Cite: Remove more Parsoid internals knowledge

* Remove use of $env from ReferencesData and RefGroup by
  providing high-level helpers in ParsoidExtensionAPI.

  - Given a fragment id, provide helpers to fetch fragment DOM
    or fragment HTML
  - Fetch the URI for the current page (being parsed)

* There is still a lot of subtle knowledge Cite has about
  how data-parsoid and data-mw attributes are held off to the
  side in a bag and all the pp* and load/store manipulation
  of those attributes. It would be an interesting exercise
  to purge this implementation of those notions OR figure out
  high-level concepts that we document as being part of Parsoid
  reality that we'll forever support.

Bug: T242746
Change-Id: I29ff154f2f17123b9756dfd2f3b422f0b30222b1

src/Parsoid/RefGroup.php

@@ -6,10 +6,9 @@ namespace Wikimedia\Parsoid\Ext\Cite;
 use DOMDocument;
 use DOMElement;
 use stdClass;
-use Wikimedia\Parsoid\Config\Env;
+use Wikimedia\Parsoid\Config\ParsoidExtensionAPI;
 use Wikimedia\Parsoid\Utils\DOMDataUtils;
 use Wikimedia\Parsoid\Utils\DOMUtils;
-use Wikimedia\Parsoid\Utils\Title;
 
 /**
  * Helper class used by `<references>` implementation.
@@ -43,24 +42,21 @@ class RefGroup {
 
 	/**
 	 * Generate leading linkbacks
+	 * @param ParsoidExtensionAPI $extApi
 	 * @param string $href
 	 * @param string|null $group
 	 * @param string $text
 	 * @param DOMDocument $ownerDoc
-	 * @param Env $env
 	 * @return DOMElement
 	 */
 	private static function createLinkback(
-		string $href, ?string $group, string $text, DOMDocument $ownerDoc, Env $env
+		ParsoidExtensionAPI $extApi,
+		string $href, ?string $group, string $text, DOMDocument $ownerDoc
 	): DOMElement {
 		$a = $ownerDoc->createElement( 'a' );
 		$s = $ownerDoc->createElement( 'span' );
 		$textNode = $ownerDoc->createTextNode( $text . ' ' );
-		$title = Title::newFromText(
-			$env->getPageConfig()->getTitle(),
-			$env->getSiteConfig()
-		);
-		$a->setAttribute( 'href', $env->makeLink( $title ) . '#' . $href );
+		$a->setAttribute( 'href', $extApi->getPageUri() . '#' . $href );
 		$s->setAttribute( 'class', 'mw-linkback-text' );
 		if ( $group ) {
 			$a->setAttribute( 'data-mw-group', $group );
@@ -71,11 +67,13 @@ class RefGroup {
 	}
 
 	/**
-	 * @param Env $env
+	 * @param ParsoidExtensionAPI $extApi
 	 * @param DOMElement $refsList
 	 * @param stdClass $ref
 	 */
-	public function renderLine( Env $env, DOMElement $refsList, stdClass $ref ): void {
+	public function renderLine(
+		ParsoidExtensionAPI $extApi, DOMElement $refsList, stdClass $ref
+	): void {
 		$ownerDoc = $refsList->ownerDocument;
 
 		// Generate the li and set ref content first, so the HTML gets parsed.
@@ -83,7 +81,7 @@ class RefGroup {
 		$li = $ownerDoc->createElement( 'li' );
 		$refDir = $ref->dir;
 		$refTarget = $ref->target;
-		$refContent = $ref->content;
+		$refContentId = $ref->contentId;
 		$refGroup = $ref->group;
 		DOMDataUtils::addAttributes( $li, [
 				'about' => '#' . $refTarget,
@@ -99,15 +97,15 @@ class RefGroup {
 				'class' => 'mw-reference-text',
 			]
 		);
-		if ( $refContent ) {
-			$content = $env->getFragment( $refContent )[0];
+		if ( $refContentId ) {
+			$content = $extApi->getContentDOM( $refContentId );
 			DOMUtils::migrateChildrenBetweenDocs( $content, $reftextSpan );
 			DOMDataUtils::visitAndLoadDataAttribs( $reftextSpan );
 		}
 		$li->appendChild( $reftextSpan );
 
 		if ( count( $ref->linkbacks ) === 1 ) {
-			$linkback = self::createLinkback( $ref->id, $refGroup, "↑", $ownerDoc, $env );
+			$linkback = self::createLinkback( $extApi, $ref->id, $refGroup, "↑", $ownerDoc );
 			$linkback->setAttribute( 'rel', 'mw:referencedBy' );
 			$li->insertBefore( $linkback, $reftextSpan );
 		} else {
@@ -118,7 +116,7 @@ class RefGroup {
 
 			foreach ( $ref->linkbacks as $i => $lb ) {
 				$span->appendChild(
-					self::createLinkback( $lb, $refGroup, (string)( $i + 1 ), $ownerDoc, $env )
+					self::createLinkback( $extApi, $lb, $refGroup, (string)( $i + 1 ), $ownerDoc )
 				);
 			}
 		}

src/Parsoid/RefProcessor.php

@@ -4,7 +4,6 @@ declare( strict_types = 1 );
 namespace Wikimedia\Parsoid\Ext\Cite;
 
 use DOMElement;
-use Wikimedia\Parsoid\Config\Env;
 use Wikimedia\Parsoid\Config\ParsoidExtensionAPI;
 
