/**
* Takes a list of words and returns them reduced to their stems.
*
* $words can be either a string or an array. If it is a string, it will
* be split into separate words on whitespace, commas, or semicolons. If
* an array, it assumes one word per element.
*
* @param mixed $words
* String or array of word(s) to reduce
* @param bool $commonwords
* Remove common words prior to stemming
* @access public
* @return array List of word stems
*/
public function getStemmed($words, $commonwords = false)
{
if ($commonwords === true) {
if (is_array($words)) {
$words = implode(" ", $words);
$words = \Rss\Text\CommonWords\CommonWords::removeCommonWords($words);
$words = preg_replace('/\\s{2,}/', ' ', $words);
$words = explode(" ", $words);
}
if (is_string($words)) {
$words = \Rss\Text\CommonWords\CommonWords::removeCommonWords($words);
$words = preg_replace('/\\s{2,}/', ' ', $words);
}
}
$stemmer = new \Rss\Util\Contrib\Nlp\Stemmers\en\Stemmer\Stemmer();
return $stemmer->stem_list($words);
}
/**
*
* @param array $item
* @param bool $extinfo
*/
private function process_item($item, $extinfo)
{
$item['title'] = $this->get_title($item['title']);
$title = $item['title'];
$item['description'] = $this->get_text($item['description']);
$description = $item['description'];
$item['content'] = $this->get_text($item['content']);
$content = $item['content'];
$item['link'] = $this->get_link($item['link']);
$link = $item['link'];
if (empty($link)) {
$this->logger->logWarn("Empty item url ({$link})");
return null;
}
$zuri = \Zend\Uri\UriFactory::factory($link);
if (!$zuri->isValid()) {
$this->logger->logWarn("Invalid item url ({$link})");
return null;
}
$linkparts = parse_url($link);
if (!isset($linkparts['host']) || empty($linkparts['host'])) {
$this->logger->logWarn("Invalid item host ({$link})");
return null;
}
if ($extinfo) {
// Ext Title
$titlenocommon = \Rss\Text\CommonWords\CommonWords::removeCommonWords($title);
$stemmer = new \Rss\Text\Stemmer\PorterStemmer();
$titlestemmed = $stemmer->getStemmed($titlenocommon);
// Strip multiply spaces and remove line breaks
$titlestemmed = preg_replace('/\\s+/m', ' ', $titlestemmed);
$misc = new \Rss\Text\Misc\Misc();
$titlemetaphone = $misc->getMetaphone($titlestemmed);
$item['titlenocommon'] = $titlenocommon;
$item['titlestemmed'] = $titlestemmed;
$item['titlemetaphone'] = $titlemetaphone;
}
ksort($item);
return $item;
}
