading, $options )
223 Adds a new relationship of the type given in $options between the two readings,
224 which may be specified by object or ID. Returns a value of ( $status, @vectors)
225 where $status is true on success, and @vectors is a list of relationship edges
226 that were ultimately added.
227 See L<Text::Tradition::Collation::Relationship> for the available options.
233 $self->_set_relations( Text::Tradition::Collation::RelationshipStore->new( 'collation' => $self ) );
234 $self->_set_start( $self->add_reading( { 'collation' => $self, 'is_start' => 1 } ) );
235 $self->_set_end( $self->add_reading( { 'collation' => $self, 'is_end' => 1 } ) );
238 ### Reading construct/destruct functions
241 my( $self, $reading ) = @_;
242 unless( ref( $reading ) eq 'Text::Tradition::Collation::Reading' ) {
243 my %args = %$reading;
244 $reading = Text::Tradition::Collation::Reading->new(
245 'collation' => $self,
248 # First check to see if a reading with this ID exists.
249 if( $self->reading( $reading->id ) ) {
250 throw( "Collation already has a reading with id " . $reading->id );
252 $self->_add_reading( $reading->id => $reading );
253 # Once the reading has been added, put it in both graphs.
254 $self->sequence->add_vertex( $reading->id );
255 $self->relations->add_reading( $reading->id );
259 around del_reading => sub {
264 if( ref( $arg ) eq 'Text::Tradition::Collation::Reading' ) {
267 # Remove the reading from the graphs.
268 $self->sequence->delete_vertex( $arg );
269 $self->relations->delete_reading( $arg );
272 $self->$orig( $arg );
275 # merge_readings( $main, $to_be_deleted );
280 # We only need the IDs for adding paths to the graph, not the reading
281 # objects themselves.
282 my( $kept, $deleted, $combine_char ) = $self->_stringify_args( @_ );
284 # The kept reading should inherit the paths and the relationships
285 # of the deleted reading.
286 foreach my $path ( $self->sequence->edges_at( $deleted ) ) {
287 my @vector = ( $kept );
288 push( @vector, $path->[1] ) if $path->[0] eq $deleted;
289 unshift( @vector, $path->[0] ) if $path->[1] eq $deleted;
290 next if $vector[0] eq $vector[1]; # Don't add a self loop
291 my %wits = %{$self->sequence->get_edge_attributes( @$path )};
292 $self->sequence->add_edge( @vector );
293 my $fwits = $self->sequence->get_edge_attributes( @vector );
294 @wits{keys %$fwits} = values %$fwits;
295 $self->sequence->set_edge_attributes( @vector, \%wits );
297 $self->relations->merge_readings( $kept, $deleted, $combine_char );
299 # Do the deletion deed.
300 if( $combine_char ) {
301 my $kept_obj = $self->reading( $kept );
302 my $new_text = join( $combine_char, $kept_obj->text,
303 $self->reading( $deleted )->text );
304 $kept_obj->alter_text( $new_text );
306 $self->del_reading( $deleted );
310 # Helper function for manipulating the graph.
311 sub _stringify_args {
312 my( $self, $first, $second, $arg ) = @_;
314 if ref( $first ) eq 'Text::Tradition::Collation::Reading';
315 $second = $second->id
316 if ref( $second ) eq 'Text::Tradition::Collation::Reading';
317 return( $first, $second, $arg );
320 # Helper function for manipulating the graph.
321 sub _objectify_args {
322 my( $self, $first, $second, $arg ) = @_;
323 $first = $self->reading( $first )
324 unless ref( $first ) eq 'Text::Tradition::Collation::Reading';
325 $second = $self->reading( $second )
326 unless ref( $second ) eq 'Text::Tradition::Collation::Reading';
327 return( $first, $second, $arg );
334 # We only need the IDs for adding paths to the graph, not the reading
335 # objects themselves.
336 my( $source, $target, $wit ) = $self->_stringify_args( @_ );
338 # Connect the readings
339 $self->sequence->add_edge( $source, $target );
340 # Note the witness in question
341 $self->sequence->set_edge_attribute( $source, $target, $wit, 1 );
347 if( ref( $_[0] ) eq 'ARRAY' ) {
354 # We only need the IDs for adding paths to the graph, not the reading
355 # objects themselves.
356 my( $source, $target, $wit ) = $self->_stringify_args( @args );
358 if( $self->sequence->has_edge_attribute( $source, $target, $wit ) ) {
359 $self->sequence->delete_edge_attribute( $source, $target, $wit );
361 unless( keys %{$self->sequence->get_edge_attributes( $source, $target )} ) {
362 $self->sequence->delete_edge( $source, $target );
367 # Extra graph-alike utility
370 my( $source, $target, $wit ) = $self->_stringify_args( @_ );
371 return undef unless $self->sequence->has_edge( $source, $target );
372 return $self->sequence->has_edge_attribute( $source, $target, $wit );
375 =head2 clear_witness( @sigil_list )
377 Clear the given witnesses out of the collation entirely, removing references
378 to them in paths, and removing readings that belong only to them. Should only
379 be called via $tradition->del_witness.
