X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=lib%2FText%2FTradition%2FCollation%2FReading.pm;h=20b0990721750a265c0e6669b512c4f0bf7776ac;hb=a2ed66b255a1b88374b5279faa18f0c390ca5dbb;hp=146eadf6f82f5551d942bc2e66449c6f9e3bff70;hpb=8e1394aabe8d8f8c33b77de57a06e83fd75ce620;p=scpubgit%2Fstemmatology.git

diff --git a/lib/Text/Tradition/Collation/Reading.pm b/lib/Text/Tradition/Collation/Reading.pm
index 146eadf..20b0990 100644
--- a/lib/Text/Tradition/Collation/Reading.pm
+++ b/lib/Text/Tradition/Collation/Reading.pm
@@ -1,170 +1,439 @@
 package Text::Tradition::Collation::Reading;
 
 use Moose;
-use Moose::Util::TypeConstraints;
-use MooseX::NonMoose;
+use overload '""' => \&_stringify, 'fallback' => 1;
 
-extends 'Graph::Easy::Node';
+=head1 NAME
 
-subtype 'Position'
-    => as 'Str',
-    => where { $_ =~ /^\d+\,\d+$/ },
-    message { 'Position must be of the form x,y' };
+Text::Tradition::Collation::Reading - represents a reading (usually a word) in a collation.
+    
+=head1 DESCRIPTION
 
-has 'position' => (
-    is => 'rw',
-    isa => 'Position',
+Text::Tradition is a library for representation and analysis of collated
+texts, particularly medieval ones.  A 'reading' refers to a unit of text,
+usually a word, that appears in one or more witnesses (manuscripts) of the
+tradition; the text of a given witness is composed of a set of readings in
+a particular sequence
+
+=head1 METHODS
+
+=head2 new
+
+Creates a new reading in the given collation with the given attributes. 
+Options include:
+
+=over 4
+
+=item collation - The Text::Tradition::Collation object to which this reading belongs.  Required.
+
+=item id - A unique identifier for this reading. Required.
+
+=item text - The word or other text of the reading.
+
+=item is_start - The reading is the starting point for the collation.
+
+=item is_end - The reading is the ending point for the collation.
+
+=item is_lacuna - The 'reading' represents a known gap in the text.
+
+=item is_ph - A temporary placeholder for apparatus parsing purposes.  Do not use unless you know what you are doing.
+
+=item rank - The sequence number of the reading. This should probably not be set manually.
+
+=back
+
+One of 'text', 'is_start', 'is_end', or 'is_lacuna' is required.
+
+=head2 collation
+
+=head2 id
+
+=head2 text
+
+=head2 is_start
+
+=head2 is_end
+
+=head2 is_lacuna
+
+=head2 rank
+
+Accessor methods for the given attributes.
+
+=cut
+
+has 'collation' => (
+	is => 'ro',
+	isa => 'Text::Tradition::Collation',
+	# required => 1,
+	weak_ref => 1,
+	);
+
+has 'id' => (
+	is => 'ro',
+	isa => 'Str',
+	required => 1,
+	);
+
+has 'text' => (
+	is => 'ro',
+	isa => 'Str',
+	required => 1,
+	writer => 'alter_text',
+	);
+	
+has 'is_start' => (
+	is => 'ro',
+	isa => 'Bool',
+	default => undef,
+	);
+
+has 'is_end' => (
+	is => 'ro',
+	isa => 'Bool',
+	default => undef,
+	);
+    
+has 'is_lacuna' => (
+    is => 'ro',
+    isa => 'Bool',
+	default => undef,
     );
+    
+has 'is_ph' => (
+	is => 'ro',
+	isa => 'Bool',
+	default => undef,
+	);
+	
+has 'is_common' => (
+	is => 'rw',
+	isa => 'Bool',
+	default => undef,
+	);
 
