Lint

paperless.conf.example

@@ -14,6 +14,7 @@
 # Paths and folders
 
 #PAPERLESS_CONSUMPTION_DIR=../consume
+#PAPERLESS_CONSUMPTION_FAILED_DIR=../consume/failed
 #PAPERLESS_DATA_DIR=../data
 #PAPERLESS_EMPTY_TRASH_DIR=
 #PAPERLESS_MEDIA_ROOT=../media

src-ui/src/app/components/admin/tasks/tasks.component.html

@@ -112,6 +112,9 @@
           </td>
           <td scope="row">
             <div class="btn-group" role="group">
+              @if (task.status === PaperlessTaskStatus.Failed) {
+                <ng-container *ngTemplateOutlet="retryDropdown; context: { task: task }"></ng-container>
+              }
               <button class="btn btn-sm btn-outline-secondary" (click)="dismissTask(task); $event.stopPropagation();" *pngxIfPermissions="{ action: PermissionAction.Change, type: PermissionType.PaperlessTask }">
                 <i-bs name="check" class="me-1"></i-bs><ng-container i18n>Dismiss</ng-container>
               </button>
@@ -184,3 +187,25 @@
   </li>
 </ul>
 <div [ngbNavOutlet]="nav"></div>
+
+<ng-template #retryDropdown let-task="task">
+  <div ngbDropdown>
+    <button class="btn btn-sm btn-outline-primary" (click)="$event.stopImmediatePropagation()" ngbDropdownToggle>
+      <i-bs name="arrow-repeat"></i-bs>&nbsp;<ng-container i18n>Retry</ng-container>
+    </button>
+    <div ngbDropdownMenu class="shadow retry-dropdown">
+      <div class="p-2">
+        <ul class="list-group list-group-flush">
+            <li class="list-group-item small" i18n>
+              <pngx-input-check [(ngModel)]="retryClean" i18n-title title="Attempt to clean pdf"></pngx-input-check>
+            </li>
+        </ul>
+        <div class="d-flex justify-content-end">
+          <button class="btn btn-sm btn-outline-primary" (click)="retryTask(task); $event.stopPropagation();">
+            <ng-container i18n>Proceed</ng-container>
+          </button>
+        </div>
+      </div>
+    </div>
+  </div>
+</ng-template>

src-ui/src/app/components/admin/tasks/tasks.component.scss

@@ -37,3 +37,7 @@ pre {
 .z-10 {
     z-index: 10;
 }
+
+.retry-dropdown {
+    width: 300px;
+}

src-ui/src/app/components/admin/tasks/tasks.component.spec.ts

@@ -16,7 +16,7 @@ import {
   NgbNavItem,
 } from '@ng-bootstrap/ng-bootstrap'
 import { allIcons, NgxBootstrapIconsModule } from 'ngx-bootstrap-icons'
-import { throwError } from 'rxjs'
+import { of, throwError } from 'rxjs'
 import { routes } from 'src/app/app-routing.module'
 import {
   PaperlessTask,
@@ -32,6 +32,7 @@ import { TasksService } from 'src/app/services/tasks.service'
 import { ToastService } from 'src/app/services/toast.service'
 import { environment } from 'src/environments/environment'
 import { ConfirmDialogComponent } from '../../common/confirm-dialog/confirm-dialog.component'
+import { CheckComponent } from '../../common/input/check/check.component'
 import { PageHeaderComponent } from '../../common/page-header/page-header.component'
 import { TasksComponent, TaskTab } from './tasks.component'
 