384 my( $self, @sigils ) = @_;
386 # Clear the witness(es) out of the paths
387 foreach my $e ( $self->paths ) {
388 foreach my $sig ( @sigils ) {
389 $self->del_path( $e, $sig );
393 # Clear out the newly unused readings
394 foreach my $r ( $self->readings ) {
395 unless( $self->reading_witnesses( $r ) ) {
396 $self->del_reading( $r );
401 sub add_relationship {
403 my( $source, $target, $opts ) = $self->_stringify_args( @_ );
404 my( @vectors ) = $self->relations->add_relationship( $source,
405 $self->reading( $source ), $target, $self->reading( $target ), $opts );
406 # Force a full rank recalculation every time. Yuck.
407 $self->calculate_ranks() if $self->end->has_rank;
411 =head2 reading_witnesses( $reading )
413 Return a list of sigils corresponding to the witnesses in which the reading appears.
417 sub reading_witnesses {
418 my( $self, $reading ) = @_;
419 # We need only check either the incoming or the outgoing edges; I have
420 # arbitrarily chosen "incoming". Thus, special-case the start node.
421 if( $reading eq $self->start ) {
422 return map { $_->sigil } $self->tradition->witnesses;
425 foreach my $e ( $self->sequence->edges_to( $reading ) ) {
426 my $wits = $self->sequence->get_edge_attributes( @$e );
427 @all_witnesses{ keys %$wits } = 1;
429 return keys %all_witnesses;
432 =head1 OUTPUT METHODS
436 Returns an SVG string that represents the graph, via as_dot and graphviz.
443 my @cmd = qw/dot -Tsvg/;
445 my $dotfile = File::Temp->new();
447 # $dotfile->unlink_on_destroy(0);
448 binmode $dotfile, ':utf8';
449 print $dotfile $self->as_dot();
450 push( @cmd, $dotfile->filename );
451 run( \@cmd, ">", binary(), \$svg );
452 $svg = decode_utf8( $svg );
456 =head2 svg_subgraph( $from, $to )
458 Returns an SVG string that represents the portion of the graph given by the
459 specified range. The $from and $to variables refer to ranks within the graph.
464 my( $self, $from, $to ) = @_;
466 my $dot = $self->as_dot( $from, $to );
468 throw( "Could not output a graph with range $from - $to" );
471 my @cmd = qw/dot -Tsvg/;
473 my $dotfile = File::Temp->new();
475 # $dotfile->unlink_on_destroy(0);
476 binmode $dotfile, ':utf8';
478 push( @cmd, $dotfile->filename );
479 run( \@cmd, ">", binary(), \$svg );
480 $svg = decode_utf8( $svg );
485 =head2 as_dot( $from, $to )
487 Returns a string that is the collation graph expressed in dot
488 (i.e. GraphViz) format. If $from or $to is passed, as_dot creates
489 a subgraph rather than the entire graph.
494 my( $self, $startrank, $endrank ) = @_;
496 # Check the arguments
498 return if $endrank && $startrank > $endrank;
499 return if $startrank > $self->end->rank;
501 if( defined $endrank ) {
502 return if $endrank < 0;
503 $endrank = undef if $endrank == $self->end->rank;
506 # TODO consider making some of these things configurable
507 my $graph_name = $self->tradition->name;
508 $graph_name =~ s/[^\w\s]//g;
509 $graph_name = join( '_', split( /\s+/, $graph_name ) );
517 'fillcolor' => 'white',
522 'arrowhead' => 'open',
523 'color' => '#000000',
524 'fontcolor' => '#000000',
527 my $dot = sprintf( "digraph %s {\n", $graph_name );
528 $dot .= "\tgraph " . _dot_attr_string( \%graph_attrs ) . ";\n";
529 $dot .= "\tnode " . _dot_attr_string( \%node_attrs ) . ";\n";
531 # Output substitute start/end readings if necessary
533 $dot .= "\t\"#SUBSTART#\" [ label=\"...\" ];\n";
536 $dot .= "\t\"#SUBEND#\" [ label=\"...\" ];\n";
538 my %used; # Keep track of the readings that actually appear in the graph
539 foreach my $reading ( $self->readings ) {
540 # Only output readings within our rank range.
541 next if $startrank && $reading->rank < $startrank;
542 next if $endrank && $reading->rank > $endrank;
543 $used{$reading->id} = 1;
544 # Need not output nodes without separate labels
545 next if $reading->id eq $reading->text;
546 my $label = $reading->text;
547 $label =~ s/\"/\\\"/g;
548 $dot .= sprintf( "\t\"%s\" [ label=\"%s\" ];\n", $reading->id, $label );
552 my @edges = $self->paths;
553 my( %substart, %subend );
554 foreach my $edge ( @edges ) {
555 # Do we need to output this edge?
