[4.0] Upgrade php-stemmer 2.0 (#27568)

Refactoring stemmer integration by reducing the seperated classes and
using the stemmer factory to create a stemmer object.
This reduces maintenance because no change in Joomla! is required to
support new languages added by the upstream php-stemmer class
and we reduce complexity in Joomla! code base.

administrator/components/com_finder/Indexer/Language.php

@@ -12,6 +12,9 @@
 defined('_JEXEC') or die;
 
 use Joomla\String\StringHelper;
+use Wamania\Snowball\NotFoundException;
+use Wamania\Snowball\StemmerFactory;
+use Wamania\Snowball\Stemmer\Stemmer;
 
 /**
  * Language support class for the Finder indexer package.
@@ -44,6 +47,42 @@ class Language
 	 */
 	public $spacer = ' ';
 
+	/**
+	 * The stemmer object.
+	 *
+	 * @var    Stemmer
+	 * @since  4.0.0
+	 */
+	protected $stemmer = null;
+
+	/**
+	 * Method to construct the language object.
+	 *
+	 * @since   4.0.0
+	 */
+	public function __construct($locale = null)
+	{
+		if ($locale !== null)
+		{
+			$this->language = $locale;
+		}
+
+		// Use our generic language handler if no language is set
+		if ($this->language === null)
+		{
+			$this->language = '*';
+		}
+
+		try
+		{
+			$this->stemmer = StemmerFactory::create($this->language);
+		}
+		catch (NotFoundException $e)
+		{
+			// We don't have a stemmer for the language
+		}
+	}
+
 	/**
 	 * Method to get a language support object.
 	 *
@@ -60,24 +99,21 @@ public static function getInstance($language)
 			return self::$instances[$language];
 		}
 
-		if ($language == '*')
-		{
-			self::$instances[$language] = new self;
+		$locale = '*';
 
-			return self::$instances[$language];
+		if ($language !== '*')
+		{
+			$locale = Helper::getPrimaryLanguage($language);
+			$class = '\\Joomla\\Component\\Finder\\Administrator\\Indexer\\Language\\' . ucfirst($locale);
 		}
 
-		$locale = Helper::getPrimaryLanguage($language);
-		$class = '\\Joomla\\Component\\Finder\\Administrator\\Indexer\\Language\\' . ucfirst($locale);
-
 		if (class_exists($class))
 		{
 			self::$instances[$language] = new $class;
 		}
 		else
 		{
-			self::$instances[$language] = new self;
-			self::$instances[$language]->language = $locale;
+			self::$instances[$language] = new self($locale);
 		}
 
 		return self::$instances[$language];
@@ -137,6 +173,11 @@ public function tokenise($input)
 	 */
 	public function stem($token)
 	{
+		if ($this->stemmer !== null)
+		{
+			return $this->stemmer->stem($token);
+		}
+
 		return $token;
 	}
 }

administrator/components/com_finder/Indexer/Language/El.php

@@ -32,13 +32,23 @@ class El extends Language
 	 */
 	public $language = 'el';
 
+	/**
+	 * Method to construct the language object.
+	 *
+	 * @since   4.0.0
+	 */
+	public function __construct($locale = null)
+	{
+		// Override parent constructor since we don't need to load an externel stemmer
+	}
+
 	/**
 	 * Method to tokenise a text string. It takes into account the odd punctuation commonly used in Greek text, mapping
 	 * it to ASCII punctuation.
 	 *
 	 * Reference: http://www.teicrete.gr/users/kutrulis/Glosika/Stixi.htm
 	 *
-	 * @param   string  $input  The input to tokenise.
+	 * @param string $input The input to tokenise.
 	 *
 	 * @return  array  An array of term strings.
 	 *