<?php
/**
* SeekQuarry/Yioop --
* Open Source Pure PHP Search Engine, Crawler, and Indexer
*
* Copyright (C) 2009 - 2015 Chris Pollett chris@pollett.org
*
* LICENSE:
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* 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, see <http://www.gnu.org/licenses/>.
*
* @author Chris Pollett chris@pollett.org
* @license http://www.gnu.org/licenses/ GPL3
* @link http://www.seekquarry.com/
* @copyright 2009 - 2015
* @filesource
*/
namespace seekquarry\yioop\locale\nl\resources;
/**
* This class has a collection of methods for Dutch locale specific
* tokenization. In particular, it has a stemmer, .
*
* @author Charles Bocage
* @package seek_quarry\locale\nl
*/
class Tokenizer
{
/**
* Words we don't want to be stemmed
* @var array
*/
public static $no_stem_list = [
"abs", "ahs", "aken", "àlle", "als", "are",
"allèen", "ate", "aten", "azen", "bse", "cfce", "curaçao",
"dègelijk", "dme", "ede", "eden", "eds", "ehs", "ems", "ene", "epe",
"eps", "ers", "eten", "ets", "even", "fme", "gedaçht", "ghe", "gve",
"hdpe", "hôte", "hpe", "hse", "ibs", "ics", "ile", "ims", "jònge",
"kwe", "ldpe", "lldpe", "lme", "lze", "maitres", "mwe", "nme", "ode",
"ogen", "oke", "ole", "ons", "ònze", "open", "ops", "oren", "ors",
"oss", "oven", "ows", "pre", "pve", "rhône", "ròme", "rwe", "ske",
"sme", "spe", "ste", "the", "tje", "uce", "uden", "uien", "uren",
"use", "uwe", "vse", "ype"
];
/**
* Stub function which could be used for a word segmenter.
* Such a segmenter on input thisisabunchofwords would output
* this is a bunch of words
*
* @param string $pre_segment before segmentation
* @return string should return string with words separated by space
* in this case does nothing
*/
public static function segment($pre_segment)
{
return $pre_segment;
}
/**
* boolean that tells the code if the e suffix was removed in step2 or not
* @var array
*/
static $removed_e_suffix = false;
/**
* Computes the stem of a Dutch word
*
* For example, lichamelijk, lichamelijke, lichamelijkheden and lichamen,
* all have licham as a stem
*
* @param string $word the string to stem
* @return string the stem of $words
*/
public static function stem($word)
{
self::$removed_e_suffix = false;
$result = $word;
if (isset($word) && !empty($word)) {
$result = self::removeAllUmlautAndAcuteAccents($word);
$result = trim(mb_strtolower($result));
$result = self::substituteIAndY($result);
$R1 = self::getRIndex($result, 1);
$R2 = self::getRIndex($result, $R1);
$result = self::step1($result, $R1);
$result = self::step2($result);
$result = self::step3a($result, $R2);
$result = self::step3b($result, $R2);
$result = self::step4($result);
$result = mb_strtolower($result);
}
return $result;
}
/**
* Remove all umlaut and acute accents that need to be removed.
*
* @param string $word the string to remove the umlauts and accents from
* @return string the string with the umlauts and accents removed
*/
private static function removeAllUmlautAndAcuteAccents($word)
{
$result = preg_replace("/\é|\ë/", "e", $word);
$result = preg_replace("/\á|ä/", "a", $result);
$result = preg_replace("/\ó|ö/", "o", $result);
$result = preg_replace("/\ï/", "i", $result);
$result = preg_replace("/\ü|\ú/", "u", $result);
return $result;
}
/**
* Put initial y, y after a vowel, and i between vowels into upper case.
*
* @param string $word the string to put initial y, y after a vowel, and
* i between vowels into upper case.
* @return string the string with an initial y, y after a vowel, and i
* between vowels into upper case.