556 if( $used{$edge->[0]} && $used{$edge->[1]} ) {;
557 my $label = $self->path_display_label( $self->path_witnesses( $edge ) );
558 my $variables = { %edge_attrs, 'label' => $label };
559 # Account for the rank gap if necessary
560 if( $self->reading( $edge->[1] )->has_rank
561 && $self->reading( $edge->[0] )->has_rank
562 && $self->reading( $edge->[1] )->rank
563 - $self->reading( $edge->[0] )->rank > 1 ) {
564 $variables->{'minlen'} = $self->reading( $edge->[1] )->rank
565 - $self->reading( $edge->[0] )->rank;
567 my $varopts = _dot_attr_string( $variables );
568 $dot .= sprintf( "\t\"%s\" -> \"%s\" %s;\n",
569 $edge->[0], $edge->[1], $varopts );
570 } elsif( $used{$edge->[0]} ) {
571 $subend{$edge->[0]} = 1;
572 } elsif( $used{$edge->[1]} ) {
573 $substart{$edge->[1]} = 1;
576 # Add substitute start and end edges if necessary
577 foreach my $node ( keys %substart ) {
578 my $witstr = $self->path_display_label ( $self->reading_witnesses( $self->reading( $node ) ) );
579 my $variables = { %edge_attrs, 'label' => $witstr };
580 my $varopts = _dot_attr_string( $variables );
581 $dot .= "\t\"#SUBSTART#\" -> \"$node\" $varopts;";
583 foreach my $node ( keys %subend ) {
584 my $witstr = $self->path_display_label ( $self->reading_witnesses( $self->reading( $node ) ) );
585 my $variables = { %edge_attrs, 'label' => $witstr };
586 my $varopts = _dot_attr_string( $variables );
587 $dot .= "\t\"$node\" -> \"#SUBEND#\" $varopts;";
594 sub _dot_attr_string {
597 foreach my $k ( sort keys %$hash ) {
599 push( @attrs, $k.'="'.$v.'"' );
601 return( '[ ' . join( ', ', @attrs ) . ' ]' );
605 my( $self, @edge ) = @_;
606 # If edge is an arrayref, cope.
607 if( @edge == 1 && ref( $edge[0] ) eq 'ARRAY' ) {
611 my @wits = keys %{$self->sequence->get_edge_attributes( @edge )};
615 sub path_display_label {
616 my( $self, @wits ) = @_;
617 my $maj = scalar( $self->tradition->witnesses ) * 0.6;
618 if( scalar @wits > $maj ) {
619 # TODO break out a.c. wits
622 return join( ', ', @wits );
629 Returns a GraphML representation of the collation. The GraphML will contain
630 two graphs. The first expresses the attributes of the readings and the witness
631 paths that link them; the second expresses the relationships that link the
632 readings. This is the native transfer format for a tradition.
641 my $datafile = 't/data/florilegium_tei_ps.xml';
642 my $tradition = Text::Tradition->new( 'input' => 'TEI',
647 ok( $tradition, "Got a tradition object" );
648 is( scalar $tradition->witnesses, 13, "Found all witnesses" );
649 ok( $tradition->collation, "Tradition has a collation" );
651 my $c = $tradition->collation;
652 is( scalar $c->readings, $READINGS, "Collation has all readings" );
653 is( scalar $c->paths, $PATHS, "Collation has all paths" );
654 is( scalar $c->relationships, 0, "Collation has all relationships" );
656 # Add a few relationships
657 $c->add_relationship( 'w123', 'w125', { 'type' => 'collated' } );
658 $c->add_relationship( 'w193', 'w196', { 'type' => 'collated' } );
659 $c->add_relationship( 'w257', 'w262', { 'type' => 'transposition' } );
661 # Now write it to GraphML and parse it again.
663 my $graphml = $c->as_graphml;
664 my $st = Text::Tradition->new( 'input' => 'Self', 'string' => $graphml );
665 is( scalar $st->collation->readings, $READINGS, "Reparsed collation has all readings" );
666 is( scalar $st->collation->paths, $PATHS, "Reparsed collation has all paths" );
667 is( scalar $st->collation->relationships, 3, "Reparsed collation has new relationships" );
677 my $graphml_ns = 'http://graphml.graphdrawing.org/xmlns';
678 my $xsi_ns = 'http://www.w3.org/2001/XMLSchema-instance';
679 my $graphml_schema = 'http://graphml.graphdrawing.org/xmlns ' .
