]> git.wh0rd.org - tt-rss.git/commitdiff
af_redditimgur updates:
authorAndrew Dolgov <noreply@fakecake.org>
Sat, 26 May 2018 07:25:39 +0000 (10:25 +0300)
committerAndrew Dolgov <noreply@fakecake.org>
Sat, 26 May 2018 07:25:39 +0000 (10:25 +0300)
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

plugins/af_redditimgur/init.php

index 59e3a760c666f68ba54939dee92ec38ef3fea384..4079a5b3053068187f6908510383eb01391503d0 100755 (executable)
@@ -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));