-# This contains an array of reading objects; the array is a pool,
-# shared by the reading objects inside the pool.  When a reading is
-# added to the pool, all the same_as attributes should be updated.
-has 'same_as' => (
+has 'rank' => (
     is => 'rw',
-    isa => 'ArrayRef[Text::Tradition::Collation::Reading]',
+    isa => 'Int',
+    predicate => 'has_rank',
+    clearer => 'clear_rank',
     );
+    
+## For morphological analysis
+
+has 'normal_form' => (
+	is => 'rw',
+	isa => 'Str',
+	predicate => 'has_normal_form',
+	);
+
+has 'lemma' => (
+	is => 'rw',
+	isa => 'Str',
+	predicate => 'has_lemma',
+	);
+
+has 'morphology' => (
+	traits => ['Array'],
+	isa => 'ArrayRef[HashRef[ArrayRef[Text::Tradition::Collation::Reading::Morphology]]]',
+	handles => {
+		lexemes => 'elements',
+		has_morphology => 'count',
+		_clear_morph => 'clear',
+		_add_morph => 'push',
+		},
+	);
+	
+## For prefix/suffix readings
+
+has 'join_prior' => (
+	is => 'ro',
+	isa => 'Bool',
+	default => undef,
+	);
+	
+has 'join_next' => (
+	is => 'ro',
+	isa => 'Bool',
+	default => undef,
+	);
 
-# # This is a hash mapping of 'relationship => reading'.
-# # TODO we should validate the relationships sometime.
-has 'relationships' => (
-    is => 'ro',
-    isa => 'HashRef[Text::Tradition::Collation::Reading]',
-    default => sub { {} },
-    );
 
-# Deal with the non-arg option for Graph::Easy's constructor.
 around BUILDARGS => sub {
-    my $orig = shift;
-    my $class = shift;
-
-    my %args;
-    if( @_ == 1 && ref( $_[0] ) ne 'HASH' ) {
-	return $class->$orig( 'name' => $_[0] );
-    } else {
-	return $class->$orig( @_ );
-    }
+	my $orig = shift;
+	my $class = shift;
+	my $args;
+	if( @_ == 1 ) {
+		$args = shift;
+	} else {
+		$args = { @_ };
+	}
+			
+	# If one of our special booleans is set, we change the text and the
+	# ID to match.
+	if( exists $args->{'is_lacuna'} && !exists $args->{'text'} ) {
+		$args->{'text'} = '#LACUNA#';
+	} elsif( exists $args->{'is_start'} ) {
+		$args->{'id'} = '#START#';  # Change the ID to ensure we have only one
+		$args->{'text'} = '#START#';
+		$args->{'rank'} = 0;
+	} elsif( exists $args->{'is_end'} ) {
+		$args->{'id'} = '#END#';	# Change the ID to ensure we have only one
+		$args->{'text'} = '#END#';
+	} elsif( exists $args->{'is_ph'} ) {
+		$args->{'text'} = $args->{'id'};
+	}
+	
+	$class->$orig( $args );
 };
 
-# Initialize the identity pool. 
-sub BUILD {
-    my( $self, $args ) = @_;
-    $self->same_as( [ $self ] );
+=head2 is_meta
+
+A meta attribute (ha ha), which should be true if any of our 'special'
+booleans are true.  Implies that the reading does not represent a bit 
+of text found in a witness.
+
+=cut
+
+sub is_meta {
+	my $self = shift;
+	return $self->is_start || $self->is_end || $self->is_lacuna || $self->is_ph;	
 }
 
-sub merge_from {
-    my( $self, $merged_node ) = @_;
-    # Adopt the identity pool of the other node.
-    my @now_identical = grep { $_ ne $merged_node } @{$merged_node->same_as};
-    my $new_pool = _merge_array_pool( \@now_identical, $self->same_as )
-	if @now_identical;
-
-    # Adopt the relationship attributes of the other node.
-    my $now_rel = $merged_node->relationships;
-    foreach my $key ( %$now_rel ) {
-	if( $self->has_relationship( $key ) ) {
-	    my $related = $self->get_relationship( $key );
-	    if( $now_rel->{$key} ne $related ) {
-		warn( sprintf( "Merged reading %s has relationship %s to reading %s instead of %s; skipping",
-			       $merged_node->name, $key,
-			       $now_rel->{$key},
-			       $related) );
-	    } # else no action needed
-	} else {
-	    $self->set_relationship( $key, $now_rel->{$key} );
-	}
-    }
+=head1 Convenience methods
+
+=head2 related_readings
+
+Calls Collation's related_readings with $self as the first argument.
+
+=cut
+
+sub related_readings {
+	my $self = shift;
+	return $self->collation->related_readings( $self, @_ );
 }
 
