Merge pull request #18742 from owncloud/mimetype-updatedb

Introduce mimetype DB update occ command

4 files changed, 217 insertions, 88 deletions

diff --git a/lib/private/files/cache/cache.php b/lib/private/files/cache/cache.php
index 8cf097421d4..5c04da1f0d5 100644
--- a/lib/private/files/cache/cache.php
+++ b/lib/private/files/cache/cache.php
@@ -35,6 +35,8 @@
 
 namespace OC\Files\Cache;
 
+use \OCP\Files\IMimeTypeLoader;
+
 /**
  * Metadata cache for a storage
  *
@@ -66,8 +68,8 @@ class Cache {
 	 */
 	protected $storageCache;
 
-	protected static $mimetypeIds = array();
-	protected static $mimetypes = array();
+	/** @var IMimeTypeLoader */
+	protected $mimetypeLoader;
 
 	/**
 	 * @param \OC\Files\Storage\Storage|string $storage
@@ -83,6 +85,7 @@ class Cache {
 		}
 
 		$this->storageCache = new Storage($storage);
+		$this->mimetypeLoader = \OC::$server->getMimeTypeLoader();
 	}
 
 	/**
@@ -95,72 +98,6 @@ class Cache {
 	}
 
 	/**
-	 * Get the numeric id for a mimetype
-	 *
-	 * Mimetypes are stored as integers in the cache to prevent duplicated data of the (usually) fairly limited amount of unique mimetypes
-	 * If the supplied mimetype does not yet have a numeric id a new one will be generated
-	 *
-	 * @param string $mime
-	 * @return int
-	 */
-	public function getMimetypeId($mime) {
-		if (empty($mime)) {
-			// Can not insert empty string into Oracle NOT NULL column.
-			$mime = 'application/octet-stream';
-		}
-		if (empty(self::$mimetypeIds)) {
-			$this->loadMimetypes();
-		}
-
-		if (!isset(self::$mimetypeIds[$mime])) {
-			try {
-				$connection = \OC_DB::getConnection();
-				$connection->insertIfNotExist('*PREFIX*mimetypes', [
-					'mimetype'	=> $mime,
-				]);
-				$this->loadMimetypes();
-			} catch (\Doctrine\DBAL\DBALException $e) {
-				\OCP\Util::writeLog('core', 'Exception during mimetype insertion: ' . $e->getmessage(), \OCP\Util::DEBUG);
-				return -1;
-			}
-		}
-
-		return self::$mimetypeIds[$mime];
-	}
-
-
-	/**
-	 * Get the mimetype (as string) from a mimetype id
-	 *
-	 * @param int $id
-	 * @return string | null the mimetype for the id or null if the id is not known
-	 */
-	public function getMimetype($id) {
-		if (empty(self::$mimetypes)) {
-			$this->loadMimetypes();
-		}
-
-		return isset(self::$mimetypes[$id]) ? self::$mimetypes[$id] : null;
-	}
-
-	/**
-	 * Load all known mimetypes and mimetype ids from the database
-	 *
-	 * @throws \OC\DatabaseException
-	 */
-	public function loadMimetypes() {
-		self::$mimetypeIds = self::$mimetypes = array();
-
-		$result = \OC_DB::executeAudited('SELECT `id`, `mimetype` FROM `*PREFIX*mimetypes`', array());
-		if ($result) {
-			while ($row = $result->fetchRow()) {
-				self::$mimetypeIds[$row['mimetype']] = $row['id'];
-				self::$mimetypes[$row['id']] = $row['mimetype'];
-			}
-		}
-	}
-
-	/**
 	 * get the stored metadata of a file or folder
 	 *
 	 * the returned cache entry contains at least the following values:
@@ -222,8 +159,8 @@ class Cache {
 			$data['storage_mtime'] = (int)$data['storage_mtime'];
 			$data['encrypted'] = (bool)$data['encrypted'];
 			$data['storage'] = $this->storageId;
-			$data['mimetype'] = $this->getMimetype($data['mimetype']);
-			$data['mimepart'] = $this->getMimetype($data['mimepart']);
+			$data['mimetype'] = $this->mimetypeLoader->getMimetypeById($data['mimetype']);
+			$data['mimepart'] = $this->mimetypeLoader->getMimetypeById($data['mimepart']);
 			if ($data['storage_mtime'] == 0) {
 				$data['storage_mtime'] = $data['mtime'];
 			}
@@ -258,8 +195,8 @@ class Cache {
 			$result = \OC_DB::executeAudited($sql, array($fileId));
 			$files = $result->fetchAll();
 			foreach ($files as &$file) {
-				$file['mimetype'] = $this->getMimetype($file['mimetype']);
-				$file['mimepart'] = $this->getMimetype($file['mimepart']);
+				$file['mimetype'] = $this->mimetypeLoader->getMimetypeById($file['mimetype']);
+				$file['mimepart'] = $this->mimetypeLoader->getMimetypeById($file['mimepart']);
 				if ($file['storage_mtime'] == 0) {
 					$file['storage_mtime'] = $file['mtime'];
 				}
@@ -385,9 +322,9 @@ class Cache {
 					$params[] = md5($value);
 					$queryParts[] = '`path_hash`';
 				} elseif ($name === 'mimetype') {
-					$params[] = $this->getMimetypeId(substr($value, 0, strpos($value, '/')));
+					$params[] = $this->mimetypeLoader->getId(substr($value, 0, strpos($value, '/')));
 					$queryParts[] = '`mimepart`';
-					$value = $this->getMimetypeId($value);
+					$value = $this->mimetypeLoader->getId($value);
 				} elseif ($name === 'storage_mtime') {
 					if (!isset($data['mtime'])) {
 						$params[] = $value;
@@ -613,7 +550,6 @@ class Cache {
 	 * @return array an array of cache entries where the name matches the search pattern
 	 */
 	public function search($pattern) {
-
 		// normalize pattern
 		$pattern = $this->normalize($pattern);
 
