X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=lib%2FText%2FTradition%2FCollation%2FReading.pm;h=0b57ff8996204600d27f813b6f8a6cad2b1d1b85;hb=a70370723615fd41e59a6b3e48d7afceb89ac288;hp=95912061ff128164b231fd37a053721e43ae5c43;hpb=85bb98307e622f456a33b7d31342b2900d47b51d;p=scpubgit%2Fstemmatology.git

diff --git a/lib/Text/Tradition/Collation/Reading.pm b/lib/Text/Tradition/Collation/Reading.pm
index 9591206..0b57ff8 100644
--- a/lib/Text/Tradition/Collation/Reading.pm
+++ b/lib/Text/Tradition/Collation/Reading.pm
@@ -1,256 +1,234 @@
 package Text::Tradition::Collation::Reading;
 
 use Moose;
-use MooseX::NonMoose;
+use overload '""' => \&_stringify, 'fallback' => 1;
 
-extends 'Graph::Easy::Node';
+=head1 NAME
 
-has 'rank' => (
-    is => 'rw',
-    isa => 'Int',
-    predicate => 'has_rank',
-    );
+Text::Tradition::Collation::Reading - represents a reading (usually a word) in a collation.
+    
+=head1 DESCRIPTION
+
+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 'punctuation' => (
+	traits => ['Array'],
+	isa => 'ArrayRef[HashRef[Str]]',
+	default => sub { [] },
+	handles => {
+			punctuation => 'elements',
+			add_punctuation => 'push',
+			},
+	);
+
+has 'separate_punctuation' => (
+	is => 'ro',
+	isa => 'Bool',
+	default => 1,
+	);
+
+has 'is_start' => (
+	is => 'ro',
+	isa => 'Bool',
+	default => undef,
+	);
+
+has 'is_end' => (
+	is => 'ro',
+	isa => 'Bool',
+	default => undef,
+	);
     
 has 'is_lacuna' => (
-    is => 'rw',
+    is => 'ro',
     isa => 'Bool',
+	default => undef,
     );
+    
+has 'is_ph' => (
+	is => 'ro',
+	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',
     );
 
-# 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( @_ );
-    }
-};
 
-# A lacuna node is also a meta node.
-before is_lacuna => sub {
-	my( $self, $arg ) = @_;
-	if( $arg ) {
-		$self->is_meta( 1 );
+around BUILDARGS => sub {
+	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. 
+# Post-process the given text, stripping punctuation if we are asked.
 sub BUILD {
-    my( $self, $args ) = @_;
-    $self->same_as( [ $self ] );
-}
-
-sub text {
-    # Wrapper function around 'label' attribute.
-    my $self = shift;
-    if( @_ ) {
-        if( defined $_[0] ) {
-        	$self->set_attribute( 'label', $_[0] );
-        } else {
-            $self->del_attribute( 'label' );
-        }
-    }
-    return $self->label;
-}
-
-sub witnessed_by {
-    my( $self, $sigil, $backup ) = @_;
-    my @wits = $self->witnesses;
-    return 1 if grep { $_ eq $sigil } @wits;
-    if( $backup ) {
-        return 1 if grep { $_ eq $backup } @wits;
-    }
-    return 0;
-}
-    
-sub witnesses {
-    my( $self ) = @_;
-    my @paths = grep { $_->get_attribute( 'class' ) eq 'path' } $self->outgoing;
-    push( @paths, grep { $_->get_attribute( 'class' ) eq 'path' } $self->incoming );
-    my %wits;
-    foreach my $p ( @paths ) {
-        if( $p->has_hidden_witnesses ) {
-            foreach ( @{$p->hidden_witnesses} ) {
-                $wits{$_} = 1;
-            }
-        } else {
-            $wits{$p->label} = 1;
-        }
-    }
-    return keys %wits;
-}
-
-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;
-
-    # TODO Adopt the relationship attributes and segment memberships of the other node.
-}
-
-## Dealing with transposed readings.  These methods are only really
-## applicable if we have a linear collation graph.
-
-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 identical_readings {
-    my $self = shift;
-    my @same = grep { $_ ne $self } @{$self->same_as};
-    return @same;
+	my $self = shift;
+	if( $self->separate_punctuation && !$self->is_meta ) {
+		my $pos = 0;
+		my $wspunct = '';  # word sans punctuation
+		foreach my $char ( split( //, $self->text ) ) {
+			if( $char =~ /^[[:punct:]]$/ ) {
+				$self->add_punctuation( { 'char' => $char, 'pos' => $pos } );
+			} else {
+				$wspunct .= $char;
+			}
+			$pos++;
+		}
+		$self->alter_text( $wspunct );
+	}
 }
 