@@ -138,6 +139,7 @@ describe('TasksComponent', () => {
         PageHeaderComponent,
         IfPermissionsDirective,
         CustomDatePipe,
+        CheckComponent,
         ConfirmDialogComponent,
       ],
       providers: [
@@ -184,8 +186,10 @@ describe('TasksComponent', () => {
       `Failed${currentTasksLength}`
     )
     expect(
-      fixture.debugElement.queryAll(By.css('table input[type="checkbox"]'))
+      fixture.debugElement.queryAll(
-    ).toHaveLength(currentTasksLength + 1)
+        By.css('table td > .form-check input[type="checkbox"]')
+      )
+    ).toHaveLength(currentTasksLength)
 
     currentTasksLength = tasks.filter(
       (t) => t.status === PaperlessTaskStatus.Complete
@@ -389,4 +393,20 @@ describe('TasksComponent', () => {
     expect(component.filterText).toEqual('')
     expect(component.filterTargetID).toEqual(0)
   })
+
+  it('should retry a task, show toast on error or success', () => {
+    const retrySpy = jest.spyOn(tasksService, 'retryTask')
+    const toastInfoSpy = jest.spyOn(toastService, 'showInfo')
+    const toastErrorSpy = jest.spyOn(toastService, 'showError')
+    retrySpy.mockReturnValueOnce(of({ task_id: '123' }))
+    component.retryTask(tasks[0])
+    expect(retrySpy).toHaveBeenCalledWith(tasks[0], false)
+    expect(toastInfoSpy).toHaveBeenCalledWith('Retrying task...')
+    retrySpy.mockReturnValueOnce(throwError(() => new Error('test')))
+    component.retryTask(tasks[0])
+    expect(toastErrorSpy).toHaveBeenCalledWith(
+      'Failed to retry task',
+      new Error('test')
+    )
+  })
 })

src-ui/src/app/components/admin/tasks/tasks.component.ts

@@ -20,12 +20,13 @@ import {
   takeUntil,
   timer,
 } from 'rxjs'
-import { PaperlessTask } from 'src/app/data/paperless-task'
+import { PaperlessTask, PaperlessTaskStatus } from 'src/app/data/paperless-task'
 import { IfPermissionsDirective } from 'src/app/directives/if-permissions.directive'
 import { CustomDatePipe } from 'src/app/pipes/custom-date.pipe'
 import { TasksService } from 'src/app/services/tasks.service'
 import { ToastService } from 'src/app/services/toast.service'
 import { ConfirmDialogComponent } from '../../common/confirm-dialog/confirm-dialog.component'
+import { CheckComponent } from '../../common/input/check/check.component'
 import { PageHeaderComponent } from '../../common/page-header/page-header.component'
 import { LoadingComponentWithPermissions } from '../../loading-component/loading.component'
 
@@ -54,6 +55,7 @@ const FILTER_TARGETS = [
     PageHeaderComponent,
     IfPermissionsDirective,
     CustomDatePipe,
+    CheckComponent,
     SlicePipe,
     FormsModule,
     ReactiveFormsModule,
@@ -75,6 +77,7 @@ export class TasksComponent
   private readonly router = inject(Router)
   private readonly toastService = inject(ToastService)
 
+  public PaperlessTaskStatus = PaperlessTaskStatus
   public activeTab: TaskTab
   public selectedTasks: Set<number> = new Set()
   public togggleAll: boolean = false
@@ -105,6 +108,8 @@ export class TasksComponent
       : FILTER_TARGETS.slice(0, 1)
   }
 
+  public retryClean: boolean = false
+
   get dismissButtonText(): string {
     return this.selectedTasks.size > 0
       ? $localize`Dismiss selected`
@@ -178,6 +183,17 @@ export class TasksComponent
     this.router.navigate(['documents', task.related_document])
   }
 
