PortableInfobox/includes/services/Parser/MediaWikiParserService.php
alex4401 eb55d89e77
Output any image candidate markers generated by PageImages (#105)
* Output any image candidate markers generated by PageImages

wikimedia/mediawiki-extensions-PageImages#768464d changed how lead images are identified within PageImages, and since then an HTML comment in the form of `MW-PAGEIMAGES-CANDIDATE-$ID` is output by`doParserModifyImageHTML`, where the ID is an index in the extension output data's list of image metadata. These comments should find a way into the ParserOutput as the extension looks for those comments in a chosen document fragment when identifying qualifying candidates out of all images gathered through `onParserModifyImageHTML`.

This change does not affect compatibility with older PageImages (only the "new" hook's being called already).

* Default htmlAfter to null and update tests to expect it

Follow-up to my previous PageImages commit.
2023-07-24 11:13:33 -06:00

125 lines
3.2 KiB
PHP

<?php
namespace PortableInfobox\Parser;
use BlockLevelPass;
use MediaWiki\MediaWikiServices;
use MediaWiki\Tidy\RemexDriver;
use PageImages\Hooks\ParserFileProcessingHookHandlers;
use Parser;
use PPFrame;
use Title;
class MediaWikiParserService implements ExternalParser {
protected $parser;
protected $frame;
protected $localParser;
protected $tidyDriver;
protected $cache = [];
public function __construct( Parser $parser, PPFrame $frame ) {
global $wgPortableInfoboxUseTidy;
$this->parser = $parser;
$this->frame = $frame;
if ( $wgPortableInfoboxUseTidy && class_exists( RemexDriver::class ) ) {
global $wgTidyConfig;
$wgTidyConfig = [
'driver' => 'RemexHtml',
'pwrap' => false
];
$this->tidyDriver = MediaWikiServices::getInstance()->getTidy();
}
}
/**
* Method used for parsing wikitext provided in infobox that might contain variables
*
* @param string $wikitext
*
* @return string HTML outcome
*/
public function parseRecursive( $wikitext ) {
if ( isset( $this->cache[$wikitext] ) ) {
return $this->cache[$wikitext];
}
$parsed = $this->parser->recursiveTagParse( $wikitext ?? '', $this->frame );
if ( in_array( substr( $parsed, 0, 1 ), [ '*', '#' ] ) ) {
// fix for first item list elements
$parsed = "\n" . $parsed;
}
// @phan-suppress-next-line PhanAccessMethodInternal
$output = BlockLevelPass::doBlockLevels( $parsed, false );
$ready = $this->parser->getStripState()->unstripBoth( $output );
// @phan-suppress-next-line PhanDeprecatedFunction
$this->parser->replaceLinkHolders( $ready );
if ( isset( $this->tidyDriver ) ) {
$ready = $this->tidyDriver->tidy( $ready );
}
$newlinesstripped = preg_replace( '|[\n\r]|Us', '', $ready );
$marksstripped = preg_replace( '|{{{.*}}}|Us', '', $newlinesstripped );
$this->cache[$wikitext] = $marksstripped;
return $marksstripped;
}
public function replaceVariables( $wikitext ) {
$output = $this->parser->replaceVariables( $wikitext, $this->frame );
return $output;
}
/**
* Add image to parser output for later usage
*
* @param Title $title
* @return ?string PageImages markers, if any.
*/
public function addImage( $title ): ?string {
$services = MediaWikiServices::getInstance();
$repoGroup = $services->getRepoGroup();
$file = $repoGroup->findFile( $title );
$tmstmp = $file ? $file->getTimestamp() : null;
$sha1 = $file ? $file->getSha1() : null;
$this->parser->getOutput()->addImage( $title->getDBkey(), $tmstmp, $sha1 );
// Pass PI images to PageImages extension if available (Popups and og:image). Since 1.38, this produces an HTML
// comment that must be present in the rendered HTML for the image to qualify for selection.
if ( method_exists(
ParserFileProcessingHookHandlers::class, 'onParserModifyImageHTML'
) ) {
// @phan-suppress-next-line PhanParamTooMany
$handler = new ParserFileProcessingHookHandlers(
$repoGroup,
$services->getMainWANObjectCache(),
$services->getHttpRequestFactory(),
$services->getDBLoadBalancerFactory(),
$services->getTitleFactory()
);
$params = [];
$html = '';
$handler->onParserModifyImageHTML(
$this->parser, $file, $params, $html
);
return $html;
}
return null;
}
}