Gracefully handle already done ocr

lib/Exception/OcrAlreadyDoneException.php

@@ -0,0 +1,32 @@
+<?php
+
+declare(strict_types=1);
+
+/**
+ * @copyright Copyright (c) 2025 Robin Windey <ro.windey@gmail.com>
+ *
+ *  @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 <http://www.gnu.org/licenses/>.
+ */
+
+namespace OCA\WorkflowOcr\Exception;
+
+use Exception;
+
+class OcrAlreadyDoneException extends Exception {
+	public function __construct(string $message) {
+		$this->message = $message;
+	}
+}

lib/OcrProcessors/Local/OcrMyPdfBasedProcessor.php

@@ -23,6 +23,7 @@
 
 namespace OCA\WorkflowOcr\OcrProcessors\Local;
 
+use OCA\WorkflowOcr\Exception\OcrAlreadyDoneException;
 use OCA\WorkflowOcr\Exception\OcrNotPossibleException;
 use OCA\WorkflowOcr\Exception\OcrResultEmptyException;
 use OCA\WorkflowOcr\Helper\ISidecarFileAccessor;
@@ -63,6 +64,10 @@ public function ocrFile(File $file, WorkflowSettings $settings, GlobalSettings $
 		$exitCode = $this->command->getExitCode();
 
 		if (!$success) {
+			# Gracefully handle OCR_MODE_SKIP_FILE (ExitCode.already_done_ocr)
+			if ($exitCode === 6) {
+				throw new OcrAlreadyDoneException('File ' . $file->getPath() . ' appears to contain text so it may not need OCR. Message: ' . $errorOutput . ' ' . $stdErr);
+			}
 			throw new OcrNotPossibleException('OCRmyPDF exited abnormally with exit-code ' . $exitCode . ' for file ' . $file->getPath() . '. Message: ' . $errorOutput . ' ' . $stdErr);
 		}
 

lib/OcrProcessors/Remote/WorkflowOcrRemoteProcessor.php

@@ -23,6 +23,7 @@
 
 namespace OCA\WorkflowOcr\OcrProcessors\Remote;
 
+use OCA\WorkflowOcr\Exception\OcrAlreadyDoneException;
 use OCA\WorkflowOcr\Exception\OcrNotPossibleException;
 use OCA\WorkflowOcr\Model\GlobalSettings;
 use OCA\WorkflowOcr\Model\WorkflowSettings;
@@ -56,7 +57,14 @@ public function ocrFile(File $file, WorkflowSettings $settings, GlobalSettings $
 		$this->logger->debug('OCR result received', ['apiResult' => $apiResult]);
 
 		if ($apiResult instanceof ErrorResult) {
-			throw new OcrNotPossibleException($apiResult->getMessage());
+			$resultMessage = $apiResult->getMessage();
+			$exitCode = $apiResult->getOcrMyPdfExitCode();
+
+			# Gracefully handle OCR_MODE_SKIP_FILE (ExitCode.already_done_ocr)
+			if ($exitCode === 6) {
+				throw new OcrAlreadyDoneException('File ' . $file->getPath() . ' appears to contain text so it may not need OCR. Message: ' . $resultMessage);
+			}
+			throw new OcrNotPossibleException($resultMessage);
 		}
 
 		return new OcrProcessorResult(

lib/Service/OcrService.php

@@ -30,7 +30,7 @@
 use OCA\Files_Versions\Versions\IMetadataVersion;
 use OCA\Files_Versions\Versions\IMetadataVersionBackend;
 use OCA\Files_Versions\Versions\IVersionManager;
-use OCA\WorkflowOcr\Exception\OcrResultEmptyException;
+use OCA\WorkflowOcr\Exception\OcrAlreadyDoneException;
 use OCA\WorkflowOcr\Helper\IProcessingFileAccessor;
 use OCA\WorkflowOcr\Model\WorkflowSettings;
 use OCA\WorkflowOcr\OcrProcessors\IOcrProcessorFactory;
@@ -158,8 +158,8 @@ public function runOcrProcess(int $fileId, string $uid, WorkflowSettings $settin
 
 			try {
 				$result = $ocrProcessor->ocrFile($file, $settings, $globalSettings);
-			} catch (OcrResultEmptyException $ex) {
-				// #232: it's okay to have an empty result if the file was skipped due to OCR mode
+			} catch (OcrAlreadyDoneException $ex) {
+				// #232: Skip file if OCR was already done
 				if ($settings->getOcrMode() === WorkflowSettings::OCR_MODE_SKIP_FILE) {
 					$this->logger->debug('Skipping empty OCR result for file with id {fileId} because OCR mode is set to \'skip file\'', ['fileId' => $fileId]);
 					return;

tests/Integration/BackendTestBase.php

@@ -97,7 +97,7 @@ protected function checkOcrMyPdfInstalled() : bool {
 		return $command->getExitCode() === 0;
 	}
 
