Add a metadata service to store file metadata

Signed-off-by: Carl Schwan <carl@carlschwan.eu>

10 files changed, 512 insertions, 5 deletions

diff --git a/lib/private/Files/ObjectStore/NoopScanner.php b/lib/private/Files/ObjectStore/NoopScanner.php
index 42e212271d5..3b8cbdb18bb 100644
--- a/lib/private/Files/ObjectStore/NoopScanner.php
+++ b/lib/private/Files/ObjectStore/NoopScanner.php
@@ -31,7 +31,7 @@ use OC\Files\Storage\Storage;
 
 class NoopScanner extends Scanner {
 	public function __construct(Storage $storage) {
-		//we don't need the storage, so do nothing here
+		// we don't need the storage, so do nothing here
 	}
 
 	/**
diff --git a/lib/private/Metadata/Capabilities.php b/lib/private/Metadata/Capabilities.php
new file mode 100644
index 00000000000..2fa0006f581
--- /dev/null
+++ b/lib/private/Metadata/Capabilities.php
@@ -0,0 +1,44 @@
+<?php
+
+declare(strict_types=1);
+
+/**
+ * @copyright Copyright 2022 Carl Schwan <carl@carlschwan.eu>
+ * @license AGPL-3.0-or-later
+ *
+ * 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\Metadata;
+
+use OCP\Capabilities\IPublicCapability;
+use OCP\IConfig;
+
+class Capabilities implements IPublicCapability {
+	private IMetadataManager $manager;
+	private IConfig $config;
+
+	public function __construct(IMetadataManager $manager, IConfig $config) {
+		$this->manager = $manager;
+		$this->config = $config;
+	}
+
+	public function getCapabilities() {
+		if ($this->config->getSystemValueBool('enable_file_metadata', true)) {
+			return ['metadataAvailable' => $this->manager->getCapabilities()];
+		}
+
+		return [];
+	}
+}
diff --git a/lib/private/Metadata/FileEventListener.php b/lib/private/Metadata/FileEventListener.php
new file mode 100644
index 00000000000..fdec891c6e2
--- /dev/null
+++ b/lib/private/Metadata/FileEventListener.php
@@ -0,0 +1,84 @@
+<?php
+
+declare(strict_types=1);
+/**
+ * @copyright Copyright 2022 Carl Schwan <carl@carlschwan.eu>
+ * @license AGPL-3.0-or-later
+ *
+ * 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\Metadata;
+
+use OC\Files\Filesystem;
+use OCP\EventDispatcher\Event;
+use OCP\EventDispatcher\IEventListener;
+use OCP\Files\Events\Node\NodeDeletedEvent;
+use OCP\Files\Events\Node\NodeWrittenEvent;
+use OCP\Files\Events\NodeRemovedFromCache;
+use OCP\Files\File;
+use OCP\Files\Node;
+use OCP\Files\NotFoundException;
+use OCP\Files\FileInfo;
+
+class FileEventListener implements IEventListener {
+	private IMetadataManager $manager;
+
+	public function __construct(IMetadataManager $manager) {
+		$this->manager = $manager;
+	}
+
+	private function shouldExtractMetadata(Node $node): bool {
+		try {
+			if ($node->getMimetype() === 'httpd/unix-directory') {
+				return false;
+			}
+		} catch (NotFoundException $e) {
+			return false;
+		}
+		if ($node->getSize(false) <= 0) {
+			return false;
+		}
+
+		$path = $node->getPath();
+		// TODO make this more dynamic, we have the same issue in other places
+		return !str_starts_with($path, 'appdata_') && !str_starts_with($path, 'files_versions/') && !str_starts_with($path, 'files_trashbin/');
+	}
+
+	public function handle(Event $event): void {
+		if ($event instanceof NodeRemovedFromCache) {
+			$view = Filesystem::getView();
+			$info = $view->getFileInfo($event->getPath());
+			if ($info && $info->getType() === FileInfo::TYPE_FILE) {