*/
private static function substituteIAndY($word)
{
$word_split = preg_split('/(?!^)(?=.)/u', $word);
for ($i = 0; $i < count($word_split); $i++) {
if ($i == 0) {
if ($word_split[$i] == 'y') {
$word_split[$i] = 'Y';
}
} else {
if ($word_split[$i] == 'i') {
if (count($word_split) > ($i + 1)) {
if (self::isVowel($word_split[$i - 1]) &&
self::isVowel($word_split[$i + 1])) {
$word_split[$i] = 'I';
}
}
} elseif ($word_split[$i] == 'y') {
if (self::isVowel($word_split[$i - 1])) {
$word_split[$i] = 'Y';
}
}
}
}
return implode('', $word_split);
}
/**
* Check that the letter is a vowel
*
* @param string $letter the character to check
* @return boolean true if it is a vowel, otherwise false
*/
private static function isVowel($letter)
{
$result = false;
switch ($letter) {
case 'e':
// no break
case 'a':
case 'o':
case 'i':
case 'u':
case 'y':
case 'è':
$result = true;
}
return $result;
}
/**
* Get the R index. The R index is the first consonent that follows a
* vowel after the $start index
*
* @param string $word the string to search for the R index
* @param int $start the index to start searching for the R index in the
* string
* @return int the R index if found, otherwise -1
*/
private static function getRIndex($word, $start)
{
$result = -1;
$word_split = preg_split('/(?!^)(?=.)/u', $word);
for ($i = $start <= 0 ? 1 : $start; $i < count($word_split); $i++) {
if (!self::isVowel($word_split[$i]) &&
self::isVowel($word_split[$i - 1])) {
$result = ($i + 1);
break;
}
}
return $result;
}
/**
* Define a valid en-ending as a non-vowel, and not gem and remove it
*
* @param string $word the string to stem
* @param int $R1 the int that represents the R index
* @return string the string with the valid en-ending as a non-vowel, and
* not gem ending removed
*/
private static function step1($word, $R1)
{
$result = $word;
if ($R1 > -1) {
$wordLength = strlen($word);
if ($wordLength > 2 && $R1 < $wordLength) {
if (self::endsWith($word, "heden")) {
$result = self::replace($word, '/heden$/', 'heid', $R1);
} else {
if (preg_match("/(?<![aeioèuy]|gem)(ene?)$/", $word,
$matches, 0, $R1) && $word != "eten" &&
$word != "even" && $word != "opene") {
$result = self::undouble(self::replace($word,
'/ene$/', '', $R1));
$result = self::undouble(self::replace($result,
'/en$/', '', $R1));
} else {
if (preg_match("/(?<![aeiouyèj])se$/", $word,
$matches, 0, $R1) && $word != "osse") {
$result = self::replace($word,
'/(?<![aeiouyèj])se$/', '', $R1);
} elseif (preg_match("/(?<![aeiouyèj])s$/", $word,
$matches, 0, $R1)) {
$result = self::replace($word,
'/(?<![aeiouyèj])s$/', '', $R1);
}
}
}
}
}
return $result;
}
/**
* Delete the suffix e if in R1 and preceded by a non-vowel, and then
* undouble the ending
*
* @param string $word the string to delete the suffix e if in R1 and
* preceded by a non-vowel, and then undouble the ending
* @return string the string with the suffix e if in R1 and preceded by
* a non-vowel deleted, and then undouble the ending
*/
private static function step2($word)
{
$result = $word;
if (self::endsWith($word, "e")
&& $word != "ene" && $word != "oce" && $word != "ohe" &&
$word != "pâte") {
$word_split = preg_split('/(?!^)(?=.)/u', $word);
$word_split_length = count($word_split);
if ($word_split_length > 2 &&
!self::isVowel($word_split[$word_split_length - 2])) {
$word_split = array_slice($word_split, 0,
$word_split_length - 1);
$result = implode('', $word_split);
$result = self::undouble($result);
self::$removed_e_suffix = true;
}
}
return $result;
}
/**
* Delete the letters heid if in R2 and not preceded by a c, and treat an a
* preceding en like in step 1
*
* @param string $word the string to delete the letters heid if in R2 and
* not preceded by a c, and treat an a preceding en like in step 1
* @param int $R2 the R index
* @return string the string with the letters heid if in R2 and not
* preceded by a c deleted, and treated an a preceding en like in step 1
*/
private static function step3a($word, $R2)
{
$result = $word;
if ($R2 > -1) {
if (preg_match("/(?<![c])(heid)$/", $word, $matches, 0, $R2)) {
$result = self::replace($word, '/(?<![c])(heid)$/', '', $R2);
if (preg_match("/(?<![aeiouyè]|gem)(ene?)$/", $result,
$matches, 0, $R2)) {
$result = self::undouble(self::replace($result,
'/(?<![aeiouyè]|gem)(ene?)$/', '', $R2));
}
}
}
return $result;
}
/**
* Search for the longest among the following suffixes, and perform the
* action indicated.