680 'http://graphml.graphdrawing.org/xmlns/1.0/graphml.xsd';
682 # Create the document and root node
683 my $graphml = XML::LibXML->createDocument( "1.0", "UTF-8" );
684 my $root = $graphml->createElementNS( $graphml_ns, 'graphml' );
685 $graphml->setDocumentElement( $root );
686 $root->setNamespace( $xsi_ns, 'xsi', 0 );
687 $root->setAttributeNS( $xsi_ns, 'schemaLocation', $graphml_schema );
689 # Add the data keys for the graph
692 my @graph_attributes = qw/ version wit_list_separator baselabel linear ac_label /;
693 foreach my $datum ( @graph_attributes ) {
694 $graph_data_keys{$datum} = 'dg'.$gdi++;
695 my $key = $root->addNewChild( $graphml_ns, 'key' );
696 $key->setAttribute( 'attr.name', $datum );
697 $key->setAttribute( 'attr.type', $key eq 'linear' ? 'boolean' : 'string' );
698 $key->setAttribute( 'for', 'graph' );
699 $key->setAttribute( 'id', $graph_data_keys{$datum} );
702 # Add the data keys for nodes
709 is_start => 'boolean',
711 is_lacuna => 'boolean',
713 foreach my $datum ( keys %node_data ) {
714 $node_data_keys{$datum} = 'dn'.$ndi++;
715 my $key = $root->addNewChild( $graphml_ns, 'key' );
716 $key->setAttribute( 'attr.name', $datum );
717 $key->setAttribute( 'attr.type', $node_data{$datum} );
718 $key->setAttribute( 'for', 'node' );
719 $key->setAttribute( 'id', $node_data_keys{$datum} );
722 # Add the data keys for edges, i.e. witnesses
726 class => 'string', # Class, deprecated soon
727 witness => 'string', # ID/label for a path
728 relationship => 'string', # ID/label for a relationship
729 extra => 'boolean', # Path key
730 scope => 'string', # Relationship key
731 non_correctable => 'boolean', # Relationship key
732 non_independent => 'boolean', # Relationship key
734 foreach my $datum ( keys %edge_data ) {
735 $edge_data_keys{$datum} = 'de'.$edi++;
736 my $key = $root->addNewChild( $graphml_ns, 'key' );
737 $key->setAttribute( 'attr.name', $datum );
738 $key->setAttribute( 'attr.type', $edge_data{$datum} );
739 $key->setAttribute( 'for', 'edge' );
740 $key->setAttribute( 'id', $edge_data_keys{$datum} );
743 # Add the collation graph itself
744 my $sgraph = $root->addNewChild( $graphml_ns, 'graph' );
745 $sgraph->setAttribute( 'edgedefault', 'directed' );
746 $sgraph->setAttribute( 'id', $self->tradition->name );
747 $sgraph->setAttribute( 'parse.edgeids', 'canonical' );
748 $sgraph->setAttribute( 'parse.edges', scalar($self->paths) );
749 $sgraph->setAttribute( 'parse.nodeids', 'canonical' );
750 $sgraph->setAttribute( 'parse.nodes', scalar($self->readings) );
751 $sgraph->setAttribute( 'parse.order', 'nodesfirst' );
753 # Collation attribute data
754 foreach my $datum ( @graph_attributes ) {
755 my $value = $datum eq 'version' ? '3.0' : $self->$datum;
756 _add_graphml_data( $sgraph, $graph_data_keys{$datum}, $value );
761 # Add our readings to the graph
762 foreach my $n ( sort { $a->id cmp $b->id } $self->readings ) {
763 # Add to the main graph
764 my $node_el = $sgraph->addNewChild( $graphml_ns, 'node' );
765 my $node_xmlid = 'n' . $node_ctr++;
766 $node_hash{ $n->id } = $node_xmlid;
767 $node_el->setAttribute( 'id', $node_xmlid );
768 foreach my $d ( keys %node_data ) {
770 _add_graphml_data( $node_el, $node_data_keys{$d}, $nval )
775 # Add the path edges to the sequence graph
777 foreach my $e ( sort { $a->[0] cmp $b->[0] } $self->sequence->edges() ) {
778 # We add an edge in the graphml for every witness in $e.
779 foreach my $wit ( $self->path_witnesses( $e ) ) {
780 my( $id, $from, $to ) = ( 'e'.$edge_ctr++,
781 $node_hash{ $e->[0] },
782 $node_hash{ $e->[1] } );
783 my $edge_el = $sgraph->addNewChild( $graphml_ns, 'edge' );
784 $edge_el->setAttribute( 'source', $from );
785 $edge_el->setAttribute( 'target', $to );
786 $edge_el->setAttribute( 'id', $id );
788 # It's a witness path, so add the witness
790 my $key = $edge_data_keys{'witness'};
791 # Is this an ante-corr witness?
