view release on metacpan or  search on metacpan

lib/Lingua/FreeLing3/HMMTagger.pm  view on Meta::CPAN

States whether words that carry retokenization information (e.g. set
by the dictionary or affix handling modules) must be retokenized (that
is, splitted in two or more words) after the tagging.

=item C<AmbiguityResolution> (option)

States whether and when the tagger must select only one analysis in
case of ambiguity. Possible values are: FORCE_NONE: no selection
forced, words ambiguous after the tagger, remain
ambiguous. FORCE_TAGGER: force selection immediately after tagging,
and before retokenization. FORCE_RETOK: force selection after
retokenization.

=item C<KBest> (integer)

This option, only available with FreeLing 3.1, states how many best
tag sequences the tagger must try to compute. If not specified, this
parameter defaults to 1. Since a sentence may have less possible tag
sequences than the given k value, the results may contain a number of
sequences smaller than k.

=back

=cut

sub new {
    my ($class, $lang, %ops) = @_;

    my $config = Lingua::FreeLing3::Config->new($lang);
    my $file = $config->config("TaggerHMMFile");

    unless (-f $file) {
        carp "Cannot find hmm_tagger data file. Tried [$file]\n";
        return undef;
    }

    my $retok = Lingua::FreeLing3::_validate_bool($ops{Retokenize},
                                                  $config->config('TaggerRetokenize')); # bool
    my $ft = $config->config("TaggerForceSelect");
    $ft = "FORCE_NONE"   if $ft eq "none";
    $ft = "FORCE_TAGGER" if $ft eq "tagger";
    $ft = "FORCE_RETOK"  if $ft eq "retok";
    my $amb   = Lingua::FreeLing3::_validate_option($ops{AmbiguityResolution},
                                                    {
                                                     FORCE_NONE   => 0,
                                                     FORCE_TAGGER => 1,
                                                     FORCE_RETOK  => 2,
                                                    }, $ft);

    my $kbest = $ops{KBest} || 1;

    my $self;

    if (Lingua::FreeLing3::ConfigData->config("fl_minor") == 0) {
        $self = $class->SUPER::new($lang, $file, $retok, $amb);
    } else {
        $self = $class->SUPER::new($file, $retok, $amb, $kbest);
    }

    return bless $self => $class
}


=head2 C<tag>

Alias to C<analyze>

=cut

sub tag { &analyze }

=head2 C<analyze>

Receives a list of sentences, and returns that same list of sentences
after tagging process. Basically, selected the most probable
(accordingly with the tagger model) analysis for each word.

=cut

sub analyze {
    my ($self, $sentences, %opts) = @_;

    unless (Lingua::FreeLing3::_is_sentence_list($sentences)) {
        carp "Error: analyze argument isn't a list of sentences";
        return undef;
    }

    $sentences = $self->SUPER::analyze($sentences);

    for my $s (@$sentences) {
	$s->ACQUIRE();
        $s = Lingua::FreeLing3::Sentence->_new_from_binding($s);
    }
    return $sentences;
}


1;

__END__

=head1 SEE ALSO

Lingua::FreeLing3 (3), freeling, perl(1)

=head1 AUTHOR

Alberto Manuel Brandão Simões, E<lt>ambs@cpan.orgE<gt>

Jorge Cunha Mendes E<lt>jorgecunhamendes@gmail.comE<gt>

=head1 COPYRIGHT AND LICENSE

Copyright (C) 2011-2013 by Projecto Natura

=cut