 /**
@@ -23,17 +22,17 @@ class RefProcessor {
 
 	/**
 	 * @param DOMElement $body
-	 * @param Env $env
+	 * @param mixed $unused unused Env object FIXME: stop passing this through?
 	 * @param array $options
 	 * @param bool $atTopLevel
 	 */
 	public function run(
-		DOMElement $body, Env $env, array $options = [], bool $atTopLevel = false
+		DOMElement $body, $unused, array $options = [], bool $atTopLevel = false
 	): void {
 		if ( $atTopLevel ) {
-			$refsData = new ReferencesData( $env );
+			$refsData = new ReferencesData();
 			References::processRefs( $this->extApi, $refsData, $body );
-			References::insertMissingReferencesIntoDOM( $refsData, $body );
+			References::insertMissingReferencesIntoDOM( $this->extApi, $refsData, $body );
 		}
 	}
 }

src/Parsoid/References.php

@@ -104,7 +104,7 @@ class References extends ExtensionTag {
 		DOMElement $node, ReferencesData $refsData, ?string $referencesAboutId = null,
 		?string $referencesGroup = '', array &$nestedRefsHTML = []
 	): void {
-		$env = $refsData->getEnv();
+		$env = $extApi->getEnv();
 		$doc = $node->ownerDocument;
 		$nestedInReferences = $referencesAboutId !== null;
 
@@ -114,12 +114,12 @@ class References extends ExtensionTag {
 		$typeOf = $node->getAttribute( 'typeof' );
 		$isTplWrapper = DOMUtils::matchTypeOf( $node, '/^mw:Transclusion$/' );
 		$nodeType = preg_replace( '#mw:DOMFragment/sealed/ref#', '', $typeOf, 1 );
-		$content = $nodeDp->html;
+		$contentId = $nodeDp->html;
 		$tplDmw = $isTplWrapper ? DOMDataUtils::getDataMw( $node ) : null;
 
 		// This is the <sup> that's the meat of the sealed fragment
 		/** @var DOMElement $c */
-		$c = $env->getFragment( $content )[0];
+		$c = $extApi->getContentDOM( $contentId );
 		DOMUtils::assertElt( $c );
 		// All the actions that require loaded data-attributes on `c` are done
 		// here so that we can quickly store those away for later.
@@ -142,7 +142,7 @@ class References extends ExtensionTag {
 		// elt has a group attribute, what takes precedence?
 		$group = $refDmw->attrs->group ?? $referencesGroup ?? '';
 		$refName = $refDmw->attrs->name ?? '';
-		$ref = $refsData->add( $env, $group, $refName, $about, $nestedInReferences );
+		$ref = $refsData->add( $extApi, $group, $refName, $about, $nestedInReferences );
 
 		// Add ref-index linkback
 		$linkBack = $doc->createElement( 'sup' );
@@ -150,7 +150,7 @@ class References extends ExtensionTag {
 		// FIXME: Lot of useless work for an edge case
 		if ( !empty( $cDp->empty ) ) {
 			// Discard wrapper if there was no input wikitext
-			$content = null;
+			$contentId = null;
 			if ( !empty( $cDp->selfClose ) ) {
 				unset( $refDmw->body );
 			} else {
@@ -234,27 +234,29 @@ class References extends ExtensionTag {
 			$node->parentNode->replaceChild( $linkBack, $node );
 		} else {
 			// We don't need to delete the node now since it'll be removed in
-			// `insertReferencesIntoDOM` when all the children all cleaned out.
+			// `insertReferencesIntoDOM` when all the children are cleaned out.
 			array_push( $nestedRefsHTML, ContentUtils::ppToXML( $linkBack ), "\n" );
 		}
 
 		// Keep the first content to compare multiple <ref>s with the same name.
-		if ( !$ref->content ) {
-			$ref->content = $content;
+		if ( !$ref->contentId ) {
+			$ref->contentId = $contentId;
 			$ref->dir = strtolower( $refDmw->attrs->dir ?? '' );
 		}
 	}
 
 	/**
+	 * @param ParsoidExtensionAPI $extApi
 	 * @param DOMElement $refsNode
 	 * @param ReferencesData $refsData
 	 * @param array $nestedRefsHTML
 	 * @param bool $autoGenerated
 	 */
 	private static function insertReferencesIntoDOM(
-		DOMElement $refsNode, ReferencesData $refsData, array $nestedRefsHTML, bool $autoGenerated = false
+		ParsoidExtensionAPI $extApi, DOMElement $refsNode,
+		ReferencesData $refsData, array $nestedRefsHTML, bool $autoGenerated = false
 	): void {
-		$env = $refsData->getEnv();
+		$env = $extApi->getEnv();
 		$isTplWrapper = DOMUtils::matchTypeOf( $refsNode, '/^mw:Transclusion$/' );
 		$dp = DOMDataUtils::getDataParsoid( $refsNode );
 		$group = $dp->group ?? '';
@@ -311,7 +313,7 @@ class References extends ExtensionTag {
 
 		if ( $refGroup ) {
 			foreach ( $refGroup->refs as $ref ) {
-				$refGroup->renderLine( $env, $refsNode, $ref );
+				$refGroup->renderLine( $extApi, $refsNode, $ref );
 			}
 		}
 
