Pagecache for frequently fetched pages

database.sql

@@ -1,6 +1,6 @@
 -- ------------------------------------------
 -- Friendica 2022.09-rc (Giant Rhubarb)
--- DB_UPDATE_VERSION 1482
+-- DB_UPDATE_VERSION 1483
 -- ------------------------------------------
 
 
@@ -1007,6 +1007,17 @@ CREATE TABLE IF NOT EXISTS `openwebauth-token` (
 	FOREIGN KEY (`uid`) REFERENCES `user` (`uid`) ON UPDATE RESTRICT ON DELETE CASCADE
 ) DEFAULT COLLATE utf8mb4_general_ci COMMENT='Store OpenWebAuth token to verify contacts';
 
+--
+-- TABLE pagecache
+--
+CREATE TABLE IF NOT EXISTS `pagecache` (
+	`page` varbinary(255) NOT NULL COMMENT 'Page',
+	`content` mediumtext COMMENT 'Page content',
+	`fetched` datetime COMMENT 'date when the page had been fetched',
+	 PRIMARY KEY(`page`),
+	 INDEX `fetched` (`fetched`)
+) DEFAULT COLLATE utf8mb4_general_ci COMMENT='Stores temporary data';
+
 --
 -- TABLE parsed_url
 --

doc/database.md

@@ -47,6 +47,7 @@ Database Tables
 | [notify-threads](help/database/db_notify-threads) |  |
 | [oembed](help/database/db_oembed) | cache for OEmbed queries |
 | [openwebauth-token](help/database/db_openwebauth-token) | Store OpenWebAuth token to verify contacts |
+| [pagecache](help/database/db_pagecache) | Stores temporary data |
 | [parsed_url](help/database/db_parsed_url) | cache for 'parse_url' queries |
 | [pconfig](help/database/db_pconfig) | personal (per user) configuration storage |
 | [permissionset](help/database/db_permissionset) |  |

doc/database/db_pagecache.md

@@ -0,0 +1,24 @@
+Table pagecache
+===========
+
+Stores temporary data
+
+Fields
+------
+
+| Field   | Description                         | Type           | Null | Key | Default | Extra |
+| ------- | ----------------------------------- | -------------- | ---- | --- | ------- | ----- |
+| page    | Page                                | varbinary(255) | NO   | PRI | NULL    |       |
+| content | Page content                        | mediumtext     | YES  |     | NULL    |       |
+| fetched | date when the page had been fetched | datetime       | YES  |     | NULL    |       |
+
+Indexes
+------------
+
+| Name    | Fields  |
+| ------- | ------- |
+| PRIMARY | page    |
+| fetched | fetched |
+
+
+Return to [database documentation](help/database)

src/Model/Contact.php

@@ -3321,7 +3321,7 @@ class Contact
 				continue;
 			}
 			$contact = self::getByURL($url, false, ['id', 'updated']);
-			if (empty($contact['id'])) {
+			if (empty($contact['id']) && Network::isValidHttpUrl($url)) {
 				Worker::add(PRIORITY_LOW, 'AddContact', 0, $url);
 				++$added;
 			} elseif ($contact['updated'] < DateTimeFormat::utc('now -7 days')) {

src/Model/ItemURI.php

@@ -21,8 +21,10 @@
 
 namespace Friendica\Model;
 
+use Friendica\Core\Logger;
 use Friendica\Database\Database;
 use Friendica\Database\DBA;
+use Friendica\DI;
 
 class ItemURI
 {
@@ -35,14 +37,16 @@ class ItemURI
 	 */
 	public static function insert(array $fields)
 	{
+		$fields = DI::dbaDefinition()->truncateFieldsForTable('item-uri', $fields);
+
 		if (!DBA::exists('item-uri', ['uri' => $fields['uri']])) {
-			DBA::insert('item-uri', $fields, Database::INSERT_UPDATE);
+			DBA::insert('item-uri', $fields, Database::INSERT_IGNORE);
 		}
 
 		$itemuri = DBA::selectFirst('item-uri', ['id', 'guid'], ['uri' => $fields['uri']]);
-
 		if (!DBA::isResult($itemuri)) {
 			// This shouldn't happen
+			Logger::warning('Item-uri not found', $fields);
 			return null;
 		}
 