+  retryTask(task: PaperlessTask) {
+    this.tasksService.retryTask(task, this.retryClean).subscribe({
+      next: () => {
+        this.toastService.showInfo($localize`Retrying task...`)
+      },
+      error: (e) => {
+        this.toastService.showError($localize`Failed to retry task`, e)
+      },
+    })
+  }
+
   expandTask(task: PaperlessTask) {
     this.expandedTask = this.expandedTask == task.id ? undefined : task.id
   }

src-ui/src/app/services/tasks.service.spec.ts

@@ -147,4 +147,33 @@ describe('TasksService', () => {
       result: 'success',
     })
   })
+
+  it('should call retry task api endpoint', () => {
+    const task = {
+      id: 1,
+      type: PaperlessTaskType.File,
+      status: PaperlessTaskStatus.Failed,
+      acknowledged: false,
+      task_id: '1234',
+      task_file_name: 'file1.pdf',
+      date_created: new Date(),
+    }
+
+    tasksService.retryTask(task, true).subscribe()
+    const reloadSpy = jest.spyOn(tasksService, 'reload')
+    const req = httpTestingController.expectOne(
+      `${environment.apiBaseUrl}tasks/${task.id}/retry/`
+    )
+    expect(req.request.method).toEqual('POST')
+    expect(req.request.body).toEqual({
+      clean: true,
+    })
+    req.flush({ task_id: 12345 })
+    expect(reloadSpy).toHaveBeenCalled()
+    httpTestingController
+      .expectOne(
+        `${environment.apiBaseUrl}tasks/?task_name=consume_file&acknowledged=false`
+      )
+      .flush([])
+  })
 })

src-ui/src/app/services/tasks.service.ts

@@ -81,6 +81,20 @@ export class TasksService {
       )
   }
 
+  public retryTask(task: PaperlessTask, clean: boolean): Observable<any> {
+    return this.http
+      .post(`${this.baseUrl}tasks/${task.id}/retry/`, {
+        clean,
+      })
+      .pipe(
+        takeUntil(this.unsubscribeNotifer),
+        first(),
+        tap(() => {
+          this.reload()
+        })
+      )
+  }
+
   public cancelPending(): void {
     this.unsubscribeNotifer.next(true)
   }

src/documents/consumer.py

@@ -174,6 +174,17 @@ class ConsumerPluginMixin:
     ):
         self._send_progress(100, 100, ProgressStatusOptions.FAILED, message)
         self.log.error(log_message or message, exc_info=exc_info)
+        # Move the file to the failed directory
+        if (
+            self.input_doc.original_file.exists()
+            and not Path(
+                settings.CONSUMPTION_FAILED_DIR / self.input_doc.original_file.name,
+            ).exists()
+        ):
+            copy_file_with_basic_stats(
+                self.input_doc.original_file,
+                settings.CONSUMPTION_FAILED_DIR / self.input_doc.original_file.name,
+            )
         raise ConsumerError(f"{self.filename}: {log_message or message}") from exception
 
 

src/documents/serialisers.py

@@ -2411,6 +2411,14 @@ class TasksViewSerializer(OwnedObjectSerializer):
         return list(duplicates.values("id", "title", "deleted_at"))
 
 
+class RetryTaskSerializer(serializers.Serializer):
+    clean = serializers.BooleanField(
+        default=False,
+        write_only=True,
+        required=False,
+    )
+
+
 class RunTaskViewSerializer(serializers.Serializer[dict[str, Any]]):
     task_name = serializers.ChoiceField(
         choices=PaperlessTask.TaskName.choices,

src/documents/signals/handlers.py

@@ -631,6 +631,19 @@ def update_filename_and_move_files(
             )
 
 
+@receiver(models.signals.post_save, sender=PaperlessTask)
+def cleanup_failed_documents(sender, instance: PaperlessTask, **kwargs):
+    if instance.status != states.FAILURE or not instance.acknowledged:
+        return
+
+    if instance.task_file_name:
+        try:
+            Path(settings.CONSUMPTION_FAILED_DIR / instance.task_file_name).unlink()
+            logger.debug(f"Cleaned up failed file {instance.task_file_name}")
+        except FileNotFoundError:
+            logger.warning(f"Failed to clean up failed file {instance.task_file_name}")
+
+
 @shared_task
 def process_cf_select_update(custom_field: CustomField) -> None:
     """

