[scpubgit/stemmatology.git] / lib / Text / Tradition / Collation / Reading / WordForm.pm

package Text::Tradition::Collation::Reading::WordForm;

use Lingua::Features::Structure;
use JSON ();
use Moose;

=head1 NAME

Text::Tradition::Collation::Reading::WordForm - represents a
language/lemma/morphology triplet that can be associated with a Reading.

=head1 DESCRIPTION

Text::Tradition is a library for representation and analysis of collated
texts, particularly medieval ones.  A word form is used for the analysis of
Reading objects; it consists of a lemma, a language, and a code to
represent its part of speech.  In general the word forms for a particular
language should be read from / written to some morphological database.

=head1 METHODS

=head2 new

Creates a new word form from the passed options.

=head2 language

Returns the language to which this word form belongs.

=head2 lemma

Returns the lemma for the word form.

=head2 morphology

Returns an array representing this word's morphology. The contents of the
array depend on the language being used.

=cut

has 'language' => (
	is => 'ro',
	isa => 'Str',
	required => 1,
	);
	
# TODO do we need this?
has 'form' => (
	is => 'ro',
	isa => 'Str',
	# required => 1,
	);
	
has 'lemma' => (
	is => 'ro',
	isa => 'Str',
	required => 1,
	);
	
has 'morphology' => (
	is => 'ro',
	isa => 'Lingua::Features::Structure',
	required => 1,
	);
	
around BUILDARGS => sub {
	my $orig = shift;
	my $class = shift;
	my $args = @_ == 1 ? $_[0] : { @_ };
	if( exists $args->{'JSON'} ) {
		$DB::single = 1;
		my @data = split( / \/\/ /, $args->{'JSON'} );
		print STDERR "Attempting to parse " . $data[2] . " into structure";
		my $morph = Lingua::Features::Structure->from_string( $data[2] );
		$args = { 'language' => $data[0], 'lemma' => $data[1],
			'morphology' => $morph };
	}
	$class->$orig( $args );
};
	
=head2 to_string

Returns a string combination of language/lemma/morphology that can be used
in equivalence testing.

=cut

sub to_string {
	my $self = shift;
	return JSON->new->convert_blessed(1)->encode( $self );
}

# Rather than spitting it out as a JSON hash, encode it as a string so that
# the XML serialization doesn't become insane.
sub TO_JSON {
	my $self = shift;
	return sprintf( "%s // %s // %s", $self->language, $self->lemma,
		$self->morphology->to_string() );
}
	
no Moose;
__PACKAGE__->meta->make_immutable;

1;

=head1 LICENSE

This package is free software and is provided "as is" without express
or implied warranty.  You can redistribute it and/or modify it under
the same terms as Perl itself.

=head1 AUTHOR

Tara L Andrews E<lt>aurum@cpan.orgE<gt>
Commit	Line	Data
cca4f996	1	package Text::Tradition::Collation::Reading::WordForm;
cca4f996	2
70745e70	3	use Lingua::Features::Structure;
7604424b	4	use JSON ();
7604424b	5	use Moose;
cca4f996	6
	7	=head1 NAME
	8
	9	Text::Tradition::Collation::Reading::WordForm - represents a
	10	language/lemma/morphology triplet that can be associated with a Reading.
	11
	12	=head1 DESCRIPTION
	13
	14	Text::Tradition is a library for representation and analysis of collated
	15	texts, particularly medieval ones. A word form is used for the analysis of
	16	Reading objects; it consists of a lemma, a language, and a code to
	17	represent its part of speech. In general the word forms for a particular
	18	language should be read from / written to some morphological database.
	19
	20	=head1 METHODS
	21
	22	=head2 new
	23
	24	Creates a new word form from the passed options.
	25
	26	=head2 language
	27
	28	Returns the language to which this word form belongs.
	29
	30	=head2 lemma
	31
	32	Returns the lemma for the word form.
	33
	34	=head2 morphology
	35
	36	Returns an array representing this word's morphology. The contents of the
	37	array depend on the language being used.
	38
	39	=cut
	40
	41	has 'language' => (
	42	is => 'ro',
	43	isa => 'Str',
	44	required => 1,
	45	);
	46
	47	# TODO do we need this?
	48	has 'form' => (
	49	is => 'ro',
	50	isa => 'Str',
	51	# required => 1,
	52	);
	53
	54	has 'lemma' => (
	55	is => 'ro',
	56	isa => 'Str',
	57	required => 1,
	58	);
	59
	60	has 'morphology' => (
	61	is => 'ro',
6ad2ce78	62	isa => 'Lingua::Features::Structure',
cca4f996	63	required => 1,
	64	);
	65
70745e70	66	around BUILDARGS => sub {
	67	my $orig = shift;
	68	my $class = shift;
	69	my $args = @_ == 1 ? $_[0] : { @_ };
7604424b	70	if( exists $args->{'JSON'} ) {
da83693e	71	$DB::single = 1;
	72	my @data = split( / \/\/ /, $args->{'JSON'} );
	73	print STDERR "Attempting to parse " . $data[2] . " into structure";
	74	my $morph = Lingua::Features::Structure->from_string( $data[2] );
	75	$args = { 'language' => $data[0], 'lemma' => $data[1],
	76	'morphology' => $morph };
70745e70	77	}
	78	$class->$orig( $args );
	79	};
	80
6ad2ce78	81	=head2 to_string
	82
	83	Returns a string combination of language/lemma/morphology that can be used
	84	in equivalence testing.
	85
	86	=cut
	87
	88	sub to_string {
cca4f996	89	my $self = shift;
7604424b	90	return JSON->new->convert_blessed(1)->encode( $self );
	91	}
	92
da83693e	93	# Rather than spitting it out as a JSON hash, encode it as a string so that
da83693e	94	# the XML serialization doesn't become insane.
7604424b	95	sub TO_JSON {
7604424b	96	my $self = shift;
da83693e	97	return sprintf( "%s // %s // %s", $self->language, $self->lemma,
da83693e	98	$self->morphology->to_string() );
cca4f996	99	}
6ad2ce78	100
cca4f996	101	no Moose;
	102	__PACKAGE__->meta->make_immutable;
	103
	104	1;
	105
	106	=head1 LICENSE
	107
	108	This package is free software and is provided "as is" without express
	109	or implied warranty. You can redistribute it and/or modify it under
	110	the same terms as Perl itself.
	111
	112	=head1 AUTHOR
	113
	114	Tara L Andrews E<lt>aurum@cpan.orgE<gt>