@@ -77,22 +81,4 @@ class ItemURI
 
 		return $itemuri['id'] ?? 0;
 	}
-
-	/**
-	 * Searched for an id of a given guid.
-	 *
-	 * @param string $guid
-	 * @return integer item-uri id
-	 * @throws \Exception
-	 */
-	public static function getIdByGUID(string $guid): int
-	{
-		$itemuri = DBA::selectFirst('item-uri', ['id'], ['guid' => $guid]);
-
-		if (!DBA::isResult($itemuri)) {
-			return 0;
-		}
-
-		return $itemuri['id'];
-	}
 }

src/Module/ActivityPub/Objects.php

@@ -31,6 +31,7 @@ use Friendica\Model\Item;
 use Friendica\Model\Post;
 use Friendica\Network\HTTPException;
 use Friendica\Protocol\ActivityPub;
+use Friendica\Protocol\ActivityPub\PageCache;
 use Friendica\Util\HTTPSignature;
 use Friendica\Util\Network;
 use Friendica\Util\Strings;
@@ -50,6 +51,13 @@ class Objects extends BaseModule
 			DI::baseUrl()->redirect(str_replace('objects/', 'display/', DI::args()->getQueryString()));
 		}
 
+		$data = PageCache::fetch($_SERVER['REQUEST_URI']);
+		if (!empty($data)) {
+			header('Access-Control-Allow-Origin: *');
+
+			System::jsonExit($data, 'application/activity+json');
+		}
+
 		$itemuri = DBA::selectFirst('item-uri', ['id'], ['guid' => $this->parameters['guid']]);
 
 		if (DBA::isResult($itemuri)) {
@@ -127,6 +135,10 @@ class Objects extends BaseModule
 			throw new HTTPException\NotFoundException();
 		}
 
+		if (in_array($item['private'], [Item::PUBLIC, Item::UNLISTED])) {
+			PageCache::add($_SERVER['REQUEST_URI'], $data);
+		}
+
 		// Relaxed CORS header for public items
 		header('Access-Control-Allow-Origin: *');
 

src/Module/Settings/Account.php

@@ -38,6 +38,7 @@ use Friendica\Model\Verb;
 use Friendica\Module\BaseSettings;
 use Friendica\Network\HTTPException;
 use Friendica\Protocol\Activity;
+use Friendica\Util\Network;
 use Friendica\Util\Temporal;
 use Friendica\Worker\Delivery;
 
@@ -373,7 +374,7 @@ class Account extends BaseSettings
 						// or the handle of the account, therefore we check for either
 						// "http" or "@" to be present in the string.
 						// All other fields from the row will be ignored
-						if ((strpos($csvRow[0], '@') !== false) || in_array(parse_url($csvRow[0], PHP_URL_SCHEME), ['http', 'https'])) {
+						if ((strpos($csvRow[0], '@') !== false) || Network::isValidHttpUrl($csvRow[0])) {
 							Worker::add(PRIORITY_MEDIUM, 'AddContact', local_user(), $csvRow[0]);
 						} else {
 							Logger::notice('Invalid account', ['url' => $csvRow[0]]);