+				$this->manager->clearMetadata($info->getId());
+			}
+		}
+
+		if ($event instanceof NodeDeletedEvent) {
+			$node = $event->getNode();
+			if ($this->shouldExtractMetadata($node)) {
+				/** @var File $node */
+				$this->manager->clearMetadata($event->getNode()->getId());
+			}
+		}
+
+		if ($event instanceof NodeWrittenEvent) {
+			$node = $event->getNode();
+			if ($this->shouldExtractMetadata($node)) {
+				/** @var File $node */
+				$this->manager->generateMetadata($event->getNode(), false);
+			}
+		}
+	}
+}
diff --git a/lib/private/Metadata/FileMetadata.php b/lib/private/Metadata/FileMetadata.php
new file mode 100644
index 00000000000..c53f5d7f619
--- /dev/null
+++ b/lib/private/Metadata/FileMetadata.php
@@ -0,0 +1,43 @@
+<?php
+
+declare(strict_types=1);
+/**
+ * @copyright Copyright 2022 Carl Schwan <carl@carlschwan.eu>
+ *
+ * @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\Metadata;
+
+use OCP\AppFramework\Db\Entity;
+use OCP\DB\Types;
+
+/**
+ * @method string getGroupName()
+ * @method void setGroupName(string $groupName)
+ * @method string getMetadata()
+ * @method void setMetadata(array $metadata)
+ * @see OC\Core\Migrations\Version240000Date20220404230027
+ */
+class FileMetadata extends Entity {
+	protected ?string $groupName = null;
+	protected ?array $metadata = null;
+
+	public function __construct() {
+		$this->addType('groupName', 'string');
+		$this->addType('metadata', Types::JSON);
+	}
+}
diff --git a/lib/private/Metadata/FileMetadataMapper.php b/lib/private/Metadata/FileMetadataMapper.php
new file mode 100644
index 00000000000..53f750ae540
--- /dev/null
+++ b/lib/private/Metadata/FileMetadataMapper.php
@@ -0,0 +1,105 @@
+<?php
+
+declare(strict_types=1);
+/**
+ * @copyright Copyright 2022 Carl Schwan <carl@carlschwan.eu>
+ * @license AGPL-3.0-or-later
+ *
+ * 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\Metadata;
+
+use OCP\AppFramework\Db\DoesNotExistException;
+use OCP\AppFramework\Db\MultipleObjectsReturnedException;
+use OCP\AppFramework\Db\QBMapper;
+use OCP\DB\Exception;
+use OCP\DB\QueryBuilder\IQueryBuilder;
+use OCP\IDBConnection;
+
+class FileMetadataMapper extends QBMapper {
+	public function __construct(IDBConnection $db) {
+		parent::__construct($db, 'file_metadata', FileMetadata::class);
+	}
+
+	/**
+	 * @return FileMetadata[]
+	 * @throws Exception
+	 */
+	public function findForFile(int $fileId): array {
+		$qb = $this->db->getQueryBuilder();
+		$qb->select('*')
+			->from($this->getTableName())
+			->where($qb->expr()->eq('id', $qb->createNamedParameter($fileId, IQueryBuilder::PARAM_INT)));
+
+		return $this->findEntities($qb);
+	}
+
+	/**
+	 * @throws DoesNotExistException
+	 * @throws MultipleObjectsReturnedException
+	 * @throws Exception
+	 */
+	public function findForGroupForFile(int $fileId, string $groupName): FileMetadata {
+		$qb = $this->db->getQueryBuilder();
+		$qb->select('*')
+			->from($this->getTableName())
+			->where($qb->expr()->eq('id', $qb->createNamedParameter($fileId, IQueryBuilder::PARAM_INT_ARRAY)))
+			->andWhere($qb->expr()->eq('group_name', $qb->createNamedParameter($groupName, IQueryBuilder::PARAM_STR)));
+
+		return $this->findEntity($qb);
+	}
+
+	/**
+	 * @return array<int, FileMetadata>
+	 * @throws Exception
+	 */
+	public function findForGroupForFiles(array $fileIds, string $groupName): array {
