better lexeme serialization

diff --git a/lib/Text/Tradition/Collation/Reading.pm b/lib/Text/Tradition/Collation/Reading.pm
index 0e78cbc..29a8aa1 100644 (file)
--- a/lib/Text/Tradition/Collation/Reading.pm
+++ b/lib/Text/Tradition/Collation/Reading.pm
@@ -323,7 +323,17 @@ sub lemmatize {
 # reading lexemes.
 sub _serialize_lexemes {
        my $self = shift;
-       return Dump( [ $self->lexemes ] );
+       my @lexstrs;
+       foreach my $l ( $self->lexemes ) {
+               my @mf;
+               foreach my $wf ( $l->matching_forms ) {
+                       push( @mf, $wf->to_string );
+               }
+               my $form = $l->form ? $l->form->to_string : '';
+               push( @lexstrs, join( '|L|', $l->language, $l->string, $form, 
+                       join( '|M|', @mf ) ) );
+       }
+       return join( '|R|', @lexstrs );
 }