src/Protocol/ActivityPub/PageCache.php

@@ -0,0 +1,73 @@
+<?php
+/**
+ * @copyright Copyright (C) 2010-2022, the Friendica project
+ *
+ * @license GNU AGPL version 3 or any later version
+ *
+ * This program is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU Affero General Public License as
+ * published by the Free Software Foundation, either version 3 of the
+ * License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU Affero General Public License for more details.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program.  If not, see <https://www.gnu.org/licenses/>.
+ *
+ */
+
+namespace Friendica\Protocol\ActivityPub;
+
+use Friendica\Core\Logger;
+use Friendica\Database\Database;
+use Friendica\Database\DBA;
+use Friendica\DI;
+use Friendica\Util\DateTimeFormat;
+
+/**
+ * This class handles the page cache
+ */
+class PageCache
+{
+	/**
+	 * Add content to the page cache
+	 *
+	 * @param string $page
+	 * @param mixed $content
+	 * @return void
+	 */
+	public static function add(string $page, $content)
+	{
+		if (!DI::config()->get('system', 'pagecache')) {
+			return;
+		}
+
+		DBA::delete('pagecache', ["`fetched` < ?", DateTimeFormat::utc('now - 5 minutes')]);
+		DBA::insert('pagecache', ['page' => $page, 'content' => serialize($content), 'fetched' => DateTimeFormat::utcNow()], Database::INSERT_UPDATE);
+
+		Logger::debug('Page added', ['page' => $page]);
+	}
+
+	/**
+	 * Fetch data from the page cache
+	 *
+	 * @param string $page
+	 * @return mixed
+	 */
+	public static function fetch(string $page)
+	{
+		$pagecache = DBA::selectFirst('pagecache', [], ['page' => $page]);
+		if (empty($pagecache['content'])) {
+			return null;
+		}
+
+		DBA::update('pagecache', ['fetched' => DateTimeFormat::utcNow()], ['page' => $page]);
+
+		Logger::debug('Page fetched', ['page' => $page]);
+
+		return unserialize($pagecache['content']);
+	}
+}

src/Worker/OptimizeTables.php

@@ -43,6 +43,7 @@ class OptimizeTables
 		DBA::e("OPTIMIZE TABLE `cache`");
 		DBA::e("OPTIMIZE TABLE `locks`");
 		DBA::e("OPTIMIZE TABLE `oembed`");
+		DBA::e("OPTIMIZE TABLE `pagecache`");
 		DBA::e("OPTIMIZE TABLE `parsed_url`");
 		DBA::e("OPTIMIZE TABLE `session`");
 

static/dbstructure.config.php

@@ -55,7 +55,7 @@
 use Friendica\Database\DBA;
 
 if (!defined('DB_UPDATE_VERSION')) {
-	define('DB_UPDATE_VERSION', 1482);
+	define('DB_UPDATE_VERSION', 1483);
 }
 
 return [
@@ -795,7 +795,7 @@ return [
 			"conversation" => ["type" => "varbinary(383)", "comment" => ""],
 			"type" => ["type" => "varchar(64)", "comment" => "Type of the activity"],
 			"object-type" => ["type" => "varchar(64)", "comment" => "Type of the object activity"],
-			"object-object-type" => ["type" => "varchar(64)", "comment" => "Type of the object's object activity"],			
+			"object-object-type" => ["type" => "varchar(64)", "comment" => "Type of the object's object activity"],
 			"received" => ["type" => "datetime", "comment" => "Receiving date"],
 			"activity" => ["type" => "mediumtext", "comment" => "The JSON activity"],
 			"signer" => ["type" => "varchar(255)", "comment" => ""],
@@ -1051,6 +1051,18 @@ return [
 			"uid" => ["uid"],
 		]
 	],
+	"pagecache" => [
+		"comment" => "Stores temporary data",
+		"fields" => [
+			"page" => ["type" => "varbinary(255)", "not null" => "1", "primary" => "1", "comment" => "Page"],
+			"content" => ["type" => "mediumtext", "comment" => "Page content"],
+			"fetched" => ["type" => "datetime", "comment" => "date when the page had been fetched"],
+		],
+		"indexes" => [
+			"PRIMARY" => ["page"],
+			"fetched" => ["fetched"],
+		],
+	],
 	"parsed_url" => [
 		"comment" => "cache for 'parse_url' queries",
 		"fields" => [

static/defaults.config.php

@@ -478,6 +478,10 @@ return [
 		// Don't show smilies.
 		'no_smilies' => false,
 
+		// pagecache (Boolean)
+		// Cache frequently fetched pages.
+		'pagecache' => false,
+
 		// paranoia (Boolean)
 		// Log out users if their IP address changed.
 		'paranoia' => false,
@@ -643,7 +647,7 @@ return [
 		'worker_load_exponent' => 3,
 
 		// worker_processes_cooldown (Integer)
-		// Maximum number pro processes that causes a cooldown before each worker function call.
+		// Maximum number per processes that causes a cooldown before each worker function call.
 		'worker_processes_cooldown' => 0,
 
 		// worker_multiple_fetch (Boolean)