File Coverage

blib/lib/Lingua/Stem/Snowball/Se.pm

Criterion	Covered	Total	%
statement	91	97	93.8
branch	40	50	80.0
condition	1	3	33.3
subroutine	9	9	100.0
pod	0	4	0.0
total	141	163	86.5

package Lingua::Stem::Snowball::Se;
use strict;
use bytes;
# -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
# Lingua::Stem::Snowball::Se - Swedish stemmer
# :: based upon the swedish stemmer algorithm at snowball.tartarus.org.
#	 by Martin Porter.
# (c) 2001-2007 Ask Solem Hoel
#
#   This program is free software; you can redistribute it and/or modify
#   it under the terms of the GNU General Public License version 2,
#   *NOT* "earlier versions", as published by the Free Software Foundation.
#
#   This program is distributed in the hope that it will be useful,
#   but WITHOUT ANY WARRANTY; without even the implied warranty of
#   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#   GNU General Public License for more details.
#
#   You should have received a copy of the GNU General Public License
#   along with this program; if not, write to the Free Software
#   Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
# -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
#####


use vars qw($VERSION);
$VERSION = 1.2;

# special characters
my $ae = chr(0xe4); # a"
my $ao = chr(0xe5); # ao
my $oe = chr(0xf6); # o"

# delete the s if a "s ending" is preceeded by one
# of these characters.
my $s_ending = "bcdfghjklmnoprtvy";

# swedish vowels.
my $vowels = "aeiouy$ae$ao$oe";

# ####
# the endings in step 1
# XXX: these must be sorted by length
# to save time we've done it already, you can do it like this:
#	my $bylength = sub {
#		length $a <=> length $b;
#	}
#	@endings = reverse sort $bylength @endings;
my @endings = qw/
   heterna hetens ornas ernas andet heter arnas arens heten andes
   anden orna erna erns arna ades aren aste arne ande ast het ens
   ern ade are er at es ad as en or ar e a s
/;

# the endings in step 2
# XXX: these must be sorted by length, like @endings in step 1.
my @endings2 = ('fullt', "l${oe}st", 'els', 'lig', 'ig');

my %cache = ( );

sub new {
	my $pkg = shift;
	$pkg = ref $pkg || $pkg;
	my %arg = @_;
	my $self = {};
	bless $self, $pkg;
	if($arg{use_cache}) {
		$self->use_cache(1);
	}
	return $self;
}

sub use_cache {
	my($self, $use_cache) = @_;
	if($use_cache) {
		$self->{USE_CACHE} = 1;
	}
	return $self->{USE_CACHE};
}

sub stem {
	my ($self, $word) = @_;
    no warnings;
	my $orig_word;

	if($self->use_cache()) {
		$orig_word = $word;
		my $cached_word = $cache{$word};
		return $cached_word if $cached_word;
	}

	my ($ls, $rs, $wlen, $lslen, $rslen) = getsides($word);
	return $word unless $lslen >= 3;

	# ### STEP 1
	# only need to refresh wlen each time we change the word.
	foreach my $ending (@endings)  {
		my $endinglen = length $ending; # do this once.

		# only continue if the word has this ending at all.
		if(substr($rs, $rslen - $endinglen, $rslen) eq $ending) {
			if($ending eq 's') { # b)
				# check if it has a valid "s ending"...
				my $valid_s_ending = 0;
				if($rslen == 1) {
					my $wmr1 = substr($word, 0, $wlen - $rslen);
					if($wmr1 =~ /[$s_ending]$/o) {
						$valid_s_ending = 1;
					}
				}
				else {
					if(substr($rs, $rslen - 2, $rslen - 1) =~ /[$s_ending]/o) {
						$valid_s_ending = 1;
					}
				}
				if($valid_s_ending) {
					# ...delete the last character (which is a s)
					$word = substr($word, 0, $wlen - 1);
					($ls, $rs, $wlen, $lslen, $rslen) = getsides($word);
					last;
				}
			}
			else { # a)
				# delete the ending.
				$word = substr($word, 0, $wlen - $endinglen);
				($ls, $rs, $wlen, $lslen, $rslen) = getsides($word);
				last;
			}
		}
	}
	return $word unless $lslen >= 3;

	# ### STEP 2
	my $ending = substr($rs, $rslen - 2, $rslen);
	if($ending =~ /dd|gd|nn|dt|gt|kt|tt/) {
		$word = substr($word, 0, $wlen - 1);
		($ls, $rs, $wlen, $lslen, $rslen) = getsides($word);
	}
	return $word unless $lslen >= 3;

	# ### STEP 3
	foreach my $ending (@endings2) {
		if($rs =~ /\Q$ending\E$/)
		{
			if($ending =~ /lig|ig|els/)
			{
				my $res = substr($word, 0, $wlen - length($ending));
				$word = $res; #if length $res > 2;
				last
			}
			if($ending eq "l${oe}st")
			{
				my $res = substr($word, 0, $wlen - 1);
				$word = $res; #if length $res > 2;
				last
			}
			if($ending eq 'fullt')
			{
				my $res = substr($word, 0, $wlen - 1);
				$word = $res; #if length $res > 2;
				last
			}
		}
	}

	if($self->use_cache()) {
		$cache{$orig_word} = $word;
	}

	return $word;
}

sub getsides {
    my $word = shift;
    no warnings;
    my $wlen = length $word;

    my($ls, $rs) = (undef, undef); # left side and right side.

    # ###
    # find the first vowel with a non-vowel after it.
    my($found_vowel, $nonv_position, $curpos) = (-1, -1, 0);
    foreach(split//, $word) {
        if($found_vowel> 0) {
			if(/[^$vowels]/o) {
				if($curpos > 0) {
				$nonv_position = $curpos + 1;
				last;
				}
			}
        }
        if(/[$vowels]/o) {
            $found_vowel = 1;
        }
        $curpos++;
    }

	# got nothing: return false
	return undef if $nonv_position < 0;

    # ###
    # length of the left side must be atleast 3 chars.
    my $leftlen = $wlen - ($wlen - $nonv_position);
    if($leftlen < 3) {
        $ls = substr($word, 0, 3);
        $rs = substr($word, 3, $wlen);
    }
    else {
        $ls = substr($word, 0, $leftlen);
        $rs = substr($word, $nonv_position, $wlen);
    }
    return($ls, $rs, $wlen, length $ls, length $rs);
}

1;

__END__