792 my $aclabel = $self->ac_label;
793 if( $wit =~ /^(.*)\Q$aclabel\E$/ ) {
794 # Keep the base witness
796 # ...and record that this is an 'extra' reading path
797 _add_graphml_data( $edge_el, $edge_data_keys{'extra'}, $aclabel );
799 _add_graphml_data( $edge_el, $edge_data_keys{'witness'}, $base );
800 _add_graphml_data( $edge_el, $edge_data_keys{'class'}, 'path' );
804 # Add the relationship graph to the XML
805 $self->relations->as_graphml( $graphml_ns, $root, \%node_hash,
806 $node_data_keys{'id'}, \%edge_data_keys );
808 # Save and return the thing
809 my $result = decode_utf8( $graphml->toString(1) );
813 sub _add_graphml_data {
814 my( $el, $key, $value ) = @_;
815 return unless defined $value;
816 my $data_el = $el->addNewChild( $el->namespaceURI, 'data' );
817 $data_el->setAttribute( 'key', $key );
818 $data_el->appendText( $value );
823 Returns a CSV alignment table representation of the collation graph, one
824 row per witness (or witness uncorrected.)
830 my $table = $self->make_alignment_table;
831 my $csv = Text::CSV_XS->new( { binary => 1, quote_null => 0 } );
833 # Make the header row
834 $csv->combine( map { $_->{'witness'} } @{$table->{'alignment'}} );
835 push( @result, decode_utf8( $csv->string ) );
836 # Make the rest of the rows
837 foreach my $idx ( 0 .. $table->{'length'} - 1 ) {
838 my @rowobjs = map { $_->{'tokens'}->[$idx] } @{$table->{'alignment'}};
839 my @row = map { $_ ? $_->{'t'} : $_ } @rowobjs;
840 $csv->combine( @row );
841 push( @result, decode_utf8( $csv->string ) );
843 return join( "\n", @result );
846 =head2 make_alignment_table( $use_refs, $include_witnesses )
848 Return a reference to an alignment table, in a slightly enhanced CollateX
849 format which looks like this:
851 $table = { alignment => [ { witness => "SIGIL",
852 tokens => [ { t => "TEXT" }, ... ] },
854 tokens => [ { t => "TEXT" }, ... ] },
858 If $use_refs is set to 1, the reading object is returned in the table
859 instead of READINGTEXT; if not, the text of the reading is returned.
861 If $include_witnesses is set to a hashref, only the witnesses whose sigil
862 keys have a true hash value will be included.
866 sub make_alignment_table {
867 my( $self, $noderefs, $include ) = @_;
868 unless( $self->linear ) {
869 throw( "Need a linear graph in order to make an alignment table" );
871 my $table = { 'alignment' => [], 'length' => $self->end->rank - 1 };
872 my @all_pos = ( 1 .. $self->end->rank - 1 );
873 foreach my $wit ( sort { $a->sigil cmp $b->sigil } $self->tradition->witnesses ) {
875 next unless $include->{$wit->sigil};
877 # print STDERR "Making witness row(s) for " . $wit->sigil . "\n";
878 my @wit_path = $self->reading_sequence( $self->start, $self->end, $wit->sigil );
879 my @row = _make_witness_row( \@wit_path, \@all_pos, $noderefs );
880 push( @{$table->{'alignment'}},
881 { 'witness' => $wit->sigil, 'tokens' => \@row } );
882 if( $wit->is_layered ) {
883 my @wit_ac_path = $self->reading_sequence( $self->start, $self->end,
884 $wit->sigil.$self->ac_label );
885 my @ac_row = _make_witness_row( \@wit_ac_path, \@all_pos, $noderefs );
886 push( @{$table->{'alignment'}},
887 { 'witness' => $wit->sigil.$self->ac_label, 'tokens' => \@ac_row } );
893 sub _make_witness_row {
894 my( $path, $positions, $noderefs ) = @_;
896 map { $char_hash{$_} = undef } @$positions;
898 foreach my $rdg ( @$path ) {
899 my $rtext = $rdg->text;
900 $rtext = '#LACUNA#' if $rdg->is_lacuna;
901 print STDERR "rank " . $rdg->rank . "\n" if $debug;
902 # print STDERR "No rank for " . $rdg->id . "\n" unless defined $rdg->rank;
903 $char_hash{$rdg->rank} = $noderefs ? { 't' => $rdg }
906 my @row = map { $char_hash{$_} } @$positions;
907 # Fill in lacuna markers for undef spots in the row
908 my $last_el = shift @row;
909 my @filled_row = ( $last_el );
910 foreach my $el ( @row ) {
911 # If we are using node reference, make the lacuna node appear many times
912 # in the table. If not, use the lacuna tag.
