Extract description algorithm into a new class

Separating the extract algorithm from integration code. This results in a slightly cleaner code structure (at least in my opinion) and enables adding alternate algorithms without devolving into spaghetti.

The DescriptionProvider (name of the new base interface) is exposed as a service through dependency injection to avoid factories. The implementation can be swapped at service instantiation time.

Depends-On: I73c61ce045dcf31ac1ca5888f1548de8fd8b56ff
Change-Id: I97fd065c9554837747021ba9fff26005e33270f4

extension.json

@@ -17,6 +17,9 @@
 	"ConfigRegistry": {
 		"Description2": "GlobalVarConfig::newInstance"
 	},
+	"ServiceWiringFiles": [
+		"includes/ServiceWiring.php"
+	],
 	"AutoloadNamespaces": {
 		"MediaWiki\\Extension\\Description2\\": "includes/"
 	},
@@ -32,7 +35,8 @@
 		"Description2": {
 			"class": "MediaWiki\\Extension\\Description2\\Hooks",
 			"services": [
-				"ConfigFactory"
+				"ConfigFactory",
+				"Description2.DescriptionProvider"
 			]
 		}
 	},

includes/Description2.php

@@ -17,7 +17,6 @@ use PPFrame;
  */
 
 class Description2 {
-
 	/**
 	 * @param Parser $parser The parser.
 	 * @param string $desc The description text.

includes/DescriptionProvider.php

@@ -0,0 +1,13 @@
+<?php
+
+namespace MediaWiki\Extension\Description2;
+
+interface DescriptionProvider {
+	/**
+	 * Extracts description from the HTML representation of a page.
+	 *
+	 * @param string $text HTML to extract the description from.
+	 * @return ?string
+	 */
+	public function derive( string $text ): ?string;
+}

includes/Hooks.php

@@ -25,15 +25,21 @@ class Hooks implements
 	\MediaWiki\Hook\OutputPageParserOutputHook
 {
 
+	/** @var Config */
 	private Config $config;
 
+	/** @var DescriptionProvider */
+	private DescriptionProvider $descriptionProvider;
+
 	/**
 	 * @param ConfigFactory $configFactory
 	 */
 	public function __construct(
-		ConfigFactory $configFactory
+		ConfigFactory $configFactory,
+		DescriptionProvider $descriptionProvider
 	) {
 		$this->config = $configFactory->makeConfig( 'Description2' );
+		$this->descriptionProvider = $descriptionProvider;
 	}
 
 	/**
@@ -43,22 +49,7 @@ class Hooks implements
 	 * @return bool
 	 */
 	public function onParserAfterTidy( $parser, &$text ) {
-		$desc = '';
+		$desc = $this->descriptionProvider->derive( $text );
-
-		$pattern = '%<table\b[^>]*+>(?:(?R)|[^<]*+(?:(?!</?table\b)<[^<]*+)*+)*+</table>%i';
-		$myText = preg_replace( $pattern, '', $text );
-
-		$paragraphs = [];
-		if ( preg_match_all( '#<p>.*?</p>#is', $myText, $paragraphs ) ) {
-			foreach ( $paragraphs[0] as $paragraph ) {
-				$paragraph = trim( strip_tags( $paragraph ) );
-				if ( !$paragraph ) {
-					continue;
-				}
-				$desc = $paragraph;
-				break;
-			}
-		}
 
 		if ( $desc ) {
 			Description2::setDescription( $parser, $desc );

includes/ServiceWiring.php

@@ -0,0 +1,13 @@
+<?php
+
+use MediaWiki\Extension\Description2\DescriptionProvider;
+use MediaWiki\Extension\Description2\SimpleDescriptionProvider;
+use MediaWiki\MediaWikiServices;
+
+return [
+	'Description2.DescriptionProvider' => static function (
+		MediaWikiServices $services
+	): DescriptionProvider {
+		return new SimpleDescriptionProvider();
+	},
+];

includes/SimpleDescriptionProvider.php

@@ -0,0 +1,36 @@
+<?php
+
+namespace MediaWiki\Extension\Description2;
+
+class SimpleDescriptionProvider implements DescriptionProvider {
+
+	/**
+	 * Extracts description from the HTML representation of a page.
+	 *
+	 * The algorithm:
+	 * 1. Removes all <table> elements and their contents.
+	 * 2. Selects all <p> elements.
+	 * 3. Iterates over those paragraphs, strips out all HTML tags and trims white-space around.
+	 * 4. Then the first non-empty paragraph is picked as the result.
+	 *
+	 * @param string $text
+	 * @return string
+	 */
+	public function derive( string $text ): ?string {
+		$pattern = '%<table\b[^>]*+>(?:(?R)|[^<]*+(?:(?!</?table\b)<[^<]*+)*+)*+</table>%i';
+		$myText = preg_replace( $pattern, '', $text );
+
+		$paragraphs = [];
+		if ( preg_match_all( '#<p>.*?</p>#is', $myText, $paragraphs ) ) {
+			foreach ( $paragraphs[0] as $paragraph ) {
+				$paragraph = trim( strip_tags( $paragraph ) );
+				if ( !$paragraph ) {
+					continue;
+				}
+				return $paragraph;
+			}
+		}
+
+		return null;
+	}
+}