feat: ParsoidModule implementation

extension.json

@@ -7,15 +7,19 @@
 	],
 	"url": "https://github.com/Universal-Omega/PortableInfobox",
 	"descriptionmsg": "portable-infobox-desc",
-	"version": "0.8",
+	"version": "1.0",
 	"type": "parserhook",
 	"license-name": "GPL-3.0-or-later",
 	"requires": {
 		"MediaWiki": ">= 1.43.0"
 	},
 	"config": {
 		"AllInfoboxesExcludedSubpages": {
-			"value": [ "doc", "draft", "test" ]
+			"value": [
+				"doc",
+				"draft",
+				"test"
+			]
 		},
 		"PortableInfoboxCacheRenderers": {
 			"value": false
@@ -125,5 +129,6 @@
 	"APIListModules": {
 		"allinfoboxes": "PortableInfobox\\Controllers\\ApiQueryAllInfoboxes"
 	},
+	"ParsoidModules": [ "PortableInfobox\\Parsoid\\InfoboxTag" ],
 	"manifest_version": 2
-}
+}

includes/Parsoid/InfoboxTag.php

@@ -0,0 +1,57 @@
+<?php
+
+namespace PortableInfobox\Parsoid;
+
+use Wikimedia\Parsoid\DOM\Element;
+use Wikimedia\Parsoid\Ext\DOMUtils;
+use Wikimedia\Parsoid\Ext\ExtensionModule;
+use Wikimedia\Parsoid\Ext\ExtensionTagHandler;
+use Wikimedia\Parsoid\Ext\ParsoidExtensionAPI;
+
+class InfoboxTag extends ExtensionTagHandler implements ExtensionModule {
+
+	/**
+	 * @inheritDoc
+	 */
+	public function getConfig(): array
+	{
+		return [
+			'name' => 'PortableInfobox',
+			'tags' => [
+				[
+					'name' => 'infobox',
+					'handler' => self::class
+				]
+			],
+			'domProcessors' => [
+				'PortableInfobox\\Parsoid\\PortableInfoboxDOMProcessor'
+			]
+		];
+	}
+
+    /**
+     * @inheritDoc
+     */
+    public function sourceToDom( ParsoidExtensionAPI $api, string $src, array $args )
+	{
+		$domFragments = $api->extTagToDOM( $args, $src, [
+			'wrapperTag' => 'aside',
+			'parseOpts' => [
+				'extTag' => 'infobox',
+				'context' => 'inline'
+			]
+		]);
+
+
+		// this is a bit messed up as these methods are deprecated, but the documentation
+		// for the replacement methods doesn't exist or make sense
+		// this is commented out at present, as these scripts and styles will be added
+		// by the legacy parser (might need it here to stick these in the parser cache also?)
+		$api->getMetadata()->addModules( [ 'ext.PortableInfobox.scripts' ] );
+		$api->getMetadata()->addModuleStyles( [ 'ext.PortableInfobox.styles' ] );
+
+		// return this back. At this point, we have constructed the outer tag (<aside class=...</aside>)
+		// and this function is done with its work. The rest of the work will happen in the DOMProcessor
+		return $domFragments;
+	}
+}

includes/Parsoid/ParsoidMediaWikiParser.php

@@ -0,0 +1,117 @@
+<?php
+
+namespace PortableInfobox\Parsoid;
+
+use MediaWiki\Title\Title;
+use PortableInfobox\Services\Parser\ExternalParser;
+use ReflectionClass;
+use Wikimedia\Parsoid\Ext\ParsoidExtensionAPI;
+use Wikimedia\Parsoid\Fragments\WikitextPFragment;
+use Wikimedia\Parsoid\Utils\DOMCompat;
+
+class ParsoidMediaWikiParser implements ExternalParser {
+
+    public ParsoidExtensionAPI $api;
+
+    public function __construct(
+        ParsoidExtensionAPI $api
+    ) {
+        $this->api = $api;
+    }
+
+    public function parseRecursive( $wikitext ) 
+    {
+        if ( $wikitext === null ) {
+            return null; 
+        }
+
+        $paramParsed = $this->api->wikitextToDOM( $wikitext, [
+			// this differs from earlier as we need the frame to be able to grab the 
+            // params the user passed - parsoid handles this internally it appears
+            'processInNewFrame' => false,
+			'parseOpts' => [ 'context' => 'inline' ]
+		], true );
+
+        // we don't want Parsoid to wrap in a span or add a typeof here, 
+        // just interested in the content
+        return DOMCompat::getOuterHTML( $paramParsed );
+    }
+
+    public function replaceVariables( $wikitext ) {
+		// no-op - I think handled by ->wikiTextToDOM? 
+	}
+
+    public function addImage($title, array $sizeParams): ?string
+    {
+        // no-op at present. Used to extend the image tag with specific information for the PageImages extension. 
+        // that extension relies on Parser hooks and therefore is not safe to assume that will work indefinitely.
+        // could potentially just do it for now whilst the hooks still exist, and maybe remove at a later date if
+        // PageImages is not made Parsoid-compat. 
+        return '';
+    }
+
+    public function getParsoidExtensionApi(): ParsoidExtensionAPI {
+        return $this->api;
+    }
+
+    /**
+     * Extract the gallery and return the filename -> captions. PortableInfobox currently does this
+     * a lot cleaner as it piggybacks on onAfterParserFetchFileAndTitle hook to set the images into the data bag.
+     * This hook is NOT available on Parsoid, and we have no other way to get the resultant class which PortableInfobox
+     * currently relies on. So we need to fake it as best we can and hope WMF comes up with something later down
+     * the line.
+     * @param mixed $wikitext
+     * @return array an array of the images
+     */
+    public function extractGallery( string $wikitext ): array
+    {
+        if ( $wikitext === null ) {
+            return [];
+        }
+
+        // the legacy implementation reuturns an array where each element is an array of the caption
+        // and the title object for that specific image. We don't have access to this by default,
+        // since there is no concept of half parsing in Parsoid - we either ask Parsoid for the Parsed wt->html
+        // or we work with the WT and grab what we need.
+        $result = [];
+
+        // this is quicker than passing the wikitext to Parsoid and extracting
+        // the images etc from it. 
+        if ( preg_match( '/<gallery[^>]*>(.*?)<\/gallery>/s', $wikitext, $matches ) ) {
+            $galleryContent = trim( $matches[1] );
+
+            if ( empty( $galleryContent ) ) {
+                return [];
+            }
+
+            $lines = explode( "\n", $galleryContent );
+
+            foreach ( $lines as $line ) {
+                $line = trim( $line );
+
+                if ( empty( $line ) ) {
+                    continue;
+                }
+
+                $parts = explode( '|', $line, 2 );
+                $filename = trim( $parts[0] );
+                $caption = isset( $parts[1] ) ? trim( $parts[1] ) : '';
+
+                if ( empty( $filename ) ) {
+                    continue;
+                }
+
+                $title = Title::newFromText( $filename, NS_FILE );
+
+                if ( $title !== null ) {
+                    $result[] = [
+                        'label' => $caption,
+                        'title' => $title
+                    ];
+                }
+            }
+        }
+
+        return $result;
+    }
+}