+=head2 witnesses 
+
+Calls Collation's reading_witnesses with $self as the first argument.
+
+=cut
+
+sub witnesses {
+	my $self = shift;
+	return $self->collation->reading_witnesses( $self, @_ );
+}
+
+=head2 predecessors
+
+Returns a list of Reading objects that immediately precede $self in the collation.
+
+=cut
+
+sub predecessors {
+	my $self = shift;
+	my @pred = $self->collation->sequence->predecessors( $self->id );
+	return map { $self->collation->reading( $_ ) } @pred;
+}
+
+=head2 successors
+
+Returns a list of Reading objects that immediately follow $self in the collation.
+
+=cut
+
+sub successors {
+	my $self = shift;
+	my @succ = $self->collation->sequence->successors( $self->id );
+	return map { $self->collation->reading( $_ ) } @succ;
+}
+
+=head2 set_identical( $other_reading)
+
+Backwards compatibility method, to add a transposition relationship
+between $self and $other_reading.  Don't use this.
+
+=cut
+
 sub set_identical {
-    my( $self, $other_node ) = @_; 
-    my $enlarged_pool = _merge_array_pool( $self->same_as, 
-					   $other_node->same_as );
-
-    # ...and set this node to point to the enlarged pool.
-    $self->same_as( $enlarged_pool );
-}   
-
-sub _merge_array_pool {
-    my( $pool, $main_pool ) = @_;
-    my %poolhash;
-    foreach ( @$main_pool ) {
-	# Note which nodes are already in the main pool so that we
-	# don't re-add them.
-	$poolhash{$_->name} = 1;
-    }
-
-    foreach( @$pool ) {
-	# Add the remaining nodes to the main pool...
-	push( @$main_pool, $_ ) unless $poolhash{$_->name};
-    }
-    return $main_pool;
+	my( $self, $other ) = @_;
+	return $self->collation->add_relationship( $self, $other, 
+		{ 'type' => 'transposition' } );
 }
 
-sub has_primary {
-    my $self = shift;
-    my $pool = $self->same_as;
-    return $pool->[0]->name eq $self->name;
+sub _stringify {
+	my $self = shift;
+	return $self->id;
 }
 
-sub primary {
-    my $self = shift;
-    return $self->same_as->[0];
+=head1 MORPHOLOGY
+
+A few methods to try to tack on morphological information.
+
+=head2 is_disambiguated
+
+Returns true if there is only one tag per lexeme in this reading.
+
+=cut
+
+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;
 }
 
-# Much easier to do this with a hash than with an array of Relationship objects,
-# which would be the proper OO method.
+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 );
+}
 
-sub has_relationship {
-    my( $self, $rel ) = @_;
-    return exists( $self->relationships->{ $rel } );
+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";
+		return;
+	}
+	# Keep the object at the selected index
+	my $selected = $struct->{$lexeme}->[$index];
+	$struct->{$lexeme} = [ $selected ];
 }
 
-sub get_relationship {
-    my( $self, $rel ) = @_;
-    if( $self->has_relationship( $rel ) ) {
-	return $self->relationships->{ $rel };
-    }
-    return undef;
+sub is_disambiguated {
+	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;
+	}
+	return 1;
 }
 
-sub set_relationship {
-    my( $self, $rel, $value ) = @_;
-    $self->relationships->{ $rel } = $value;
+## Utility methods
+
+sub TO_JSON {
+	my $self = shift;
+	return $self->text;
 }
 
+## TODO will need a throw() here
+
 no Moose;
 __PACKAGE__->meta->make_immutable;
 
-1;
+###################################################
+### Morphology objects, to be attached to readings
+###################################################
 
-######################################################
-## copied from Graph::Easy::Parser docs
-######################################################
-# when overriding nodes, we also need ::Anon
+package Text::Tradition::Collation::Reading::Morphology;
 
-package Text::Tradition::Collation::Reading::Anon;
 use Moose;
-use MooseX::NonMoose;
-extends 'Text::Tradition::Collation::Reading';
-extends 'Graph::Easy::Node::Anon';
-no Moose;
-__PACKAGE__->meta->make_immutable;
 
-1;
-# use base qw/Text::Tradition::Collation::Reading/;
-# use base qw/Graph::Easy::Node::Anon/;
+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.
 
-######################################################
-# and :::Empty
+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 );
+	}
+};
 
-package Text::Tradition::Collation::Reading::Empty;
-use Moose;
-use MooseX::NonMoose;
-extends 'Graph::Easy::Node::Empty';
 no Moose;
 __PACKAGE__->meta->make_immutable;
 
 1;
-# use base qw/Text::Tradition::Collation::Reading/;
 
-######################################################