start using witness->text and ->layertext for consistency checking

[scpubgit/stemmatology.git] / lib / Text / Tradition / Collation.pm

diff --git a/lib/Text/Tradition/Collation.pm b/lib/Text/Tradition/Collation.pm
index f713311..10d215e 100644 (file)
--- a/lib/Text/Tradition/Collation.pm
+++ b/lib/Text/Tradition/Collation.pm
@@ -66,12 +66,6 @@ has 'linear' => (
     default => 1,
     );
     
-has 'collapse_punctuation' => (
-       is => 'rw',
-       isa => 'Bool',
-       default => 1,
-       );
-
 has 'ac_label' => (
     is => 'rw',
     isa => 'Str',
@@ -133,8 +127,6 @@ belongs. Required.
 transposed readings should be treated as two linked readings rather than one, 
 and therefore whether the collation graph is acyclic.  Defaults to true.
 
-=item * collapse_punctuation - TODO
-
 =item * baselabel - The default label for the path taken by a base text 
 (if any). Defaults to 'base text'.
 
@@ -154,8 +146,6 @@ the like.  Defaults to ' (a.c.)'.
 
 =head2 linear
 
-=head2 collapse_punctuation
-
 =head2 wit_list_separator
 
 =head2 baselabel
@@ -554,7 +544,7 @@ sub as_dot {
         $used{$reading->id} = 1;
         # Need not output nodes without separate labels
         next if $reading->id eq $reading->text;
-        my $label = $reading->punctuated_form;
+        my $label = $reading->text;
         $label =~ s/\"/\\\"/g;
         $dot .= sprintf( "\t\"%s\" [ label=\"%s\" ];\n", $reading->id, $label );
     }
@@ -568,9 +558,13 @@ sub as_dot {
                my $label = $self->path_display_label( $self->path_witnesses( $edge ) );
                        my $variables = { %edge_attrs, 'label' => $label };
                        # Account for the rank gap if necessary
-                       my $rankgap = $self->reading( $edge->[1] )->rank 
+                       if( $self->reading( $edge->[1] )->has_rank 
+                               && $self->reading( $edge->[0] )->has_rank
+                               && $self->reading( $edge->[1] )->rank 
+                               - $self->reading( $edge->[0] )->rank > 1 ) {
+                               $variables->{'minlen'} = $self->reading( $edge->[1] )->rank 
                                - $self->reading( $edge->[0] )->rank;
-                       $variables->{'minlen'} = $rankgap if $rankgap > 1;
+                       }
                        my $varopts = _dot_attr_string( $variables );
                        $dot .= sprintf( "\t\"%s\" -> \"%s\" %s;\n", 
                                $edge->[0], $edge->[1], $varopts );
@@ -774,7 +768,6 @@ sub as_graphml {
         $node_el->setAttribute( 'id', $node_xmlid );
         foreach my $d ( keys %node_data ) {
                my $nval = $n->$d;
-               $nval = $n->punctuated_form if $d eq 'text';
                _add_graphml_data( $node_el, $node_data_keys{$d}, $nval )
                        if defined $nval;
         }
@@ -965,6 +958,7 @@ used wherever no path exists for $sigil or $backup.
 =cut
 
 # TODO Think about returning some lazy-eval iterator.
+# TODO Get rid of backup; we should know from what witness is whether we need it.
 
 sub reading_sequence {
     my( $self, $start, $end, $witness, $backup ) = @_;
@@ -1083,8 +1077,25 @@ sub _witnesses_of_label {
     my $regex = $self->wit_list_separator;
     my @answer = split( /\Q$regex\E/, $label );
     return @answer;
-}    
+}
 
+=head2 path_text( $sigil, $mainsigil [, $start, $end ] )
+
+Returns the text of a witness (plus its backup, if we are using a layer)
+as stored in the collation.  The text is returned as a string, where the
+individual readings are joined with spaces and the meta-readings (e.g.
+lacunae) are omitted.  Optional specification of $start and $end allows
+the generation of a subset of the witness text.
+
+=cut
+
+sub path_text {
+       my( $self, $wit, $backup, $start, $end ) = @_;
+       $start = $self->start unless $start;
+       $end = $self->end unless $end;
+       my @path = grep { !$_->is_meta } $self->reading_sequence( $start, $end, $wit, $backup );
+       return join( ' ', map { $_->text } @path );
+}
 
 =head1 INITIALIZATION METHODS