913 if( $last_el && _el_is_lacuna( $last_el ) && !defined $el ) {
914 $el = $noderefs ? $last_el : { 't' => '#LACUNA#' };
916 push( @filled_row, $el );
922 # Tiny utility function to say if a table element is a lacuna
925 return 1 if $el->{'t'} eq '#LACUNA#';
926 return 1 if ref( $el->{'t'} ) eq 'Text::Tradition::Collation::Reading'
927 && $el->{'t'}->is_lacuna;
931 # Helper to turn the witnesses along columns rather than rows. Assumes
936 return $result unless scalar @$table;
937 my $nrows = scalar @{$table->[0]};
938 foreach my $idx ( 0 .. $nrows - 1 ) {
939 foreach my $wit ( 0 .. $#{$table} ) {
940 $result->[$idx]->[$wit] = $table->[$wit]->[$idx];
946 =head1 NAVIGATION METHODS
948 =head2 reading_sequence( $first, $last, $sigil, $backup )
950 Returns the ordered list of readings, starting with $first and ending
951 with $last, for the witness given in $sigil. If a $backup sigil is
952 specified (e.g. when walking a layered witness), it will be used wherever
953 no $sigil path exists. If there is a base text reading, that will be
954 used wherever no path exists for $sigil or $backup.
958 # TODO Think about returning some lazy-eval iterator.
959 # TODO Get rid of backup; we should know from what witness is whether we need it.
961 sub reading_sequence {
962 my( $self, $start, $end, $witness ) = @_;
964 $witness = $self->baselabel unless $witness;
965 my @readings = ( $start );
968 while( $n && $n->id ne $end->id ) {
969 if( exists( $seen{$n->id} ) ) {
970 throw( "Detected loop for $witness at " . $n->id );
974 my $next = $self->next_reading( $n, $witness );
976 throw( "Did not find any path for $witness from reading " . $n->id );
978 push( @readings, $next );
981 # Check that the last reading is our end reading.
982 my $last = $readings[$#readings];
983 throw( "Last reading found from " . $start->text .
984 " for witness $witness is not the end!" ) # TODO do we get this far?
985 unless $last->id eq $end->id;
990 =head2 next_reading( $reading, $sigil );
992 Returns the reading that follows the given reading along the given witness
998 # Return the successor via the corresponding path.
1000 my $answer = $self->_find_linked_reading( 'next', @_ );
1001 return undef unless $answer;
1002 return $self->reading( $answer );
1005 =head2 prior_reading( $reading, $sigil )
1007 Returns the reading that precedes the given reading along the given witness
1013 # Return the predecessor via the corresponding path.
1015 my $answer = $self->_find_linked_reading( 'prior', @_ );
1016 return $self->reading( $answer );
1019 sub _find_linked_reading {
1020 my( $self, $direction, $node, $path ) = @_;
1022 # Get a backup if we are dealing with a layered witness
1024 my $aclabel = $self->ac_label;
1025 if( $path && $path =~ /^(.*)\Q$aclabel\E$/ ) {
1029 my @linked_paths = $direction eq 'next'
1030 ? $self->sequence->edges_from( $node )
1031 : $self->sequence->edges_to( $node );
1032 return undef unless scalar( @linked_paths );
1034 # We have to find the linked path that contains all of the
1035 # witnesses supplied in $path.
1036 my( @path_wits, @alt_path_wits );
1037 @path_wits = sort( $self->_witnesses_of_label( $path ) ) if $path;
1038 @alt_path_wits = sort( $self->_witnesses_of_label( $alt_path ) ) if $alt_path;
1041 foreach my $le ( @linked_paths ) {
1042 if( $self->sequence->has_edge_attribute( @$le, $self->baselabel ) ) {
1045 my @le_wits = $self->path_witnesses( $le );
1046 if( _is_within( \@path_wits, \@le_wits ) ) {
1047 # This is the right path.
1048 return $direction eq 'next' ? $le->[1] : $le->[0];
1049 } elsif( _is_within( \@alt_path_wits, \@le_wits ) ) {
1053 # Got this far? Return the alternate path if it exists.
1054 return $direction eq 'next' ? $alt_le->[1] : $alt_le->[0]
1057 # Got this far? Return the base path if it exists.
1058 return $direction eq 'next' ? $base_le->[1] : $base_le->[0]
1061 # Got this far? We have no appropriate path.
1062 warn "Could not find $direction node from " . $node->id
1063 . " along path $path";
1069 my( $set1, $set2 ) = @_;
1070 my $ret = @$set1; # will be 0, i.e. false, if set1 is empty
1071 foreach my $el ( @$set1 ) {
1072 $ret = 0 unless grep { /^\Q$el\E$/ } @$set2;
1077 # Return the string that joins together a list of witnesses for
1078 # display on a single path.
