X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=lib%2FText%2FTradition%2FCollation%2FReading.pm;h=f445420829ba97d6230bc5d2d1f958cc2530ea8a;hb=339786dd0b0c493786af4df1ec3bc7ef2bf497e2;hp=7a991024f520860afd683e0b49ff38fb2006a09f;hpb=2acf0892d2cfca8036e2208e4f83d0404d761695;p=scpubgit%2Fstemmatology.git diff --git a/lib/Text/Tradition/Collation/Reading.pm b/lib/Text/Tradition/Collation/Reading.pm index 7a99102..f445420 100644 --- a/lib/Text/Tradition/Collation/Reading.pm +++ b/lib/Text/Tradition/Collation/Reading.pm @@ -1,12 +1,26 @@ package Text::Tradition::Collation::Reading; use Moose; +use Moose::Util::TypeConstraints; +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; +subtype 'ReadingID', + as 'Str', + where { $_ =~ /\A$xml10_name_rx\z/ }, + message { 'Reading ID must be a valid XML attribute string' }; + +no Moose::Util::TypeConstraints; + =head1 NAME -Text::Tradition::Collation::Reading - represents a reading (usually a word) in a collation. - +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 @@ -19,12 +33,13 @@ a particular sequence =head2 new -Creates a new reading in the given collation with the given attributes. +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 collation - The Text::Tradition::Collation object to which this +reading belongs. Required. =item id - A unique identifier for this reading. Required. @@ -36,9 +51,11 @@ Options include: =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 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. +=item rank - The sequence number of the reading. This should probably not +be set manually. =back @@ -71,7 +88,7 @@ has 'collation' => ( has 'id' => ( is => 'ro', - isa => 'Str', + isa => 'ReadingID', required => 1, ); @@ -82,6 +99,12 @@ has 'text' => ( writer => 'alter_text', ); +has 'language' => ( + is => 'ro', + isa => 'Str', + predicate => 'has_language', + ); + has 'is_start' => ( is => 'ro', isa => 'Bool', @@ -116,7 +139,59 @@ has 'rank' => ( is => 'rw', isa => 'Int', predicate => 'has_rank', + clearer => 'clear_rank', ); + +## For morphological analysis + +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 + +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 { @@ -131,22 +206,54 @@ around BUILDARGS => sub { # 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'} ) { + if( exists $args->{'is_lacuna'} && $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 + } 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->{'id'} = '#END#'; # Change the ID to ensure we have only one + } 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'} ) { + } 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 ) = @_; + if( exists $args->{'lexemes'} ) { + $self->_deserialize_lexemes( $args->{'lexemes'} ); + } +} + +# 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 { + $self->$orig( @_ ); + } +}; + =head2 is_meta A meta attribute (ha ha), which should be true if any of our 'special' @@ -226,13 +333,101 @@ sub _stringify { 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 for more information +on Lexeme objects and their attributes. + +=head2 has_lexemes + +Returns a true value if the reading has any attached lexemes. + +=head2 lexemes + +Returns the Lexeme objects (if any) attached to the reading. + +=head2 clear_lexemes + +Wipes any associated Lexeme objects out of the 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 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; -