<?php
    
/**
    * o------------------------------------------------------------------------------o
    * | This package is licensed under the Phpguru license 2008. A quick summary is  |
    * | that the code is free to use for non-commercial purposes. For commercial     |
    * | purposes of any kind there is a small license fee to pay. You can read more  |
    * | at:                                                                          |
    * |                  http://www.phpguru.org/static/license.html                  |
    * o------------------------------------------------------------------------------o
    *
    * © Copyright 2008 Richard Heyes
    */

    
    /**
    * PHP5 Implementation of the Porter Stemmer algorithm. Certain elements
    * were borrowed from the (broken) implementation by Jon Abernathy.
    * 
    * Usage:
    * 
    *  $stem = PorterStemmer::Stem($word);
    * 
    * How easy is that?
    */
    
    
class PorterStemmer
    
{
        
/**
        * Regex for matching a consonant
        * @var string
        */
        
private static $regex_consonant '(?:[bcdfghjklmnpqrstvwxz]|(?<=[aeiou])y|^y)';
    
    
        
/**
        * Regex for matching a vowel
        * @var string
        */
        
private static $regex_vowel '(?:[aeiou]|(?<![aeiou])y)';


        
/**
        * Cache for mass stemmings. Do not use for mass stemmings of *different* words. Use for
        * stemming words from documents for example, where the same word may crop up multiple times.
        * @var array
        */
        
private static $cache = array();


        
/**
        * Stems a word. Simple huh?
        * 
        * @param  string $word  Word to stem
        * @param  bool   $cache Whether to use caching
        * @return string        Stemmed word
        */
        
public static function Stem($word$cache false)
        {
            if (
strlen($word) <= 2) {
                return 
$word;
            }

            
// Check cache
            
if ($cache AND !empty(self::$cache[$word])) {
                return 
self::$cache[$word];
            }
            
            
/**
            * Remove: 've, n't, 'd
            */
            
$word preg_replace("/('ve|n't|'d)$/"''$word);

            
$stem self::step1ab($word);
            
$stem self::step1c($stem);
            
$stem self::step2($stem);
            
$stem self::step3($stem);
            
$stem self::step4($stem);
            
$stem self::step5($stem);

            
// Store in cache
            
if ($cache) {
                
self::$cache[$word] = $stem;
            }

            return 
$stem;
        }
    
    
        
/**
        * Step 1
        */
        
private static function step1ab($word)
        {
            
// Part a
            
if (substr($word, -1) == 's') {

                   
self::replace($word'sses''ss')
                OR 
self::replace($word'ies''i')
                OR 
self::replace($word'ss''ss')
                OR 
self::replace($word's''');
            }

            
// Part b
            
if (substr($word, -21) != 'e' OR !self::replace($word'eed''ee'0)) { // First rule
                
$v self::$regex_vowel;

                
// ing and ed
                
if (   preg_match("#$v+#"substr($word0, -3)) && self::replace($word'ing''')
                    OR 
preg_match("#$v+#"substr($word0, -2)) && self::replace($word'ed''')) { // Note use of && and OR, for precedence reasons

                    // If one of above two test successful
                    
if (    !self::replace($word'at''ate')
                        AND !
self::replace($word'bl''ble')
                        AND !
self::replace($word'iz''ize')) {

                        
// Double consonant ending
                        
if (    self::doubleConsonant($word)
                            AND 
substr($word, -2) != 'll'
                            
AND substr($word, -2) != 'ss'
                            
AND substr($word, -2) != 'zz') {
                            
                            
$word substr($word0, -1);
                        
                        } else if (
self::m($word) == AND self::cvc($word)) {
                            
$word .= 'e';
                        }
                    }
                }
            }

            return 
$word;
        }
    
    
        
/**
        * Step 1c
        * 
        * @param string $word Word to stem
        */
        
private static function step1c($word)
        {
            
$v self::$regex_vowel;

            if (
substr($word, -1) == 'y' && preg_match("#$v+#"substr($word0, -1))) {
                
self::replace($word'y''i');
            }

            return 
$word;
        }


        
/**
        * Step 2
        * 
        * @param string $word Word to stem
        */
        
private static function step2($word)
        {
            switch (
substr($word, -21)) {
                case 
'a':
                       
self::replace($word'ational''ate'0)
                    OR 
self::replace($word'tional''tion'0);
                    break;

                case 
'c':
                       
self::replace($word'enci''ence'0)
                    OR 
self::replace($word'anci''ance'0);
                    break;

                case 
'e':
                    
self::replace($word'izer''ize'0);
                    break;

                case 
'g':
                    
self::replace($word'logi''log'0);
                    break;

                case 
'l':
                       
self::replace($word'entli''ent'0)
                    OR 
self::replace($word'ousli''ous'0)
                    OR 
self::replace($word'alli''al'0)
                    OR 
self::replace($word'bli''ble'0)
                    OR 
self::replace($word'eli''e'0);
                    break;

                case 
'o':
                       
self::replace($word'ization''ize'0)
                    OR 
self::replace($word'ation''ate'0)
                    OR 
self::replace($word'ator''ate'0);
                    break;

                case 
's':
                       
self::replace($word'iveness''ive'0)
                    OR 
self::replace($word'fulness''ful'0)
                    OR 
self::replace($word'ousness''ous'0)
                    OR 
self::replace($word'alism''al'0);
                    break;

                case 
't':
                       
self::replace($word'biliti''ble'0)
                    OR 
self::replace($word'aliti''al'0)
                    OR 
self::replace($word'iviti''ive'0);
                    break;
            }

            return 
$word;
        }

    
        