1079 sub _witnesses_of_label {
1080 my( $self, $label ) = @_;
1081 my $regex = $self->wit_list_separator;
1082 my @answer = split( /\Q$regex\E/, $label );
1086 =head2 path_text( $sigil, $mainsigil [, $start, $end ] )
1088 Returns the text of a witness (plus its backup, if we are using a layer)
1089 as stored in the collation. The text is returned as a string, where the
1090 individual readings are joined with spaces and the meta-readings (e.g.
1091 lacunae) are omitted. Optional specification of $start and $end allows
1092 the generation of a subset of the witness text.
1097 my( $self, $wit, $start, $end ) = @_;
1098 $start = $self->start unless $start;
1099 $end = $self->end unless $end;
1100 my @path = grep { !$_->is_meta } $self->reading_sequence( $start, $end, $wit );
1101 return join( ' ', map { $_->text } @path );
1104 =head1 INITIALIZATION METHODS
1106 These are mostly for use by parsers.
1108 =head2 make_witness_path( $witness )
1110 Link the array of readings contained in $witness->path (and in
1111 $witness->uncorrected_path if it exists) into collation paths.
1112 Clear out the arrays when finished.
1114 =head2 make_witness_paths
1116 Call make_witness_path for all witnesses in the tradition.
1120 # For use when a collation is constructed from a base text and an apparatus.
1121 # We have the sequences of readings and just need to add path edges.
1122 # When we are done, clear out the witness path attributes, as they are no
1124 # TODO Find a way to replace the witness path attributes with encapsulated functions?
1126 sub make_witness_paths {
1128 foreach my $wit ( $self->tradition->witnesses ) {
1129 # print STDERR "Making path for " . $wit->sigil . "\n";
1130 $self->make_witness_path( $wit );
1134 sub make_witness_path {
1135 my( $self, $wit ) = @_;
1136 my @chain = @{$wit->path};
1137 my $sig = $wit->sigil;
1138 foreach my $idx ( 0 .. $#chain-1 ) {
1139 $self->add_path( $chain[$idx], $chain[$idx+1], $sig );
1141 if( $wit->is_layered ) {
1142 @chain = @{$wit->uncorrected_path};
1143 foreach my $idx( 0 .. $#chain-1 ) {
1144 my $source = $chain[$idx];
1145 my $target = $chain[$idx+1];
1146 $self->add_path( $source, $target, $sig.$self->ac_label )
1147 unless $self->has_path( $source, $target, $sig );
1151 $wit->clear_uncorrected_path;
1154 =head2 calculate_ranks
1156 Calculate the reading ranks (that is, their aligned positions relative
1157 to each other) for the graph. This can only be called on linear collations.
1161 sub calculate_ranks {
1163 # Walk a version of the graph where every node linked by a relationship
1164 # edge is fundamentally the same node, and do a topological ranking on
1165 # the nodes in this graph.
1166 my $topo_graph = Graph->new();
1170 foreach my $r ( $self->readings ) {
1171 next if exists $rel_containers{$r->id};
1172 my @rels = $r->related_readings( 'colocated' );
1174 # Make a relationship container.
1176 my $rn = 'rel_container_' . $rel_ctr++;
1177 $topo_graph->add_vertex( $rn );
1179 $rel_containers{$_->id} = $rn;
1182 # Add a new node to mirror the old node.
1183 $rel_containers{$r->id} = $r->id;
1184 $topo_graph->add_vertex( $r->id );
1189 foreach my $r ( $self->readings ) {
1190 foreach my $n ( $self->sequence->successors( $r->id ) ) {
1191 my( $tfrom, $tto ) = ( $rel_containers{$r->id},
1192 $rel_containers{$n} );
1193 # $DB::single = 1 unless $tfrom && $tto;
1194 $topo_graph->add_edge( $tfrom, $tto );
1198 # Now do the rankings, starting with the start node.
1199 my $topo_start = $rel_containers{$self->start->id};
1200 my $node_ranks = { $topo_start => 0 };
1201 my @curr_origin = ( $topo_start );
1202 # A little iterative function.
1203 while( @curr_origin ) {
1204 @curr_origin = _assign_rank( $topo_graph, $node_ranks, @curr_origin );
1206 # Transfer our rankings from the topological graph to the real one.
1207 foreach my $r ( $self->readings ) {
1208 if( defined $node_ranks->{$rel_containers{$r->id}} ) {
1209 $r->rank( $node_ranks->{$rel_containers{$r->id}} );
1211 # Die. Find the last rank we calculated.
1212 my @all_defined = sort { $node_ranks->{$rel_containers{$a->id}}
1213 <=> $node_ranks->{$rel_containers{$b->id}} }
1215 my $last = pop @all_defined;
1216 throw( "Ranks not calculated after $last - do you have a cycle in the graph?" );
1222 my( $graph, $node_ranks, @current_nodes ) = @_;
1223 # Look at each of the children of @current_nodes. If all the child's
1224 # parents have a rank, assign it the highest rank + 1 and add it to
1225 # @next_nodes. Otherwise skip it; we will return when the highest-ranked
1226 # parent gets a rank.