-	protected function addOperation(string $mimeType) {
+	protected function addOperation(string $mimeType, string $operationJson = '') {
 		// NOTE :: we're creating the workflow operation via
 		// REST API because if we'd use the manager directly, we'd
 		// face some issues because of caching etc (test ist running
@@ -118,10 +118,9 @@ protected function addOperation(string $mimeType) {
 					"invalid":false
 				}
 			],
-			"operation":"",
+			"operation":"' . str_replace('"', '\"', $operationJson) . '",
 			"valid":true
 		}';
-
 		$ch = curl_init($url);
 		curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
 		curl_setopt($ch, CURLOPT_POST, true);
@@ -135,6 +134,18 @@ protected function addOperation(string $mimeType) {
 	}
 
 	private function deleteOperation() {
+		// Clear managers cache and operations to ensure "deleteOperation" works
+		$reflection = new \ReflectionClass($this->workflowEngineManager);
+
+		$operationsByScopeProperty = $reflection->getProperty('operationsByScope');
+		$operationsByScopeProperty->setAccessible(true);
+		$operationsByScope = $operationsByScopeProperty->getValue($this->workflowEngineManager);
+		$operationsByScope->clear();
+
+		$operationsProperty = $reflection->getProperty('operations');
+		$operationsProperty->setAccessible(true);
+		$operationsProperty->setValue($this->workflowEngineManager, []);
+
 		$operations = $this->workflowEngineManager->getOperations($this->operationClass, $this->context);
 		foreach ($operations as $operation) {
 			try {

tests/Integration/IntegrationTestApiClient.php

@@ -61,4 +61,9 @@ public function getRequests(): array {
 	public function getResponses(): array {
 		return $this->responses;
 	}
+
+	public function reset(): void {
+		$this->requests = [];
+		$this->responses = [];
+	}
 }

tests/Integration/LocalBackendTest.php

@@ -24,6 +24,7 @@
 namespace OCA\WorkflowOcr\Tests\Integration;
 
 use OCA\WorkflowOcr\Events\TextRecognizedEvent;
+use OCA\WorkflowOcr\Model\WorkflowSettings;
 use OCA\WorkflowOcr\OcrProcessors\Remote\Client\IApiClient;
 use OCP\EventDispatcher\IEventDispatcher;
 
@@ -76,6 +77,7 @@ protected function setUp(): void {
 	protected function tearDown(): void {
 		parent::tearDown();
 		$this->dispatcher->removeListener(TextRecognizedEvent::class, $this->eventListener);
+		$this->capturedEvents = [];
 	}
 
 	/**
@@ -92,4 +94,33 @@ public function testWorkflowOcrLocalBackend(): void {
 		$this->assertInstanceOf(TextRecognizedEvent::class, $textRecognizedEvent, 'Expected TextRecognizedEvent instance');
 		$this->assertEquals('This document is ready for OCR', trim($textRecognizedEvent->getRecognizedText()), 'Expected recognized text');
 	}
+
+	/**
+	 * Test processing a file via ocrmypdf CLI where file already contains OCR and mode is set to "skip file".
+	 */
+	public function testWorkflowOcrLocalBackendSkipFile(): void {
+		$this->addOperation('application/pdf', '{"ocrMode":' . WorkflowSettings::OCR_MODE_SKIP_FILE . '}');
+		$this->uploadTestFile('document-has-ocr.pdf');
+		$this->runOcrBackgroundJob();
+
+		$this->assertEmpty($this->apiClient->getRequests(), 'Expected no OCR Backend Service requests');
+		$this->assertEquals(0, count($this->capturedEvents), 'Expected no TextRecognizedEvent');
+	}
+
+	/**
+	 * Test processing a file via ocrmypdf CLI where file already contains OCR and mode is set to "skip text".
+	 */
+	public function testWorkflowOcrLocalBackendSkipText(): void {
+		$this->addOperation('application/pdf', '{"ocrMode":' . WorkflowSettings::OCR_MODE_SKIP_TEXT . '}');
+		$this->uploadTestFile('document-has-ocr.pdf');
+		$this->runOcrBackgroundJob();
+
+		$this->assertEmpty($this->apiClient->getRequests(), 'Expected no OCR Backend Service requests');
+		$this->assertEquals(1, count($this->capturedEvents), 'Expected no TextRecognizedEvent');
+
+		// Ocrmypdf will "recognize" a special text for pages which already have OCR
+		$textRecognizedEvent = $this->capturedEvents[0];
+		$this->assertInstanceOf(TextRecognizedEvent::class, $textRecognizedEvent, 'Expected TextRecognizedEvent instance');
+		$this->assertEquals('[OCR skipped on page(s) 1]', trim($textRecognizedEvent->getRecognizedText()), 'Expected recognized text');
+	}
 }

tests/Integration/OcrBackendServiceTest.php

@@ -23,7 +23,9 @@
 
 namespace OCA\WorkflowOcr\Tests\Integration;
 