/**
        * Step 3
        * 
        * @param string $word String to stem
        */
        
private static function step3($word)
        {
            switch (
substr($word, -21)) {
                case 
'a':
                    
self::replace($word'ical''ic'0);
                    break;
                    
                case 
's':
                       
self::replace($word'alise''al'0)
                    OR 
self::replace($word'ness'''0);
                    break;
                    
                case 
't':
                       
self::replace($word'icate''ic'0)
                    OR 
self::replace($word'iciti''ic'0);
                    break;
                    
                case 
'u':
                    
self::replace($word'ful'''0);
                    break;
                    
                case 
'v':
                    
self::replace($word'ative'''0);
                    break;
                    
                case 
'z':
                    
self::replace($word'alize''al'0);
                    break;
            }
            
            return 
$word;
        }
    
    
        
/**
        * Step 4
        * 
        * @param string $word Word to stem
        */
        
private static function step4($word)
        {
            switch (
substr($word, -21)) {
                case 
'a':
                    
self::replace($word'al'''1);
                    break;

                case 
'c':
                       
self::replace($word'ance'''1)
                    OR 
self::replace($word'ence'''1);
                    break;

                case 
'e':
                    
self::replace($word'er'''1);
                    break;

                case 
'i':
                    
self::replace($word'ic'''1);
                    break;

                case 
'l':
                       
self::replace($word'able'''1)
                    OR 
self::replace($word'ible'''1);
                    break;

                case 
'n':
                       
self::replace($word'ant'''1)
                    OR 
self::replace($word'ement'''1)
                    OR 
self::replace($word'ment'''1)
                    OR 
self::replace($word'ent'''1);
                    break;

                case 
'o':
                    if (
substr($word, -4) == 'tion' OR substr($word, -4) == 'sion') {
                       
self::replace($word'ion'''1);
                    } else {
                        
self::replace($word'ou'''1);
                    }
                    break;

                case 
's':
                    
self::replace($word'ism'''1);
                    break;

                case 
't':
                       
self::replace($word'ate'''1)
                    OR 
self::replace($word'iti'''1);
                    break;

                case 
'u':
                    
self::replace($word'ous'''1);
                    break;

                case 
'v':
                    
self::replace($word'ive'''1);
                    break;

                case 
'z':
                    
self::replace($word'ize'''1);
                    break;
            }
            
            return 
$word;
        }

    
        
/**
        * Step 5
        * 
        * @param string $word Word to stem
        */
        
private static function step5($word)
        {
            
// Part a
            
if (substr($word, -1) == 'e') {
                if (
self::m(substr($word0, -1)) > 1) {
                    
self::replace($word'e''');
    
                } else if (
self::m(substr($word0, -1)) == 1) {

                    if (!
self::cvc(substr($word0, -1))) {
                        
self::replace($word'e''');
                    }
                }
            }

            
// Part b
            
if (self::m($word) > AND self::doubleConsonant($word) AND substr($word, -1) == 'l') {
                
$word substr($word0, -1);
            }

            return 
$word;
        }
    
    
        
/**
        * Replaces the first string with the second, at the end of the string. If third
        * arg is given, then the preceding string must match that m count at least.
        * 
        * @param  string $str   String to check
        * @param  string $check Ending to check for
        * @param  string $repl  Replacement string
        * @param  int    $m     Optional minimum number of m() to meet
        * @return bool          Whether the $check string was at the end
        *                       of the $str string. True does not necessarily mean
        *                       that it was replaced.
        */
        
private static function replace(&$str$check$repl$m null)
        {
            
$len strlen($check);
    
            if (
substr($str$len) == $check) {
                
$substr substr($str0$len);
                if (
is_null($m) OR self::m($substr) > $m) {
                    
$str $substr $repl;
                }

                return 
true;
            }
    
            return 
false;
        }
    
    
        
/**
        * What, you mean it's not obvious from the name?
        * 
        * m() measures the number of consonant sequences in $str. if c is
        * a consonant sequence and v a vowel sequence, and <..> indicates arbitrary
        * presence,
        * 
        * <c><v>       gives 0
        * <c>vc<v>     gives 1
        * <c>vcvc<v>   gives 2
        * <c>vcvcvc<v> gives 3
        * 
        * @param  string $str The string to return the m count for
        * @return int         The m count
        */
        
private static function m($str)
        {
            
$c self::$regex_consonant;
            
$v self::$regex_vowel;

            
$str preg_replace("#^$c+#"''$str);
            
$str preg_replace("#$v+$#"''$str);
    
            
preg_match_all("#($v+$c+)#"$str$matches);

            return 
count($matches[1]);
        }

    
        
/**
        * Returns true/false as to whether the given string contains two
        * of the same consonant next to each other at the end of the string.
        * 
        * @param  string $str String to check
        * @return bool        Result
        */
        
private static function doubleConsonant($str)
        {
            
$c self::$regex_consonant;
            
            return 
preg_match("#$c{2}$#"$str$matches) AND $matches[0]{0} == $matches[0]{1};
        }


        
/**
        * Checks for ending CVC sequence where second C is not W, X or Y
        * 
        * @param  string $str String to check
        * @return bool        Result
        */
        
private static function cvc($str)
        {
            
$c self::$regex_consonant;
            
$v self::$regex_vowel;

            return     
preg_match("#($c$v$c)$#"$str$matches)
                   AND 
strlen($matches[1]) == 3
                   
AND $matches[1]{2} != 'w'
                   
AND $matches[1]{2} != 'x'
                   
AND $matches[1]{2} != 'y';
        }
    }
?>