Browse Source

af_redditimgur updates:
1. remove special handling for imgur albums/pages because its full on cancerous json garbage now with no scrapeable layout
2. skip all urls leading to reddit.com while trying to embed content
3. implement generic last resort embedding based on meta og:image element

Andrew Dolgov 6 years ago
parent
commit
b70c3daffe
1 changed files with 43 additions and 6 deletions
  1. 43 6
      plugins/af_redditimgur/init.php

+ 43 - 6
plugins/af_redditimgur/init.php

@@ -84,15 +84,16 @@ class Af_RedditImgur extends Plugin {
 		$img_entries = $xpath->query("(//img[@src])");
 
 		$found = false;
+		//$debug = 1;
 
 		foreach ($entries as $entry) {
-			if ($entry->hasAttribute("href")) {
+			if ($entry->hasAttribute("href") && strpos($entry->getAttribute("href"), "reddit.com") === FALSE) {
 
 				_debug("processing href: " . $entry->getAttribute("href"), $debug);
 
 				$matches = array();
 
-				if (preg_match("/^https?:\/\/twitter.com\/(.*?)\/status\/(.*)/", $entry->getAttribute("href"), $matches)) {
+				if (!$found && preg_match("/^https?:\/\/twitter.com\/(.*?)\/status\/(.*)/", $entry->getAttribute("href"), $matches)) {
 					_debug("handling as twitter: " . $matches[1] . " " . $matches[2], $debug);
 
 					$oembed_result = fetch_file_contents("https://publish.twitter.com/oembed?url=" . urlencode($entry->getAttribute("href")));
@@ -285,7 +286,7 @@ class Af_RedditImgur extends Plugin {
 
 				// linked albums & pages
 
-				if (!$found && preg_match("/^https?:\/\/(m\.)?imgur.com\/([^\.\/]+$)/", $entry->getAttribute("href"), $matches) ||
+				/*if (!$found && preg_match("/^https?:\/\/(m\.)?imgur.com\/([^\.\/]+$)/", $entry->getAttribute("href"), $matches) ||
 					preg_match("/^https?:\/\/(m\.)?imgur.com\/(a|album|gallery)\/[^\.]+$/", $entry->getAttribute("href"), $matches)) {
 
 					_debug("Handling as an imgur page/album/gallery", $debug);
@@ -339,7 +340,7 @@ class Af_RedditImgur extends Plugin {
 							if ($debug) print_r($urls);
 						}
 					}
-				}
+				} */
 
 				// wtf is this even
 				if (!$found && preg_match("/^https?:\/\/gyazo\.com\/([^\.\/]+$)/", $entry->getAttribute("href"), $matches)) {
@@ -356,6 +357,41 @@ class Af_RedditImgur extends Plugin {
 
 					$found = true;
 				}
+
+				// let's try meta properties
+				if (!$found) {
+					_debug("looking for meta og:image", $debug);
+
+					$content = fetch_file_contents(["url" => $entry->getAttribute("href"),
+						"http_accept" => "text/*"]);
+
+					if ($content) {
+						$cdoc = new DOMDocument();
+
+						if (@$cdoc->loadHTML($content)) {
+							$cxpath = new DOMXPath($cdoc);
+
+							$og_image = $cxpath->query("//meta[@property='og:image']")->item(0);
+
+							if ($og_image) {
+
+								$og_src = $og_image->getAttribute("content");
+
+								if ($og_src) {
+									$img = $doc->createElement('img');
+									$img->setAttribute("src", $og_src);
+
+									$br = $doc->createElement('br');
+									$entry->parentNode->insertBefore($img, $entry);
+									$entry->parentNode->insertBefore($br, $entry);
+
+									$found = true;
+								}
+							}
+						}
+					}
+				}
+
 			}
 
 			// remove tiny thumbnails
@@ -534,8 +570,9 @@ class Af_RedditImgur extends Plugin {
 
 				if ($content_type && strpos($content_type, "text/html") !== FALSE) {
 
-					$tmp = fetch_file_contents(array("url" => $url,
-						"useragent" => $useragent_compat));
+					$tmp = fetch_file_contents(["url" => $url,
+						"useragent" => $useragent_compat,
+						"http_accept" => "text/html"]);
 
 					if ($debug) _debug("tmplen: " . mb_strlen($tmp));