1228 foreach my $c ( @current_nodes ) {
1229 warn "Current reading $c has no rank!"
1230 unless exists $node_ranks->{$c};
1231 # print STDERR "Looking at child of node $c, rank "
1232 # . $node_ranks->{$c} . "\n";
1233 foreach my $child ( $graph->successors( $c ) ) {
1234 next if exists $node_ranks->{$child};
1235 my $highest_rank = -1;
1237 foreach my $parent ( $graph->predecessors( $child ) ) {
1238 if( exists $node_ranks->{$parent} ) {
1239 $highest_rank = $node_ranks->{$parent}
1240 if $highest_rank <= $node_ranks->{$parent};
1247 my $c_rank = $highest_rank + 1;
1248 # print STDERR "Assigning rank $c_rank to node $child \n";
1249 $node_ranks->{$child} = $c_rank;
1250 push( @next_nodes, $child );
1256 =head2 flatten_ranks
1258 A convenience method for parsing collation data. Searches the graph for readings
1259 with the same text at the same rank, and merges any that are found.
1265 my %unique_rank_rdg;
1266 foreach my $rdg ( $self->readings ) {
1267 next unless $rdg->has_rank;
1268 my $key = $rdg->rank . "||" . $rdg->text;
1269 if( exists $unique_rank_rdg{$key} ) {
1271 # print STDERR "Combining readings at same rank: $key\n";
1272 $self->merge_readings( $unique_rank_rdg{$key}, $rdg );
1274 $unique_rank_rdg{$key} = $rdg;
1279 =head2 text_from_paths
1281 Calculate the text array for all witnesses from the path, for later consistency
1282 checking. Only to be used if there is no non-graph-based way to know the
1287 sub text_from_paths {
1289 foreach my $wit ( $self->tradition->witnesses ) {
1290 my @text = split( /\s+/,
1291 $self->reading_sequence( $self->start, $self->end, $wit->sigil ) );
1292 $wit->text( \@text );
1293 if( $wit->is_layered ) {
1294 my @uctext = split( /\s+/,
1295 $self->reading_sequence( $self->start, $self->end,
1296 $wit->sigil.$self->ac_label ) );
1297 $wit->text( \@uctext );
1302 =head1 UTILITY FUNCTIONS
1304 =head2 common_predecessor( $reading_a, $reading_b )
1306 Find the last reading that occurs in sequence before both the given readings.
1308 =head2 common_successor( $reading_a, $reading_b )
1310 Find the first reading that occurs in sequence after both the given readings.
1314 use Text::Tradition;
1316 my $cxfile = 't/data/Collatex-16.xml';
1317 my $t = Text::Tradition->new(
1319 'input' => 'CollateX',
1322 my $c = $t->collation;
1324 is( $c->common_predecessor( 'n9', 'n23' )->id,
1325 'n20', "Found correct common predecessor" );
1326 is( $c->common_successor( 'n9', 'n23' )->id,
1327 '#END#', "Found correct common successor" );
1329 is( $c->common_predecessor( 'n19', 'n17' )->id,
1330 'n16', "Found correct common predecessor for readings on same path" );
1331 is( $c->common_successor( 'n21', 'n26' )->id,
1332 '#END#', "Found correct common successor for readings on same path" );
1338 ## Return the closest reading that is a predecessor of both the given readings.
1339 sub common_predecessor {
1341 my( $r1, $r2 ) = $self->_objectify_args( @_ );
1342 return $self->common_in_path( $r1, $r2, 'predecessors' );
1345 sub common_successor {
1347 my( $r1, $r2 ) = $self->_objectify_args( @_ );
1348 return $self->common_in_path( $r1, $r2, 'successors' );
1351 sub common_in_path {
1352 my( $self, $r1, $r2, $dir ) = @_;
1353 my $iter = $r1->rank > $r2->rank ? $r1->rank : $r2->rank;
1354 $iter = $self->end->rank - $iter if $dir eq 'successors';
1356 my @last_checked = ( $r1, $r2 );
1358 while( !@candidates ) {
1360 foreach my $lc ( @last_checked ) {
1361 foreach my $p ( $lc->$dir ) {
1362 if( $all_seen{$p->id} ) {
1363 push( @candidates, $p );
1365 $all_seen{$p->id} = 1;
1366 push( @new_lc, $p );
1370 @last_checked = @new_lc;
1372 my @answer = sort { $a->rank <=> $b->rank } @candidates;
1373 return $dir eq 'predecessors' ? pop( @answer ) : shift ( @answer );
1377 Text::Tradition::Error->throw(
1378 'ident' => 'Collation error',
1384 __PACKAGE__->meta->make_immutable;
1390 =item * Get rid of $backup in reading_sequence