X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=lib%2FText%2FTradition%2FCollation%2FReading.pm;h=29a8aa17891441840d94a407a58ce617246f376a;hb=c3e04fb5fc2dfb0a8451c2a12538440548ccedbd;hp=a57e68f115e219b64872670a0d334ef448a6fd20;hpb=4aea6e9bf2a38188f3e0aaef7c1044e4d2313c82;p=scpubgit%2Fstemmatology.git

diff --git a/lib/Text/Tradition/Collation/Reading.pm b/lib/Text/Tradition/Collation/Reading.pm
index a57e68f..29a8aa1 100644
--- a/lib/Text/Tradition/Collation/Reading.pm
+++ b/lib/Text/Tradition/Collation/Reading.pm
@@ -1,6 +1,8 @@
 package Text::Tradition::Collation::Reading;
 
 use Moose;
+use Module::Load;
+use YAML::XS;
 use overload '""' => \&_stringify, 'fallback' => 1;
 
 =head1 NAME
@@ -89,7 +91,7 @@ has 'text' => (
 has 'language' => (
 	is => 'ro',
 	isa => 'Str',
-	default => 'Default',
+	predicate => 'has_language',
 	);
 	
 has 'is_start' => (
@@ -137,21 +139,17 @@ has 'normal_form' => (
 	predicate => 'has_normal_form',
 	);
 
-has 'lemma' => (
-	is => 'rw',
-	isa => 'Str',
-	predicate => 'has_lemma',
-	);
-
-has 'morphology' => (
+# Holds the lexemes for the reading.
+has 'reading_lexemes' => (
 	traits => ['Array'],
-	isa => 'ArrayRef[HashRef[ArrayRef[Text::Tradition::Collation::Reading::Morphology]]]',
+	isa => 'ArrayRef[Text::Tradition::Collation::Reading::Lexeme]',
 	handles => {
 		lexemes => 'elements',
-		has_morphology => 'count',
-		_clear_morph => 'clear',
-		_add_morph => 'push',
+		has_lexemes => 'count',
+		clear_lexemes => 'clear',
+		add_lexeme => 'push',
 		},
+	default => sub { [] },
 	);
 	
 ## For prefix/suffix readings
@@ -278,90 +276,66 @@ sub _stringify {
 
 =head1 MORPHOLOGY
 
-A few methods to try to tack on morphological information.
+Methods for the morphological information (if any) attached to readings.
+A reading may be made up of multiple lexemes; the concatenated lexeme
+strings ought to match the reading's normalized form.
+ 
+See L<Text::Tradition::Collation::Reading::Lexeme> for more information
+on Lexeme objects and their attributes.
 
-=head2 is_disambiguated
+=head2 has_lexemes
 
-Returns true if there is only one tag per lexeme in this reading.
+Returns a true value if the reading has any attached lexemes.
 
-=head2 use_lexemes
+=head2 lexemes
 
-TBD
+Returns the Lexeme objects (if any) attached to the reading.
 
-=head2 add_morphological_tag
+=head2 clear_lexemes
 
-TBD
+Wipes any associated Lexeme objects out of the reading.
 
-=head2 disambiguate
+=head2 add_lexeme( $lexobj )
 
-TBD
+Adds the Lexeme in $lexobj to the list of lexemes.
 
-=cut
+=head2 lemmatize
 
-sub use_lexemes {
-	my( $self, @lexemes ) = @_;
-	# The lexemes need to be the same as $self->text.
-	my $cmpstr = $self->has_normal_form ? lc( $self->normal_form ) : lc( $self->text );
-	$cmpstr =~ s/[\s-]+//g;
-	my $lexstr = lc( join( '', @lexemes ) );
-	$lexstr =~ s/[\s-]+//g;
-	unless( $lexstr eq $cmpstr ) {
-		warn "Cannot split " . $self->text . " into " . join( '.', @lexemes );
-		return;
-	}
-	$self->_clear_morph;
-	map { $self->_add_morph( { $_ => [] } ) } @lexemes;
-}
+If the language of the reading is set, this method will use the appropriate
+Language model to determine the lexemes that belong to this reading.  See
+L<Text::Tradition::lemmatize> if you wish to lemmatize an entire tradition.
 
