feat: Implement background job to remove unused models

Signed-off-by: David Dreschner <david.dreschner@nextcloud.com>

Author: DerDreschner

appinfo/info.xml

@@ -32,6 +32,7 @@
 		<job>OCA\SuspiciousLogin\BackgroundJob\ETLJob</job>
 		<job>OCA\SuspiciousLogin\BackgroundJob\TrainJobIpV4</job>
 		<job>OCA\SuspiciousLogin\BackgroundJob\TrainJobIpV6</job>
+		<job>OCA\SuspiciousLogin\BackgroundJob\CleanupModelsJob</job>
 	</background-jobs>
 
 	<commands>

lib/BackgroundJob/CleanupModelsJob.php

@@ -0,0 +1,50 @@
+<?php
+
+declare(strict_types=1);
+
+/**
+ * SPDX-FileCopyrightText: 2026 Nextcloud GmbH and Nextcloud contributors
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+
+namespace OCA\SuspiciousLogin\BackgroundJob;
+
+use OCA\SuspiciousLogin\Service\ModelStore;
+use OCP\AppFramework\Utility\ITimeFactory;
+use OCP\BackgroundJob\TimedJob;
+use Psr\Log\LoggerInterface;
+use Throwable;
+
+class CleanupModelsJob extends TimedJob {
+
+	/**
+	 * @var int Number of most recent models to keep per address type.
+	 *          Should match the number of models consumed by the statistics
+	 *          API.
+	 */
+	private const MODELS_TO_KEEP = 14;
+
+	public function __construct(
+		private readonly ModelStore $modelStore,
+		private readonly LoggerInterface $logger,
+		ITimeFactory $time,
+	) {
+		parent::__construct($time);
+
+		// Run once per day
+		$this->setInterval(24 * 60 * 60);
+		$this->setTimeSensitivity(self::TIME_INSENSITIVE);
+		$this->allowParallelRuns = false;
+	}
+
+	#[\Override]
+	protected function run($argument): void {
+		try {
+			$this->modelStore->cleanup(self::MODELS_TO_KEEP);
+		} catch (Throwable $e) {
+			$this->logger->error('Error during model cleanup: ' . $e->getMessage(), [
+				'exception' => $e,
+			]);
+		}
+	}
+}

lib/Db/ModelMapper.php

@@ -66,4 +66,20 @@ public function findMostRecent(int $max, string $addressType): array {
 
 		return $this->findEntities($query);
 	}
+
+	/**
+	 * Find all models older than `$numberOfModelsToKeep` per `$addressType`
+	 *
+	 * @return Model[]
+	 */
+	public function findOld(int $numberOfModelsToKeep, string $addressType): array {
+		$qb = $this->db->getQueryBuilder();
+		$query = $qb->select('*')
+			->from($this->getTableName())
+			->where($qb->expr()->eq('address_type', $qb->createNamedParameter($addressType)))
+			->setFirstResult($numberOfModelsToKeep)
+			->orderBy('created_at', 'desc');
+
+		return $this->findEntities($query);
+	}
 }

lib/Service/ModelStore.php

@@ -165,4 +165,59 @@ public function persist(Learner $estimator, Model $model) {
 			throw $e;
 		}
 	}
+
+	/**
+	 * Remove old models, keeping only the `$numberOfModelsToKeep` ones
+	 * for each used address type (IPv4 / IPv6).
+	 *
+	 * @throws \OCP\DB\Exception
+	 * @throws \OCP\Files\NotPermittedException
+	 */
+	public function cleanup(int $numberOfModelsToKeep): void {
+		$oldModels = array_merge(
+			$this->modelMapper->findOld($numberOfModelsToKeep, Ipv4Strategy::getTypeName()),
+			$this->modelMapper->findOld($numberOfModelsToKeep, IpV6Strategy::getTypeName())
+		);
+
+		if (empty($oldModels)) {
+			$this->logger->debug('No old models to clean up');
+			return;
+		}
+
+		try {
+			$modelsFolder = $this->appData->getFolder(self::APPDATA_MODELS_FOLDER);
+		} catch (NotFoundException) {
+			$this->logger->debug('Models folder does not exist, skipping file deletion');
+		}
+
+		if ($this->cacheFactory->isLocalCacheAvailable()) {
+			$cache = $this->cacheFactory->createLocal();
+		}
+
+		foreach ($oldModels as $oldModel) {
+			$id = $oldModel->getId();
+			$this->logger->debug("Cleaning up old model $id");
+
+			// Remove the model file from app data
+			if (isset($modelsFolder)) {
+				try {
+					$modelFile = $modelsFolder->getFile((string)$id);
+					$modelFile->delete();
+				} catch (NotFoundException) {
+					$this->logger->debug("Model file $id not found in app data, skipping file deletion");
+				}
+			}
+
+			// Evict from cache
+			if (isset($cache)) {
+				$cache->remove($this->getCacheKey($id));
+			}
+
+			// Remove the DB record
+			$this->modelMapper->delete($oldModel);
+		}
+
+		$numberOfOldModels = count($oldModels);
+		$this->logger->info("Cleaned up {$numberOfOldModels} old model(s)");
+	}
 }