@@ -630,8 +566,8 @@ class Cache {
 
 		$files = array();
 		while ($row = $result->fetchRow()) {
-			$row['mimetype'] = $this->getMimetype($row['mimetype']);
-			$row['mimepart'] = $this->getMimetype($row['mimepart']);
+			$row['mimetype'] = $this->mimetypeLoader->getMimetypeById($row['mimetype']);
+			$row['mimepart'] = $this->mimetypeLoader->getMimetypeById($row['mimepart']);
 			$files[] = $row;
 		}
 		return $files;
@@ -652,12 +588,12 @@ class Cache {
 		}
 		$sql = 'SELECT `fileid`, `storage`, `path`, `parent`, `name`, `mimetype`, `mimepart`, `size`, `mtime`, `encrypted`, `etag`, `permissions`
 				FROM `*PREFIX*filecache` WHERE ' . $where . ' AND `storage` = ?';
-		$mimetype = $this->getMimetypeId($mimetype);
+		$mimetype = $this->mimetypeLoader->getId($mimetype);
 		$result = \OC_DB::executeAudited($sql, array($mimetype, $this->getNumericStorageId()));
 		$files = array();
 		while ($row = $result->fetchRow()) {
-			$row['mimetype'] = $this->getMimetype($row['mimetype']);
-			$row['mimepart'] = $this->getMimetype($row['mimepart']);
+			$row['mimetype'] = $this->mimetypeLoader->getMimetypeById($row['mimetype']);
+			$row['mimepart'] = $this->mimetypeLoader->getMimetypeById($row['mimepart']);
 			$files[] = $row;
 		}
 		return $files;
diff --git a/lib/private/files/type/detection.php b/lib/private/files/type/detection.php
index ba286637df3..3dc3975fb2a 100644
--- a/lib/private/files/type/detection.php
+++ b/lib/private/files/type/detection.php
@@ -6,6 +6,7 @@
  * @author Robin Appelman <icewind@owncloud.com>
  * @author Roeland Jago Douma <roeland@famdouma.nl>
  * @author Thomas Tanghus <thomas@tanghus.net>
+ * @author Robin McCorkell <rmccorkell@owncloud.com>
  *
  * @copyright Copyright (c) 2015, ownCloud, Inc.
  * @license AGPL-3.0
@@ -101,16 +102,23 @@ class Detection implements IMimeTypeDetector {
 			return;
 		}
 
