Your IP : 216.73.216.54
<?php
/**
* ArticleMetaDescription - adds meta-description tag containing snippet of the article
*
* It takes first paragraph of the article and put into <meta description="..." /> tag inside
* page header section. All templates (infoboxes, tables etc.) are ignored by default. It's
* possible to set predefined description for main page (configured in MediaWiki:Mainpage) by
* putting desired text into MediaWiki:Description message.
*
* @file
* @ingroup Extensions
* @version 1.3.1
* @date 2 December 2019
* @author Adrian 'ADi' Wieczorek <adi@wikia-inc.com>
* @link https://www.mediawiki.org/wiki/Extension:ArticleMetaDescription Documentation
* @license GPL-2.0-or-later
*/
class ArticleMetaDescription {
/**
* @param OutputPage &$out
* @param string &$text
*/
public static function onOutputPageBeforeHTML( &$out, &$text ) {
$message = null;
$titleObj = $out->getTitle();
if ( $titleObj->isMainPage() ) {
// we're on Main Page, check MediaWiki:Description message
$message = wfMessage( 'description' )->inContentLanguage()->text();
}
if ( ( $message == null ) || wfMessage( 'description' )->inContentLanguage()->isDisabled() ) {
// ashley 6 February 2023: filthy but potentially functional hack: first and foremost get rid of the initial
// <div class="mw-parser-output"> so that the "remove all divs" regex below doesn't mutilate
// the intro section
$tmp = str_replace( '<div class="mw-parser-output">', '', $text );
$tmp = preg_replace( '/<table[^>]*>.*<\/table>/siU', '', $tmp );
$tmp = preg_replace( '/<div[^>]*>.*<\/div>/siU', '', $tmp );
$tmp = preg_replace( '/<style[^>]*>.*<\/style>/siU', '', $tmp );
$tmp = preg_replace( '/<script[^>]*>.*<\/script>/siU', '', $tmp );
$tmp = preg_replace( '/\n|\t/', ' ', $tmp );
// @todo FIXME: per discussion with bawolff in December 2019, this should
// use Sanitizer::stripAllTags instead
$tmp = strip_tags( $tmp, '<p>' );
$matches = null;
preg_match_all( '/<p>(.*)<\/p>/siU', $tmp, $matches );
if ( count( $matches ) ) {
$paragraphs = $matches[1];
$description = '';
foreach ( $paragraphs as $paragraph ) {
$paragraph = trim( $paragraph );
if ( !empty( $paragraph ) ) {
$description = $paragraph;
break;
}
}
if ( !empty( $description ) ) {
// FILTHY HACK - clean up Cite citation numbers from the output
// e.g. if our $description looks like this before:
// 'This is a sentence.[9]', the numbers would then show up in Google
// search results, which we obviously don't want. So just regex it out...
// @todo Should probably be active only when Cite is? (A fair amount of
// wikis have Cite installed, though, which is why no such check exists
// here currently.
$description = preg_replace(
// 1st part matches numbers like [1]
// 2nd part matches HTML-encoded versions, like &#91;1&#93;
// 3rd is the one that actually works (thanks, bawolff!)
// <bawolff> ashley: maybe your regex is happening before double encoding, but after single
'/(\[\d{0,}\]|\&\#91;\d{0,}\&\#93\;|\&\#91;\d*\&\#93;)/',
'',
$description
);
}
}
} else {
// MediaWiki:Description message found, use it
$description = $message;
}
if ( !empty( $description ) ) {
$out->addMeta( 'description', $description );
}
}
}