implement experimental article on-import data filters

classes/filter.php

@@ -0,0 +1,14 @@
+<?php
+class Filter {
+	protected $link;
+
+	function __construct($link) {
+		$this->link = $link;
+	}
+
+	function filter_article($article) {
+		return $article;
+	}
+
+}
+?>

classes/filter/redditimgur.php

@@ -0,0 +1,47 @@
+<?php
+class Filter_RedditImgur {
+
+	function filter_article($article) {
+
+		if (strpos($article["link"], "reddit.com/r/") !== FALSE) {
+			if (strpos($article["content"], "i.imgur.com") !== FALSE) {
+
+				$doc = new DOMDocument();
+				@$doc->loadHTML($article["content"]);
+
+				if ($doc) {
+					$xpath = new DOMXPath($doc);
+					$entries = $xpath->query('(//a[@href]|//img[@src])');
+
+					foreach ($entries as $entry) {
+						if ($entry->hasAttribute("href")) {
+							if (preg_match("/i.imgur.com\/.*?.jpg/", $entry->getAttribute("href"))) {
+
+							 	$img = $doc->createElement('img');
+								$img->setAttribute("src", $entry->getAttribute("href"));
+
+								$entry->parentNode->replaceChild($img, $entry);
+							}
+						}
+
+						// remove tiny thumbnails
+						if ($entry->hasAttribute("src")) {
+							if ($entry->parentNode && $entry->parentNode->parentNode) {
+								$entry->parentNode->parentNode->removeChild($entry->parentNode);
+							}
+						}
+					}
+
+					$node = $doc->getElementsByTagName('body')->item(0);
+
+					if ($node) {
+						$article["content"] = $doc->saveXML($node, LIBXML_NOEMPTYTAG);
+					}
+				}
+			}
+		}
+
+		return $article;
+	}
+}
+?>

include/functions.php

@@ -7,11 +7,19 @@
 	function __autoload($class) {
 		$class_file = str_replace("_", "/", strtolower(basename($class)));
 
+		$file = dirname(__FILE__)."/../plugins/$class_file.php";
+
+		if (file_exists($file)) {
+			require $file;
+			return;
+		}
+
 		$file = dirname(__FILE__)."/../classes/$class_file.php";
 
 		if (file_exists($file)) {
 			require $file;
 		}
+
 	}
 
 	mb_internal_encoding("UTF-8");

include/rssfuncs.php

@@ -398,6 +398,23 @@
 				_debug("update_rss_feed: " . count($filters) . " filters loaded.");
 			}
 
+			$filter_plugins = array();
+
+			if (defined('_ARTICLE_FILTER_PLUGINS')) {
+				foreach (explode(",", _ARTICLE_FILTER_PLUGINS) as $p) {
+					$pclass = "filter_" . trim($p);
+
+					if (class_exists($pclass)) {
+						$plugin = new $pclass($link);
+						array_push($filter_plugins, $plugin);
+					}
+				}
+			}
+
+			if ($debug_enabled) {
+				_debug("update_rss_feed: " . count($filter_plugins) . " filter plugins loaded.");
+			}
+
 			if ($use_simplepie) {
 				$iterator = $rss->get_items();
 			} else {
@@ -651,8 +668,6 @@
 
 				$entry_content = db_escape_string($entry_content, false);
 
-				$content_hash = "SHA1:" . sha1(strip_tags($entry_content));
-
 				$entry_title = db_escape_string($entry_title);
 				$entry_link = db_escape_string($entry_link);
 				$entry_comments = mb_substr(db_escape_string($entry_comments), 0, 250);
@@ -757,6 +772,31 @@
 					_debug("update_rss_feed: done collecting data [TITLE:$entry_title]");
 				}
 
+				// TODO: less memory-hungry implementation
+				if (count($filter_plugins) > 0) {
+					if ($debug_enabled) {
+						_debug("update_rss_feed: applying plugin filters...");
+					}
+
+					$article = array("owner_uid" => $owner_uid,
+						"title" => $entry_title,
+						"content" => $entry_content,
+						"link" => $entry_link,
+						"tags" => $entry_tags,
+						"author" => $entry_author);
+
+					foreach ($filter_plugins as $plugin) {
+						$article = $plugin->filter_article($article);
+					}
+
+					$entry_title = $article["title"];
+					$entry_content = $article["content"];
+					$entry_tags = $article["tags"];
+					$entry_author = $article["author"];
+				}
+
+				$content_hash = "SHA1:" . sha1(strip_tags($entry_content));
+
 				db_query($link, "BEGIN");
 
 				if (db_num_rows($result) == 0) {