package Plucene::Analysis::PorterStemFilter;
=head1 NAME
Plucene::Analysis::PorterStemFilter - Porter stemming on the token stream
=head1 SYNOPSIS
# isa Plucene::Analysis:::TokenFilter
my $token = $porter_stem_filter->next;
=head1 DESCRIPTION
This class transforms the token stream as per the Porter stemming algorithm.
Note: the input to the stemming filter must already be in lower case, so you
will need to use LowerCaseFilter or LowerCaseTokenizer farther down the
Tokenizer chain in order for this to work properly!
The Porter Stemmer implements Porter Algorithm for normalization of English
words by stripping their extensions and is used to generalize the searches.
For example, the Porter algorithm maps both 'search' and 'searching'
(as well as 'searchnessing') to 'search' such that a query for 'search' will
also match documents that contains the word 'searching'.
Note that the Porter algorithm is specific to the English language and may give
unpredictable results for other languages. Also, make sure to use the same
analyzer during the indexing and the searching.
You can find more information on the Porter algorithm at
www.tartarus.org/~martin/PorterStemmer.
A nice online demonstration of the Porter algorithm is available at
www.scs.carleton.ca/~dquesnel/java/stuff/PorterApplet.html.
=head1 METHODS
=cut
use strict;
use warnings;
use Lingua::Stem::En;
Lingua::Stem::En::stem_caching({ -level => 2 });
use base 'Plucene::Analysis::TokenFilter';
=head2 next
my $token = $porter_stem_filter->next;
Returns the next input token, after being stemmed.
=cut
sub next {
my $self = shift;
my $t = $self->input->next or return;
$t->text(@{ Lingua::Stem::En::stem({ -words => [ $t->text ] }) });
return $t;
}
1;