* If in R2 and ends with eigend, eigingm igend or iging remove it
* If in R2 and ends with ig preceded by an e remove it
* If in R2 and ends with lijk, baar or bar then remove it
*
* @param string $word the string to stem
* @param int $R2 the R index
* @return string the string with the various endings removed if they exist
*/
public static function step3b($word, $R2)
{
$result = $word;
$word_split = preg_split('/(?!^)(?=.)/u', $word);
if (count($word_split) > 2) {
if ($R2 > -1) {
if (preg_match('/(end|ing)$/', $word, $matches, 0, $R2)) {
if (preg_match('/eig(end|ing)$/', $word, $matches, 0,
$R2)) {
$result = self::replace($word, '/(end|ing)$/', '',
$R2);
}
elseif (preg_match('/ig(end|ing)$/', $word, $matches, 0,
$R2)) {
$result = self::replace($word, '/(igend|iging)$/', '',
$R2);
$result = self::undouble($result);
}
else {
$result = self::replace($word, '/(end|ing)$/', '',
$R2);
$result = self::undouble($result);
}
}
elseif (preg_match("/(?<![e])ig$/", $word, $matches, 0, $R2)) {
$result = self::replace($word, '/(?<![e])ig$/', '', $R2);
}
elseif (preg_match("/lijk$/", $word, $matches, 0, $R2)) {
$result = self::replace($word, '/lijk$/', '', $R2);
$result = self::step2($result);
}
elseif (preg_match("/baar$/", $word, $matches, 0, $R2)) {
$result = self::replace($word, '/baar$/', '', $R2);
}
elseif (preg_match("/bar$/", $word, $matches, 0, $R2) &&
self::$removed_e_suffix) {
$result = self::replace($word, '/bar$/', '', $R2);
}
}
}
return $result;
}
/**
* If the words ends CVD, where C is a non-vowel, D is a non-vowel
* other than I, and V is double a, e, o or u, remove one of the
* vowels from V (for example, moom -> mon, weed -> wed).
*
* @param string $word the string to check for the CVD combination
* @return string the string with the CVD combination removed otherwise
* the original string
*/
private static function step4($word)
{
$result = $word;
$word_split = preg_split('/(?!^)(?=.)/u', $word);
$numberOfLetters = count($word_split);
if ($numberOfLetters > 3) {
$c = $word_split[$numberOfLetters - 4];
$v1 = $word_split[$numberOfLetters - 3];
$v2 = $word_split[$numberOfLetters - 2];
$d = $word_split[$numberOfLetters - 1];
if (!self::isVowel($c) &&
self::isVowel($v1) &&
self::isVowel($v2) &&
!self::isVowel($d) &&
$v1 == $v2 &&
$d != 'I' &&
$v1 != 'i') {
unset($word_split[$numberOfLetters - 2]);
$result = implode('', $word_split);
}
}
return $result;
}
/**
* Replace a string based on a regex expression
*
* @param string $word the string to search for regex replacement
* @param string $reges the regex to use to find and replacement
* @param string $replace the string to replace if the pattern is matched
* @param int $offset the int to start to look for the regex replacement
* @return string the string with the characters replaced if the regex
* matches, otherwise the original string
*/
private static function replace($word, $regex, $replace, $offset)
{
$result = "";
if ($offset > 0) {
$part1 = substr($word, 0, $offset);
$part2 = substr($word, $offset, strlen($word));
$part2 = preg_replace($regex, $replace, $part2);
$result = $part1 . "" . $part2;
} else {
$result = preg_replace($regex, $replace, $word);
}
return $result;
}
/**
* Checks to see if a string ends with a certain string
*
* @param string $haystack the string to check
* @param string $needle the string to match at the end
* @return boolean true if it ends with $needle, otherwise false
*/
private static function endsWith($haystack, $needle, $case = true)
{
if ($case) {
return (strcmp(substr($haystack,
strlen($haystack) - strlen($needle)), $needle) === 0);
}
return (strcasecmp(substr($haystack,
strlen($haystack) - strlen($needle)), $needle) === 0);
}
/**
* undoubles the end of a string. If the string ends in kk, tt, dd remove
* one of the characters
*
* @param string $word the string to undouble
* @return string the undoubled string, otherwise the original string
*/
private static function undouble($word)
{
$result = $word;
if (self::endsWith($word, "kk") ||
self::endsWith($word, "tt") ||
self::endsWith($word, "dd")) {
$result = substr($word, 0, strlen($word) - 1);
}
return $result;
}
}