From 3e2423d86b0a2d5e6b53a30ee48a0ad6e41b2702 Mon Sep 17 00:00:00 2001
From: Yaman Qalieh <ybq987@gmail.com>
Date: Tue, 14 Jun 2022 09:45:46 -0400
Subject: [PATCH] [MsnMondeBridge] Fix bridge (#2813)

---
 bridges/MsnMondeBridge.php | 48 +++++++++++++++++++++-----------------
 1 file changed, 26 insertions(+), 22 deletions(-)

diff --git a/bridges/MsnMondeBridge.php b/bridges/MsnMondeBridge.php
index 9c418bd9..817f13ad 100644
--- a/bridges/MsnMondeBridge.php
+++ b/bridges/MsnMondeBridge.php
@@ -1,36 +1,40 @@
 <?php
-class MsnMondeBridge extends BridgeAbstract {
+class MsnMondeBridge extends FeedExpander {
 
 	const MAINTAINER = 'kranack';
 	const NAME = 'MSN Actu Monde';
-	const URI = 'http://www.msn.com/';
 	const DESCRIPTION = 'Returns the 10 newest posts from MSN Actualités (full text)';
+	const URI = 'https://www.msn.com/fr-fr/actualite';
+	const FEED_URL = 'https://rss.msn.com/fr-fr';
+	const JSON_URL = 'https://assets.msn.com/content/view/v2/Detail/fr-fr/';
+	const LIMIT = 10;
 
-	public function getURI(){
-		return self::URI . 'fr-fr/actualite/monde';
+	public function getName() {
+		return 'MSN Actualités';
 	}
 
-	private function msnMondeExtractContent($url, &$item){
-		$html2 = getSimpleHTMLDOM($url);
-		$item['content'] = $html2->find('#content', 0)->find('article', 0)->find('section', 0)->plaintext;
-		$item['timestamp'] = strtotime($html2->find('.authorinfo-txt', 0)->find('time', 0)->datetime);
+	public function getURI() {
+		return self::URI;
 	}
 
-	public function collectData(){
-		$html = getSimpleHTMLDOM($this->getURI());
+	public function collectData() {
+		$this->collectExpandableDatas(self::FEED_URL, self::LIMIT);
+	}
 
-		$limit = 0;
-
-		// TODO: fix why articles is empty
-		foreach($html->find('.smalla') as $article) {
-			if($limit < 10) {
-				$item = array();
-				$item['title'] = utf8_decode($article->find('h4', 0)->innertext);
-				$item['uri'] = self::URI . utf8_decode($article->find('a', 0)->href);
-				$this->msnMondeExtractContent($item['uri'], $item);
-				$this->items[] = $item;
-				$limit++;
-			}
+	protected function parseItem($newsItem) {
+		$item = parent::parseItem($newsItem);
+		if (!preg_match('#fr-fr/actualite.*/ar-(?<id>[\w]*)\?#', $item['uri'], $matches)) {
+			return;
 		}
+
+		$json = json_decode(getContents(self::JSON_URL . $matches['id']), true);
+		$item['content'] = $json['body'];
+		if (!empty($json['authors']))
+			$item['author'] = reset($json['authors'])['name'];
+		$item['timestamp'] = $json['createdDateTime'];
+		foreach($json['tags'] as $tag) {
+			$item['categories'][] = $tag['label'];
+		}
+		return $item;
 	}
 }