Your IP : 216.73.216.54
<?php
namespace MediaWiki\OutputTransform;
use MediaWiki\Parser\ParserOptions;
use MediaWiki\Parser\ParserOutput;
use MediaWiki\Parser\Parsoid\PageBundleParserOutputConverter;
use stdClass;
use Wikimedia\Parsoid\Core\PageBundle;
use Wikimedia\Parsoid\DOM\Document;
use Wikimedia\Parsoid\DOM\Element;
use Wikimedia\Parsoid\Mocks\MockEnv;
use Wikimedia\Parsoid\Utils\ContentUtils;
use Wikimedia\Parsoid\Utils\DOMCompat;
use Wikimedia\Parsoid\Utils\DOMDataUtils;
use Wikimedia\Parsoid\Utils\DOMUtils;
use Wikimedia\Parsoid\Utils\PHPUtils;
use Wikimedia\Parsoid\Utils\Utils;
/**
* OutputTransformStages that modify the content as a HTML DOM tree.
*
* Subclasses are expected to implement ::transformDOM() to mutate the
* DOM-structured content as a Document in-place.
*
* @internal
*/
abstract class ContentDOMTransformStage extends OutputTransformStage {
/**
* @inheritDoc
*/
public function transform(
ParserOutput $po, ?ParserOptions $popts, array &$options
): ParserOutput {
if ( $options['isParsoidContent'] ?? false ) {
return $this->parsoidTransform( $po, $popts, $options );
} else {
return $this->legacyTransform( $po, $popts, $options );
}
}
private function legacyTransform(
ParserOutput $po, ?ParserOptions $popts, array &$options
): ParserOutput {
$text = $po->getContentHolderText();
$doc = DOMUtils::parseHTML( $text );
$doc = $this->transformDOM( $doc, $po, $popts, $options );
$body = DOMCompat::getBody( $doc );
$text = ContentUtils::toXML( $body, [
'innerXML' => true,
] );
$po->setContentHolderText( $text );
return $po;
}
private function parsoidTransform(
ParserOutput $po, ?ParserOptions $popts, array &$options
): ParserOutput {
// TODO will use HTMLHolder in the future
$doc = null;
$hasPageBundle = PageBundleParserOutputConverter::hasPageBundle( $po );
$origPb = null;
if ( $hasPageBundle ) {
$origPb = PageBundleParserOutputConverter::pageBundleFromParserOutput( $po );
$doc = DOMUtils::parseHTML( $po->getContentHolderText() );
PageBundle::apply( $doc, $origPb );
DOMDataUtils::prepareDoc( $doc );
$body = DOMCompat::getBody( $doc );
'@phan-var Element $body'; // assert non-null
DOMDataUtils::visitAndLoadDataAttribs( $body );
} else {
$doc = ContentUtils::createAndLoadDocument(
$po->getContentHolderText(),
);
}
$doc = $this->transformDOM( $doc, $po, $popts, $options );
// TODO will use HTMLHolder in the future
$body = DOMCompat::getBody( $doc );
'@phan-var Element $body'; // assert non-null
if ( $hasPageBundle ) {
DOMDataUtils::visitAndStoreDataAttribs(
$body,
[
'storeInPageBundle' => true,
'env' => new MockEnv( [] ),
]
);
$pb = DOMDataUtils::getPageBundle( $doc );
$pb = self::workaroundT365036( $pb );
$pb = self::combinePageBundles( $pb, $origPb );
PageBundleParserOutputConverter::applyPageBundleDataToParserOutput( $pb, $po );
$text = ContentUtils::toXML( $body, [
'innerXML' => true,
] );
} else {
$text = ContentUtils::ppToXML( $body, [
'innerXML' => true,
] );
}
$po->setContentHolderText( $text );
return $po;
}
private function combinePageBundles( stdClass $pb, ?PageBundle $origPb ): stdClass {
if ( $origPb === null ) {
return $pb;
}
$headers = $origPb->headers;
$pb->headers = $headers ? Utils::clone( $headers ) : null;
$pb->version = $origPb->version;
$pb->contentmodel = $origPb->contentmodel;
return $pb;
}
/**
* This function is necessary (T365036) because PageBundle can have
* non-serializable DataParsoid inside which need
* to be converted to stdClass and string, respectively. We should replace this
* ugly workaround with something nicer.
*/
private function workaroundT365036( PageBundle $pb ): stdClass {
$arrayPageBundle = PHPUtils::jsonDecode( PHPUtils::jsonEncode( $pb ) );
return (object)$arrayPageBundle;
}
/** Applies the transformation to a DOM document */
abstract public function transformDOM(
Document $dom, ParserOutput $po, ?ParserOptions $popts, array &$options
): Document;
}