+		$qb = $this->db->getQueryBuilder();
+		$qb->select('*')
+			->from($this->getTableName())
+			->where($qb->expr()->in('id', $qb->createParameter('fileIds')))
+			->andWhere($qb->expr()->eq('group_name', $qb->createNamedParameter($groupName, IQueryBuilder::PARAM_STR)));
+
+		$metadata = [];
+		foreach (array_chunk($fileIds, 1000) as $fileIdsChunk) {
+			$qb->setParameter('fileIds', $fileIdsChunk, IQueryBuilder::PARAM_INT_ARRAY);
+			/** @var FileMetadata[] $rawEntities */
+			$rawEntities = $this->findEntities($qb);
+			foreach ($rawEntities as $entity) {
+				$metadata[$entity->getId()] = $entity;
+			}
+		}
+
+		foreach ($fileIds as $id) {
+			if (isset($metadata[$id])) {
+				continue;
+			}
+			$empty = new FileMetadata();
+			$empty->setMetadata([]);
+			$empty->setGroupName($groupName);
+			$empty->setId($id);
+			$metadata[$id] = $empty;
+		}
+		return $metadata;
+	}
+
+	public function clear(int $fileId): void {
+		$qb = $this->db->getQueryBuilder();
+		$qb->delete($this->getTableName())
+			->where($qb->expr()->eq('id', $qb->createNamedParameter($fileId, IQueryBuilder::PARAM_INT)));
+
+		$qb->executeStatement();
+	}
+}
diff --git a/lib/private/Metadata/IMetadataManager.php b/lib/private/Metadata/IMetadataManager.php
new file mode 100644
index 00000000000..d2d37f15c25
--- /dev/null
+++ b/lib/private/Metadata/IMetadataManager.php
@@ -0,0 +1,35 @@
+<?php
+
+declare(strict_types=1);
+
+namespace OC\Metadata;
+
+use OCP\Files\File;
+
+/**
+ * Interface to manage additional metadata for files
+ */
+interface IMetadataManager {
+	/**
+	 * @param class-string<IMetadataProvider> $className
+	 */
+	public function registerProvider(string $className): void;
+
+	/**
+	 * Generate the metadata for one file
+	 */
+	public function generateMetadata(File $file, bool $checkExisting = false): void;
+
+	/**
+	 * Clear the metadata for one file
+	 */
+	public function clearMetadata(int $fileId): void;
+
+	/** @return array<int, FileMetadata> */
+	public function fetchMetadataFor(string $group, array $fileIds): array;
+
+	/**
+	 * Get the capabilites as an array of mimetype regex to the type provided
+	 */
+	public function getCapabilities(): array;
+}
diff --git a/lib/private/Metadata/IMetadataProvider.php b/lib/private/Metadata/IMetadataProvider.php
new file mode 100644
index 00000000000..7cbe102a538
--- /dev/null
+++ b/lib/private/Metadata/IMetadataProvider.php
@@ -0,0 +1,41 @@
+<?php
+
+namespace OC\Metadata;
+
+use OCP\Files\File;
+
+/**
+ * Interface for the metadata providers. If you want an application to provide
+ * some metadata, you can use this to store them.
+ */
+interface IMetadataProvider {
+	/**
+	 * The list of groups that this metadata provider is able to provide.
+	 *
+	 * @return string[]
+	 */
+	public static function groupsProvided(): array;
+
+	/**
+	 * Check if the metadata provider is available. A metadata provider might be
+	 * unavailable due to a php extension not being installed.
+	 */
+	public static function isAvailable(): bool;
+
+	/**
+	 * Get the mimetypes supported as a regex.
+	 */
+	public static function getMimetypesSupported(): string;
+
+	/**
+	 * Execute the extraction on the specified file. The metadata should be
+	 * grouped by metadata
+	 *
+	 * Each group should be json serializable and the string representation
+	 * shouldn't be longer than 4000 characters.
