use Moose;
use Moose::Util::TypeConstraints;
-use MooseX::NonMoose;
+use JSON qw/ from_json /;
+use Module::Load;
+use Text::Tradition::Error;
+use XML::Easy::Syntax qw( $xml10_name_rx $xml10_namestartchar_rx );
+use YAML::XS;
+use overload '""' => \&_stringify, 'fallback' => 1;
-extends 'Graph::Easy::Node';
+subtype 'ReadingID',
+ as 'Str',
+ where { $_ =~ /\A$xml10_name_rx\z/ },
+ message { 'Reading ID must be a valid XML attribute string' };
+
+no Moose::Util::TypeConstraints;
-subtype 'Position'
- => as 'Str',
- => where { $_ =~ /^\d+\,\d+$/ },
- message { 'Position must be of the form x,y' };
+=head1 NAME
-has 'position' => (
- is => 'rw',
- isa => 'Position',
+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 => 'ReadingID',
+ required => 1,
+ );
+
+has 'text' => (
+ is => 'ro',
+ isa => 'Str',
+ required => 1,
+ writer => 'alter_text',
+ );
+
+has 'language' => (
+ is => 'ro',
+ isa => 'Str',
+ predicate => 'has_language',
+ );
+
+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
-# 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 { {} },
- );
+has 'grammar_invalid' => (
+ is => 'rw',
+ isa => 'Bool',
+ default => undef,
+ );
+
+has 'is_nonsense' => (
+ is => 'rw',
+ isa => 'Bool',
+ default => undef,
+ );
+
+has 'normal_form' => (
+ is => 'rw',
+ isa => 'Str',
+ predicate => '_has_normal_form',
+ clearer => '_clear_normal_form',
+ );
+
+# Holds the lexemes for the reading.
+has 'reading_lexemes' => (
+ traits => ['Array'],
+ isa => 'ArrayRef[Text::Tradition::Collation::Reading::Lexeme]',
+ handles => {
+ lexeme => 'get',
+ lexemes => 'elements',
+ has_lexemes => 'count',
+ clear_lexemes => 'clear',
+ add_lexeme => 'push',
+ },
+ default => sub { [] },
+ );
+
+## For prefix/suffix readings
-# Initialize the identity pool.
+has 'join_prior' => (
+ is => 'ro',
+ isa => 'Bool',
+ default => undef,
+ writer => '_set_join_prior',
+ );
+
+has 'join_next' => (
+ is => 'ro',
+ isa => 'Bool',
+ default => undef,
+ writer => '_set_join_next',
+ );
+
+
+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'} && $args->{'is_lacuna'} && !exists $args->{'text'} ) {
+ $args->{'text'} = '#LACUNA#';
+ } elsif( exists $args->{'is_start'} && $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->{'is_end'} ) {
+ $args->{'id'} = '__END__'; # Change the ID to ensure we have only one
+ $args->{'text'} = '#END#';
+ } elsif( exists $args->{'is_ph'} && $args->{'is_ph'} ) {
+ $args->{'text'} = $args->{'id'};
+ }
+
+ # Backwards compatibility for non-XMLname IDs
+ my $rid = $args->{'id'};
+ $rid =~ s/\#/__/g;
+ $rid =~ s/[\/,]/./g;
+ if( $rid !~ /^$xml10_namestartchar_rx/ ) {
+ $rid = 'r'.$rid;
+ }
+ $args->{'id'} = $rid;
+
+ $class->$orig( $args );
+};
+
+# Look for a lexeme-string argument in the build args.
sub BUILD {
- my( $self, $args ) = @_;
-# $self->same_as( [ $self ] );
+ my( $self, $args ) = @_;
+ if( exists $args->{'lexemes'} ) {
+ $self->_deserialize_lexemes( $args->{'lexemes'} );
+ }
}
-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
+# Make normal_form default to text, transparently.
+around 'normal_form' => sub {
+ my $orig = shift;
+ my $self = shift;
+ my( $arg ) = @_;
+ if( $arg && $arg eq $self->text ) {
+ $self->_clear_normal_form;
+ return $arg;
+ } elsif( !$arg && !$self->_has_normal_form ) {
+ return $self->text;
} else {
- $self->set_relationship( $key, $now_rel->{$key} );
+ $self->$orig( @_ );
}
- }
+};
+
+=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 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->set_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;
- }
+=head1 Convenience methods
- foreach( @$pool ) {
- # Add the remaining nodes to the main pool...