+use OCA\WorkflowOcr\Model\WorkflowSettings;
 use OCA\WorkflowOcr\OcrProcessors\Remote\Client\IApiClient;
+use OCA\WorkflowOcr\OcrProcessors\Remote\Client\Model\ErrorResult;
 use OCA\WorkflowOcr\OcrProcessors\Remote\Client\Model\OcrResult;
 
 /**
@@ -54,6 +56,7 @@ protected function setUp(): void {
 		}
 
 		$this->apiClient = $this->container->get(IntegrationTestApiClient::class);
+		$this->apiClient->reset();
 		$this->overwriteService(IApiClient::class, $this->apiClient);
 	}
 
@@ -77,4 +80,22 @@ public function testWorkflowOcrBackendService() {
 		$this->assertEquals('application/pdf', $ocrResult->getContentType(), 'Expected content type in response');
 		$this->assertTrue(strpos($ocrResult->getRecognizedText(), 'This document is ready for OCR') >= 0, 'Expected recognized text in response');
 	}
+
+	public function testWorkflowOcrBackendServiceSkipFile() {
+		$this->addOperation('application/pdf', '{"ocrMode":' . WorkflowSettings::OCR_MODE_SKIP_FILE . '}');
+		$this->uploadTestFile('document-has-ocr.pdf');
+		$this->runOcrBackgroundJob();
+
+		$requests = $this->apiClient->getRequests();
+		$this->assertCount(1, $requests, 'Expected 1 OCR request');
+		$request = $requests[0];
+		$this->assertTrue($request['ocrMyPdfParameters'] === '', 'Expected OCR parameters in request');
+
+		$responses = $this->apiClient->getResponses();
+		$this->assertCount(1, $responses, 'Expected 1 OCR response');
+		$this->assertTrue($responses[0] instanceof ErrorResult, 'Expected ErrorResult instance, type is: ' . get_class($responses[0]));
+		/** @var ErrorResult */
+		$ocrResult = $responses[0];
+		$this->assertEquals($ocrResult->getOcrMyPdfExitCode(), 6, 'Expected ocrmypdf ExitCode 6');
+	}
 }

tests/Unit/OcrProcessors/Local/PdfOcrProcessorTest.php

@@ -23,6 +23,7 @@
 
 namespace OCA\WorkflowOcr\Tests\Unit\OcrProcessors\Local;
 
+use OCA\WorkflowOcr\Exception\OcrAlreadyDoneException;
 use OCA\WorkflowOcr\Exception\OcrNotPossibleException;
 use OCA\WorkflowOcr\Exception\OcrResultEmptyException;
 use OCA\WorkflowOcr\Helper\ISidecarFileAccessor;
@@ -386,6 +387,35 @@ public function testAppliesCustomCliArgsCorrectly() {
 		$processor->ocrFile($this->fileBefore, $workflowSettings, $this->defaultGlobalSettings);
 	}
 
+	public function testThrowsOcrAlreadyDoneExceptionIfErrorCodeIsEquals6() {
+		$this->command->expects($this->once())
+			->method('setCommand')
+			->willReturn($this->command);
+		$this->command->expects($this->once())
+			->method('execute')
+			->willReturn(false);
+		$this->command->expects($this->once())
+			->method('getError')
+			->willReturn('error');
+		$this->command->expects($this->once())
+			->method('getStdErr')
+			->willReturn('stdErr');
+		$this->command->expects($this->once())
+			->method('getExitCode')
+			->willReturn(6);
+
+		$thrown = false;
+		try {
+			$processor = new PdfOcrProcessor($this->command, $this->logger, $this->sidecarFileAccessor, $this->commandLineUtils);
+			$processor->ocrFile($this->fileBefore, $this->defaultSettings, $this->defaultGlobalSettings);
+		} catch (\Throwable $t) {
+			$thrown = true;
+			$this->assertInstanceOf(OcrAlreadyDoneException::class, $t);
+		}
+
+		$this->assertTrue($thrown);
+	}
+
 	public static function dataProvider_testAppliesOcrModeParameter() {
 		return [
 			[WorkflowSettings::OCR_MODE_SKIP_TEXT, ' --skip-text '],