+	 *
+	 * @param File $file The file to extract the metadata from
+	 * @param array<string, FileMetadata> An array containing all the metadata fetched.
+	 */
+	public function execute(File $file): array;
+}
diff --git a/lib/private/Metadata/MetadataManager.php b/lib/private/Metadata/MetadataManager.php
new file mode 100644
index 00000000000..69e9cb3c852
--- /dev/null
+++ b/lib/private/Metadata/MetadataManager.php
@@ -0,0 +1,100 @@
+<?php
+/**
+ * @copyright Copyright 2022 Carl Schwan <carl@carlschwan.eu>
+ * @license AGPL-3.0-or-later
+ *
+ * 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\Metadata;
+
+use OC\Metadata\Provider\ExifProvider;
+use OCP\Files\File;
+use OCP\IConfig;
+use Psr\Log\LoggerInterface;
+
+class MetadataManager implements IMetadataManager {
+	/** @var array<string, IMetadataProvider> */
+	private array $providers;
+	private array $providerClasses;
+	private FileMetadataMapper $fileMetadataMapper;
+	private IConfig $config;
+	private LoggerInterface $logger;
+
+	public function __construct(
+		FileMetadataMapper $fileMetadataMapper,
+		IConfig $config,
+		LoggerInterface $logger
+	) {
+		$this->providers = [];
+		$this->providerClasses = [];
+		$this->fileMetadataMapper = $fileMetadataMapper;
+		$this->config = $config;
+		$this->logger = $logger;
+
+		// TODO move to another place, where?
+		$this->registerProvider(ExifProvider::class);
+	}
+
+	/**
+	 * @param class-string<IMetadataProvider> $className
+	 */
+	public function registerProvider(string $className):void {
+		if (in_array($className, $this->providerClasses)) {
+			return;
+		}
+
+		if (call_user_func([$className, 'isAvailable'])) {
+			$this->providers[call_user_func([$className, 'getMimetypesSupported'])] = \OC::$server->get($className);
+		}
+	}
+
+	public function generateMetadata(File $file, bool $checkExisting = false): void {
+		$existingMetadataGroups = [];
+
+		if ($checkExisting) {
+			$existingMetadata = $this->fileMetadataMapper->findForFile($file->getId());
+			foreach ($existingMetadata as $metadata) {
+				$existingMetadataGroups[] = $metadata->getGroupName();
+			}
+		}
+
+		foreach ($this->providers as $supportedMimetype => $provider) {
+			if (preg_match($supportedMimetype, $file->getMimeType())) {
+				if (count(array_diff($provider::groupsProvided(), $existingMetadataGroups)) > 0) {
+					$metaDataGroup = $provider->execute($file);
+					foreach ($metaDataGroup as $group => $metadata) {
+						$this->fileMetadataMapper->insertOrUpdate($metadata);
+					}
+				}
+			}
+		}
+	}
+
+	public function clearMetadata(int $fileId): void {
+		$this->fileMetadataMapper->clear($fileId);
+	}
+
+	public function fetchMetadataFor(string $group, array $fileIds): array {
+		return $this->fileMetadataMapper->findForGroupForFiles($fileIds, $group);
+	}
+
+	public function getCapabilities(): array {
+		$capabilities = [];
+		foreach ($this->providers as $supportedMimetype => $provider) {
+			$capabilities[$supportedMimetype] = $provider::groupsProvided();
+		}
+		return $capabilities;
+	}
+}
diff --git a/lib/private/Metadata/Provider/ExifProvider.php b/lib/private/Metadata/Provider/ExifProvider.php
new file mode 100644
index 00000000000..91c858f6794
--- /dev/null
+++ b/lib/private/Metadata/Provider/ExifProvider.php