@@ -324,13 +326,14 @@ class References extends ExtensionTag {
 	 * We process them as if there was an implicit `<references />` tag at
 	 * the end of the DOM.
 	 *
+	 * @param ParsoidExtensionAPI $extApi
 	 * @param ReferencesData $refsData
 	 * @param DOMNode $node
 	 */
 	public static function insertMissingReferencesIntoDOM(
-		ReferencesData $refsData, DOMNode $node
+		ParsoidExtensionAPI $extApi, ReferencesData $refsData, DOMNode $node
 	): void {
-		$env = $refsData->getEnv();
+		$env = $extApi->getEnv();
 		$doc = $node->ownerDocument;
 
 		foreach ( $refsData->getRefGroups() as $groupName => $refsGroup ) {
@@ -360,7 +363,7 @@ class References extends ExtensionTag {
 			$node->appendChild( $doc->createTextNode( "\n" ) );
 			$node->appendChild( $frag );
 
-			self::insertReferencesIntoDOM( $frag, $refsData, [ '' ], true );
+			self::insertReferencesIntoDOM( $extApi, $frag, $refsData, [ '' ], true );
 		}
 	}
 
@@ -398,7 +401,7 @@ class References extends ExtensionTag {
 						$referencesGroup,
 						$nestedRefsHTML
 					);
-					self::insertReferencesIntoDOM( $child, $refsData, $nestedRefsHTML );
+					self::insertReferencesIntoDOM( $extApi, $child, $refsData, $nestedRefsHTML );
 				} else {
 					// Look for <ref>s embedded in data attributes
 					$extApi->processHiddenHTMLInDataAttributes( $child,

src/Parsoid/ReferencesData.php

@@ -4,17 +4,11 @@ declare( strict_types = 1 );
 namespace Wikimedia\Parsoid\Ext\Cite;
 
 use stdClass;
-use Wikimedia\Parsoid\Config\Env;
-use Wikimedia\Parsoid\Utils\ContentUtils;
+use Wikimedia\Parsoid\Config\ParsoidExtensionAPI;
 use Wikimedia\Parsoid\Wt2Html\TT\Sanitizer;
 
 class ReferencesData {
 
-	/**
-	 * @var Env
-	 */
-	private $env;
-
 	/**
 	 * @var int
 	 */
@@ -27,11 +21,9 @@ class ReferencesData {
 
 	/**
 	 * ReferencesData constructor.
-	 * @param Env $env
 	 */
-	public function __construct( Env $env ) {
+	public function __construct() {
 		$this->index = 0;
-		$this->env = $env;
 		$this->refGroups = [];
 	}
 
@@ -73,7 +65,7 @@ class ReferencesData {
 	}
 
 	/**
-	 * @param Env $env
+	 * @param ParsoidExtensionAPI $extApi
 	 * @param string $groupName
 	 * @param string $refName
 	 * @param string $about
@@ -81,7 +73,7 @@ class ReferencesData {
 	 * @return stdClass
 	 */
 	public function add(
-		Env $env, string $groupName, string $refName, string $about, bool $skipLinkback
+		ParsoidExtensionAPI $extApi, string $groupName, string $refName, string $about, bool $skipLinkback
 	): stdClass {
 		$group = $this->getRefGroup( $groupName, true );
 		$refName = $this->makeValidIdAttr( $refName );
@@ -89,14 +81,11 @@ class ReferencesData {
 
 		if ( $hasRefName && isset( $group->indexByName[$refName] ) ) {
 			$ref = $group->indexByName[$refName];
-			if ( $ref->content && !$ref->hasMultiples ) {
+			if ( $ref->contentId && !$ref->hasMultiples ) {
 				$ref->hasMultiples = true;
 				// Use the non-pp version here since we've already stored attribs
 				// before putting them in the map.
-				$ref->cachedHtml = ContentUtils::toXML(
-					$env->getFragment( $ref->content )[0],
-					[ 'innerXML' => true ]
-				);
+				$ref->cachedHtml = $extApi->getContentHTML( $ref->contentId );
 			}
 		} else {
 			// The ids produced Cite.php have some particulars:
@@ -114,7 +103,7 @@ class ReferencesData {
 
 			$ref = (object)[
 				'about' => $about,
-				'content' => null,
+				'contentId' => null,
 				'dir' => '',
 				'group' => $group->name,
 				'groupIndex' => count( $group->refs ) + 1,
@@ -140,13 +129,6 @@ class ReferencesData {
 		return $ref;
 	}
 
-	/**
-	 * @return Env
-	 */
-	public function getEnv(): Env {
-		return $this->env;
-	}
-
 	/**
 	 * @return RefGroup[]
 	 */