-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;
+sub punctuated_form {
+	my $self = shift;
+	my $word = $self->text;
+	foreach my $p ( sort { $a->{pos} <=> $b->{pos} } $self->punctuation ) {
+		substr( $word, $p->{pos}, 0, $p->{char} );
+	}
+	return $word;
 }
 
-sub has_primary {
-    my $self = shift;
-    my $pool = $self->same_as;
-    return $pool->[0]->name ne $self->name;
-}
+=head2 is_meta
 
-sub primary {
-    my $self = shift;
-    return $self->same_as->[0];
-}
+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.
 
-# Looks from the outside like an accessor for a Boolean, but really 
-# sets the node's class.  Should apply to start, end, and lacunae.
+=cut
 
 sub is_meta {
 	my $self = shift;
-	my $arg = shift;
-	if( defined $arg && $arg ) {
-		$self->set_attribute( 'class', 'meta' );
-	} elsif ( defined $arg ) {
-		$self->del_attribute( 'class' );
-	}
-	return $self->sub_class eq 'meta';	
+	return $self->is_start || $self->is_end || $self->is_lacuna || $self->is_ph;	
 }
 
-# Returns all readings that adjoin this one on any path.
-sub neighbor_readings {
-    my( $self, $direction ) = @_;
-    $direction = 'both' unless $direction;
-    my @paths = grep { $_->isa( 'Text::Tradition::Collation::Path' ) } $self->edges;
-    my %connected;
-    foreach my $p ( @paths ) {
-	if( $p->to eq $self ) {
-	    next if $direction eq 'forward';
-	    $connected{$p->from->name} = $p->from;
-	} else { # $p->from eq $self
-	    next if $direction =~ /^back/;
-	    $connected{$p->to->name} = $p->to;
-	}
-    }
-    return values( %connected );
-}
-
-# Returns all readings related to the one we've got.
+# Some syntactic sugar
 sub related_readings {
-    my( $self, $colocated, $queried ) = @_;
-    $queried = { $self->name => 1 } unless $queried;
-    my @related;
-    # Get the nodes directly related to this one
-    foreach my $e ( $self->edges ) {
-        next unless $e->isa( 'Text::Tradition::Collation::Relationship' );
-        next if $colocated && $e->type eq 'repetition';
-        my $n = $e->from eq $self ? $e->to : $e->from;
-        next if $queried->{$n->name};
-        push( @related, $n );
-    }
-    # Now query those nodes for their relations, recursively
-    map { $queried->{$_->name} = 1 } @related;
-    my @also_related;
-    foreach ( @related ) {
-        push( @also_related, $_->related_readings( $colocated, $queried ) );
-    }
-    push( @related, @also_related );
-    return @related;
+	my $self = shift;
+	return $self->collation->related_readings( $self, @_ );
 }
 
-## Keep track of which readings are unchanged across witnesses.
-sub is_common {
-    my( $self ) = shift;
-    return $self->get_attribute( 'class' ) eq 'common';
+sub predecessors {
+	my $self = shift;
+	my @pred = $self->collation->sequence->predecessors( $self->id );
+	return map { $self->collation->reading( $_ ) } @pred;
 }
 
-## TODO Rationalize make_common, is_meta, etc.
-sub make_common {
-    my( $self ) = shift;
-    $self->set_attribute( 'class', 'common' );
+sub successors {
+	my $self = shift;
+	my @succ = $self->collation->sequence->successors( $self->id );
+	return map { $self->collation->reading( $_ ) } @succ;
 }
 
-sub make_variant {
-    my( $self ) = shift;
-    $self->set_attribute( 'class', 'variant' );
+sub set_identical {
+	my( $self, $other ) = @_;
+	return $self->collation->add_relationship( $self, $other, 
+		{ 'type' => 'transposition' } );
 }
 
-no Moose;
-__PACKAGE__->meta->make_immutable;
-
-1;
-
-######################################################
-## copied from Graph::Easy::Parser docs
-######################################################
-# when overriding nodes, we also need ::Anon
-
-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/;
-
-######################################################
-# and :::Empty
+sub _stringify {
+	my $self = shift;
+	return $self->id;
+}
 
-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/;
 
-######################################################