- push( @$main_pool, $_ ) unless $poolhash{$_->name};
- }
- return $main_pool;
+=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, @_ );
}
-# Much easier to do this with a hash than with an array of Relationship objects,
-# which would be the proper OO method.
+=head2 witnesses
-sub has_relationship {
- my( $self, $rel ) = @_;
- return exists( $self->relationships->{ $rel } );
+Calls Collation's reading_witnesses with $self as the first argument.
+
+=cut
+
+sub witnesses {
+ my $self = shift;
+ return $self->collation->reading_witnesses( $self, @_ );
}
-sub get_relationship {
- my( $self, $rel ) = @_;
- if( $self->has_relationship( $rel ) ) {
- return $self->relationships->{ $rel };
- }
- return undef;
+=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;
}
-sub set_relationship {
- my( $self, $rel, $value ) = @_;
- $self->relationships->{ $rel } = $value;
+=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;
}
-no Moose;
-__PACKAGE__->meta->make_immutable;
+=head2 set_identical( $other_reading)
-1;
+Backwards compatibility method, to add a transposition relationship
+between $self and $other_reading. Don't use this.
+
+=cut
+
+sub set_identical {
+ my( $self, $other ) = @_;
+ return $self->collation->add_relationship( $self, $other,
+ { 'type' => 'transposition' } );
+}
+
+sub _stringify {
+ my $self = shift;
+ return $self->id;
+}
+
+=head1 MORPHOLOGY
+
+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 has_lexemes
-######################################################
-## copied from Graph::Easy::Parser docs
-######################################################
-# when overriding nodes, we also need ::Anon
+Returns a true value if the reading has any attached lexemes.
-package Text::Tradition::Collation::Reading::Anon;
+=head2 lexemes
-use base qw/Text::Tradition::Collation::Reading/;
-use base qw/Graph::Easy::Node::Anon/;
+Returns the Lexeme objects (if any) attached to the reading.
-######################################################
-# and :::Empty
+=head2 clear_lexemes
-package Text::Tradition::Collation::Reading::Empty;
+Wipes any associated Lexeme objects out of the reading.
-use base qw/Text::Tradition::Collation::Reading/;
+=head2 add_lexeme( $lexobj )
-######################################################
+Adds the Lexeme in $lexobj to the list of lexemes.
+
+=head2 lemmatize
+
+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.
+
+=cut
+
+sub lemmatize {
+ my $self = shift;
+ unless( $self->has_language ) {
+ warn "Please set a language to lemmatize a tradition";
+ return;
+ }
+ my $mod = "Text::Tradition::Language::" . $self->language;
+ load( $mod );
+ $mod->can( 'reading_lookup' )->( $self );
+
+}
+
+# For graph serialization. Return a JSON representation of the associated
+# reading lexemes.
+sub _serialize_lexemes {
+ my $self = shift;
+ my $json = JSON->new->allow_blessed(1)->convert_blessed(1);
+ return $json->encode( [ $self->lexemes ] );
+}
+
+# Given a JSON representation of the lexemes, instantiate them and add
+# them to the reading.
+sub _deserialize_lexemes {
+ my( $self, $json ) = @_;
+ my $data = from_json( $json );
+ return unless @$data;
+
+ # Need to have the lexeme module in order to have lexemes.
+ eval { use Text::Tradition::Collation::Reading::Lexeme; };
+ throw( $@ ) if $@;
+
+ # Good to go - add the lexemes.
+ my @lexemes;
+ foreach my $lexhash ( @$data ) {
+ push( @lexemes, Text::Tradition::Collation::Reading::Lexeme->new(
+ 'JSON' => $lexhash ) );
+ }
+ $self->clear_lexemes;
+ $self->add_lexeme( @lexemes );
+}
+
+sub disambiguated {
+ my $self = shift;
+ return 0 unless $self->has_lexemes;
+ return !grep { !$_->is_disambiguated } $self->lexemes;
+}
+
+## Utility methods
+
+sub TO_JSON {
+ my $self = shift;
+ return $self->text;
+}
+
+sub throw {
+ Text::Tradition::Error->throw(
+ 'ident' => 'Reading error',
+ 'message' => $_[0],
+ );
+}
+
+no Moose;
+__PACKAGE__->meta->make_immutable;
+
+1;
projects / scpubgit/stemmatology.git / blobdiff