-		$file = file_get_contents($this->configDir . '/mimetypealiases.dist.json');
-		$this->mimeTypeAlias = get_object_vars(json_decode($file));
+		$this->mimeTypeAlias = json_decode(file_get_contents($this->configDir . '/mimetypealiases.dist.json'), true);
 
 		if (file_exists($this->configDir . '/mimetypealiases.json')) {
-			$custom = get_object_vars(json_decode(file_get_contents($this->configDir . '/mimetypealiases.json')));
+			$custom = json_decode(file_get_contents($this->configDir . '/mimetypealiases.json'), true);
 			$this->mimeTypeAlias = array_merge($this->mimeTypeAlias, $custom);
 		}
 	}
 
 	/**
+	 * @return array
+	 */
+	public function getAllAliases() {
+		$this->loadAliases();
+		return $this->mimeTypeAlias;
+	}
+
+	/**
 	 * Add mimetype mappings if they are not yet present
 	 */
 	private function loadMappings() {
@@ -118,20 +126,26 @@ class Detection implements IMimeTypeDetector {
 			return;
 		}
 
-		$dist = file_get_contents($this->configDir . '/mimetypemapping.dist.json');
-		$mimetypemapping = get_object_vars(json_decode($dist));
+		$mimetypemapping = json_decode(file_get_contents($this->configDir . '/mimetypemapping.dist.json'), true);
 
 		//Check if need to load custom mappings
 		if (file_exists($this->configDir . '/mimetypemapping.json')) {
-			$custom = file_get_contents($this->configDir . '/mimetypemapping.json');
-			$custom_mapping = get_object_vars(json_decode($custom));
-			$mimetypemapping = array_merge($mimetypemapping, $custom_mapping);
+			$custom = json_decode(file_get_contents($this->configDir . '/mimetypemapping.json'), true);
+			$mimetypemapping = array_merge($mimetypemapping, $custom);
 		}
 