-sub add_morphological_tag {
-	my( $self, $lexeme, $opts ) = @_;
-	my $struct;
-	unless( $opts ) {
-		# No lexeme was passed; use reading text.
-		$opts = $lexeme;
-		$lexeme = $self->text;
-		$self->use_lexemes( $lexeme );
-	}
-	# Get the correct container
-	( $struct ) = grep { exists $_->{$lexeme} } $self->lexemes;
-	unless( $struct ) {
-		warn "No lexeme $lexeme exists in this reading";
-		return;
-	}
-	# Now make the morph object and add it to this lexeme.
-	my $morph_obj = Text::Tradition::Collation::Reading::Morphology->new( $opts );
-	# TODO Check for existence
-	push( @{$struct->{$lexeme}}, $morph_obj );
-}
+=cut
 
-sub disambiguate {
-	my( $self, $lexeme, $index ) = @_;
-	my $struct;
-	unless( $index ) {
-		# No lexeme was passed; use reading text.
-		$index = $lexeme;
-		$lexeme = $self->text;
-	}
-	# Get the correct container
-	( $struct ) = grep { exists $_->{$lexeme} } $self->lexemes;
-	unless( $struct ) {
-		warn "No lexeme $lexeme exists in this reading";
+sub lemmatize {
+	my $self = shift;
+	unless( $self->has_language ) {
+		warn "Please set a language to lemmatize a tradition";
 		return;
 	}
-	# Keep the object at the selected index
-	my $selected = $struct->{$lexeme}->[$index];
-	$struct->{$lexeme} = [ $selected ];
+	my $mod = "Text::Tradition::Language::" . $self->language;
+	load( $mod );
+	$mod->can( 'reading_lookup' )->( $self );
+
 }
 
-sub is_disambiguated {
+# For graph serialization. Return a string representation of the associated
+# reading lexemes.
+sub _serialize_lexemes {
 	my $self = shift;
-	return undef unless $self->has_morphology;
-	foreach my $lexeme ( $self->lexemes ) {
-		my( $key ) = keys %$lexeme; # will be only one
-		return undef unless @{$lexeme->{$key}} == 1;
+	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 1;
+	return join( '|R|', @lexstrs );
 }
+		
 
 ## Utility methods
 
@@ -375,87 +349,4 @@ sub TO_JSON {
 no Moose;
 __PACKAGE__->meta->make_immutable;
 
-###################################################
-### Morphology objects, to be attached to readings
-###################################################
-
-package Text::Tradition::Collation::Reading::Morphology;
-
-use Moose;
-
-has 'lemma' => (
-	is => 'ro',
-	isa => 'Str',
-	required => 1,
-	);
-	
-has 'code' => (
-	is => 'ro',
-	isa => 'Str',
-	required => 1,
-	);
-	
-has 'language' => (
-	is => 'ro',
-	isa => 'Str',
-	required => 1,
-	);
-	
-## Transmute codes into comparison arrays for our various languages.
-
-around BUILDARGS => sub {
-	my $orig = shift;
-	my $class = shift;
-	my $args;
-	if( @_ == 1 && ref( $_[0] ) ) {
-		$args = shift;
-	} else {
-		$args = { @_ };
-	}
-	if( exists( $args->{'serial'} ) ) {
-		my( $lemma, $code ) = split( /!!/, delete $args->{'serial'} );
-		$args->{'lemma'} = $lemma;
-		$args->{'code'} = $code;
-	}
-	$class->$orig( $args );
-};
-
-sub serialization {
-	my $self = shift;
-	return join( '!!', $self->lemma, $self->code );
-};
-
-sub comparison_array {
-	my $self = shift;
-	if( $self->language eq 'French' ) {
-		my @array;
-		my @bits = split( /\+/, $self->code );
-		# First push the non k/v parts.
-		while( @bits && $bits[0] !~ /=/ ) {
-			push( @array, shift @bits );
-		}
-		while( @array < 2 ) {
-			push( @array, undef );
-		}
-		# Now push the k/v parts in a known order.
-		my @fields = qw/ Pers Nb Temps Genre Spec Fonc /;
-		my %props;
-		map { my( $k, $v ) = split( /=/, $_ ); $props{$k} = $v; } @bits;
-		foreach my $k ( @fields ) {
-			push( @array, $props{$k} );
-		}
-		# Give the answer.
-		return @array;
-	} elsif( $self->language eq 'English' ) {
-		# Do something as yet undetermined
-	} else {
-		# Latin or Greek or Armenian, just split the chars
-		return split( '', $self->code );
-	}
-};
-
-no Moose;
-__PACKAGE__->meta->make_immutable;
-
 1;
-