@@ -0,0 +1,51 @@
+<?php
+
+namespace OC\Metadata\Provider;
+
+use OC\Metadata\FileMetadata;
+use OC\Metadata\IMetadataProvider;
+use OCP\Files\File;
+
+class ExifProvider implements IMetadataProvider {
+	public static function groupsProvided(): array {
+		return ['size'];
+	}
+
+	public static function isAvailable(): bool {
+		return extension_loaded('exif');
+	}
+
+	public function execute(File $file): array {
+		$fileDescriptor = $file->fopen('rb');
+		$data = @exif_read_data($fileDescriptor, 'ANY_TAG', true);
+
+		$size = new FileMetadata();
+		$size->setGroupName('size');
+		$size->setId($file->getId());
+		$size->setMetadata([]);
+
+		if (!$data) {
+			return [
+				'size' => $size,
+			];
+		}
+
+		if (array_key_exists('COMPUTED', $data)
+			&& array_key_exists('Width', $data['COMPUTED'])
+			&& array_key_exists('Height', $data['COMPUTED'])
+		) {
+			$size->setMetadata([
+				'width' => $data['COMPUTED']['Width'],
+				'height' => $data['COMPUTED']['Height'],
+			]);
+		}
+
+		return [
+			'size' => $size,
+		];
+	}
+
+	public static function getMimetypesSupported(): string {
+		return '/image\/.*/';
+	}
+}
diff --git a/lib/private/Server.php b/lib/private/Server.php
index 7817d1beafe..e9d673d3746 100644
--- a/lib/private/Server.php
+++ b/lib/private/Server.php
@@ -122,6 +122,9 @@ use OC\Log\PsrLoggerAdapter;
 use OC\Mail\Mailer;
 use OC\Memcache\ArrayCache;
 use OC\Memcache\Factory;
+use OC\Metadata\Capabilities as MetadataCapabilities;
+use OC\Metadata\IMetadataManager;
+use OC\Metadata\MetadataManager;
 use OC\Notification\Manager;
 use OC\OCS\DiscoveryService;
 use OC\Preview\GeneratorHelper;
@@ -151,7 +154,6 @@ use OC\Template\JSCombiner;
 use OCA\Theming\ImageManager;
 use OCA\Theming\ThemingDefaults;
 use OCA\Theming\Util;
-use OCA\WorkflowEngine\Service\Logger;
 use OCP\Accounts\IAccountManager;
 use OCP\App\IAppManager;
 use OCP\Authentication\LoginCredentials\IStore;
@@ -241,15 +243,12 @@ use OCP\SystemTag\ISystemTagManager;
 use OCP\SystemTag\ISystemTagObjectMapper;
 use OCP\Talk\IBroker;
 use OCP\User\Events\BeforePasswordUpdatedEvent;
-use OCP\User\Events\BeforeUserCreatedEvent;
-use OCP\User\Events\BeforeUserDeletedEvent;
 use OCP\User\Events\BeforeUserLoggedInEvent;
 use OCP\User\Events\BeforeUserLoggedInWithCookieEvent;
 use OCP\User\Events\BeforeUserLoggedOutEvent;
 use OCP\User\Events\PasswordUpdatedEvent;
 use OCP\User\Events\PostLoginEvent;
 use OCP\User\Events\UserChangedEvent;
-use OCP\User\Events\UserDeletedEvent;
 use OCP\User\Events\UserLoggedInEvent;
 use OCP\User\Events\UserLoggedInWithCookieEvent;
 use OCP\User\Events\UserLoggedOutEvent;
@@ -1163,6 +1162,9 @@ class Server extends ServerContainer implements IServerContainer {
 			$manager->registerCapability(function () use ($c) {
 				return $c->get(\OC\Security\Bruteforce\Capabilities::class);
 			});
+			$manager->registerCapability(function () use ($c) {
+				return $c->get(MetadataCapabilities::class);
+			});
 			return $manager;
 		});
 		/** @deprecated 19.0.0 */
@@ -1433,6 +1435,8 @@ class Server extends ServerContainer implements IServerContainer {
 
 		$this->registerAlias(IBroker::class, Broker::class);
 
+		$this->registerAlias(IMetadataManager::class, MetadataManager::class);
+
 		$this->connectDispatcher();
 	}