 		$this->registerTypeArray($mimetypemapping);
 	}
 
 	/**
+	 * @return array
+	 */
+	public function getAllMappings() {
+		$this->loadMappings();
+		return $this->mimetypes;
+	}
+
+	/**
 	 * detect mimetype only based on filename, content of file is not used
 	 *
 	 * @param string $path
diff --git a/lib/private/files/type/loader.php b/lib/private/files/type/loader.php
new file mode 100644
index 00000000000..df893306615
--- /dev/null
+++ b/lib/private/files/type/loader.php
@@ -0,0 +1,165 @@
+<?php
+/**
+ * @author Robin McCorkell <rmccorkell@owncloud.com>
+ *
+ * @copyright Copyright (c) 2015, ownCloud, Inc.
+ * @license AGPL-3.0
+ *
+ * This code is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU Affero General Public License, version 3,
+ * as published by the Free Software Foundation.
+ *
+ * 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, version 3,
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>
+ *
+ */
+
+namespace OC\Files\Type;
+
+use OCP\Files\IMimeTypeLoader;
+use OCP\IDBConnection;
+
+use Doctrine\DBAL\Exception\UniqueConstraintViolationException;
+
+/**
+ * Mimetype database loader
+ *
+ * @package OC\Files\Type
+ */
+class Loader implements IMimeTypeLoader {
+
+	/** @var IDBConnection */
+	private $dbConnection;
+
+	/** @var array [id => mimetype] */
+	protected $mimetypes;
+
+	/** @var array [mimetype => id] */
+	protected $mimetypeIds;
+
+	/**
+	 * @param IDBConnection $dbConnection
+	 */
+	public function __construct(IDBConnection $dbConnection) {
+		$this->dbConnection = $dbConnection;
+		$this->mimetypes = [];
+		$this->mimetypeIds = [];
+	}
+
+	/**
+	 * Get a mimetype from its ID
+	 *
+	 * @param int $id
+	 * @return string|null
+	 */
+	public function getMimetypeById($id) {
+		if (!$this->mimetypes) {
+			$this->loadMimetypes();
+		}
+		if (isset($this->mimetypes[$id])) {
+			return $this->mimetypes[$id];
+		}
+		return null;
+	}
+
+	/**
+	 * Get a mimetype ID, adding the mimetype to the DB if it does not exist
+	 *
+	 * @param string $mimetype
+	 * @return int
+	 */
+	public function getId($mimetype) {
+		if (!$this->mimetypeIds) {
+			$this->loadMimetypes();
+		}
+		if (isset($this->mimetypeIds[$mimetype])) {
+			return $this->mimetypeIds[$mimetype];
+		}
+		return $this->store($mimetype);
+	}
+
+	/**
+	 * Test if a mimetype exists in the database
+	 *
+	 * @param string $mimetype
+	 * @return bool
+	 */
+	public function exists($mimetype) {
+		if (!$this->mimetypeIds) {
+			$this->loadMimetypes();
+		}
+		return isset($this->mimetypeIds[$mimetype]);
+	}
+
+	/**
+	 * Store a mimetype in the DB
+	 *
+	 * @param string $mimetype
+	 * @param int inserted ID
+	 */
+	protected function store($mimetype) {
+		try {
+			$qb = $this->dbConnection->getQueryBuilder();
+			$qb->insert('mimetypes')
+				->values([
+					'mimetype' => $qb->createNamedParameter($mimetype)
+				]);
+			$qb->execute();
+		} catch (UniqueConstraintViolationException $e) {
+			// something inserted it before us
+		}
+
+		$fetch = $this->dbConnection->getQueryBuilder();
+		$fetch->select('id')
+			->from('mimetypes')
+			->where(
+				$fetch->expr()->eq('mimetype', $fetch->createNamedParameter($mimetype)
+			));
+		$row = $fetch->execute()->fetch();
+
+		$this->mimetypes[$row['id']] = $mimetype;
+		$this->mimetypeIds[$mimetype] = $row['id'];
+		return $row['id'];
+	}
+
+	/**
+	 * Load all mimetypes from DB
+	 */
+	private function loadMimetypes() {
+		$qb = $this->dbConnection->getQueryBuilder();
+		$qb->select('id', 'mimetype')
+			->from('mimetypes');
+		$results = $qb->execute()->fetchAll();
+
+		foreach ($results as $row) {
+			$this->mimetypes[$row['id']] = $row['mimetype'];
+			$this->mimetypeIds[$row['mimetype']] = $row['id'];
+		}
+	}
+
+	/**
+	 * Update filecache mimetype based on file extension
+	 *
+	 * @param string $ext file extension
+	 * @param int $mimetypeId
+	 * @return int number of changed rows
+	 */
+	public function updateFilecache($ext, $mimetypeId) {
+		$update = $this->dbConnection->getQueryBuilder();
+		$update->update('filecache')
+			->set('mimetype', $update->createNamedParameter($mimetypeId))
+			->where($update->expr()->neq(
+				'mimetype', $update->createNamedParameter($mimetypeId)
+			))
+			->andWhere($update->expr()->like(
+				$update->createFunction('LOWER(`name`)'), $update->createNamedParameter($ext)
+			));
+		return $update->execute();
+	}
+
+}
diff --git a/lib/private/server.php b/lib/private/server.php
index 24674d2e3c7..393c1840973 100644
--- a/lib/private/server.php
+++ b/lib/private/server.php
@@ -470,6 +470,11 @@ class Server extends SimpleContainer implements IServerContainer {
 				$c->getURLGenerator(),
 				\OC::$configDir);
 		});
+		$this->registerService('MimeTypeLoader', function(Server $c) {
+			return new \OC\Files\Type\Loader(
+				$c->getDatabaseConnection()
+			);
+		});
 		$this->registerService('CapabilitiesManager', function (Server $c) {
 			$manager = new \OC\CapabilitiesManager();
 			$manager->registerCapability(function() use ($c) {
@@ -1011,6 +1016,15 @@ class Server extends SimpleContainer implements IServerContainer {
 	}
 
 	/**
+	 * Get the MimeTypeLoader
+	 *
+	 * @return \OCP\Files\IMimeTypeLoader
+	 */
+	public function getMimeTypeLoader() {
+		return $this->query('MimeTypeLoader');
+	}
+
+	/**
 	 * Get the manager of all the capabilities
 	 *
 	 * @return \OC\CapabilitiesManager