src/documents/tasks.py

@@ -37,6 +37,7 @@ from documents.consumer import ConsumerPreflightPlugin
 from documents.consumer import WorkflowTriggerPlugin
 from documents.data_models import ConsumableDocument
 from documents.data_models import DocumentMetadataOverrides
+from documents.data_models import DocumentSource
 from documents.double_sided import CollatePlugin
 from documents.file_handling import create_source_path_directory
 from documents.file_handling import generate_unique_filename
@@ -63,6 +64,8 @@ from documents.signals import document_updated
 from documents.signals.handlers import cleanup_document_deletion
 from documents.signals.handlers import run_workflows
 from documents.signals.handlers import send_websocket_document_updated
+from documents.utils import copy_file_with_basic_stats
+from documents.utils import run_subprocess
 from documents.workflows.utils import get_workflows_for_trigger
 from paperless.config import AIConfig
 from paperless_ai.indexing import llm_index_add_or_update_document
@@ -72,7 +75,6 @@ from paperless_ai.indexing import update_llm_index
 _T = TypeVar("_T")
 IterWrapper = Callable[[Iterable[_T]], Iterable[_T]]
 
-
 if settings.AUDIT_LOG_ENABLED:
     from auditlog.models import LogEntry
 logger = logging.getLogger("paperless.tasks")
@@ -239,6 +241,48 @@ def consume_file(
     return msg
 
 
+@shared_task
+def retry_failed_file(task_id: str, clean: bool = False, skip_ocr: bool = False):
+    task = PaperlessTask.objects.get(task_id=task_id, status=states.FAILURE)
+    if task:
+        failed_file = settings.CONSUMPTION_FAILED_DIR / task.task_file_name
+        if not failed_file.exists():
+            logger.error(f"File {failed_file} not found")
+            raise FileNotFoundError(f"File {failed_file} not found")
+        working_copy = settings.SCRATCH_DIR / failed_file.name
+        copy_file_with_basic_stats(failed_file, working_copy)
+
+        if clean:
+            try:
+                result = run_subprocess(
+                    [
+                        "qpdf",
+                        "--replace-input",
+                        "--warning-exit-0",
+                        working_copy,
+                    ],
+                    logger=logger,
+                )
+                if result.returncode != 0:
+                    raise Exception(
+                        f"qpdf failed with exit code {result.returncode}, error: {result.stderr}",
+                    )
+                else:
+                    logger.debug("PDF cleaned successfully")
+            except Exception as e:
+                logger.error(f"Error while cleaning PDF: {e}")
+                raise e
+
+        task = consume_file.delay(
+            ConsumableDocument(
+                source=DocumentSource.ConsumeFolder,
+                original_file=working_copy,
+            ),
+        )
+
+        return task.id
+
+
 @shared_task
 def sanity_check(*, scheduled=True, raise_on_error=True):
     messages = sanity_checker.check_sanity(scheduled=scheduled)

src/documents/tests/test_tasks.py

@@ -1,4 +1,5 @@
 import shutil
+import uuid
 from datetime import timedelta
 from pathlib import Path
 from unittest import mock
@@ -21,6 +22,7 @@ from documents.sanity_checker import SanityCheckMessages
 from documents.tests.test_classifier import dummy_preprocess
 from documents.tests.utils import DirectoriesMixin
 from documents.tests.utils import FileSystemAssertsMixin
+from documents.tests.utils import SampleDirMixin
 
 
 class TestIndexReindex(DirectoriesMixin, TestCase):
@@ -232,6 +234,44 @@ class TestEmptyTrashTask(DirectoriesMixin, FileSystemAssertsMixin, TestCase):
         self.assertEqual(Document.global_objects.count(), 0)
 
 
+class TestRetryConsumeTask(
+    DirectoriesMixin,
+    SampleDirMixin,
+    FileSystemAssertsMixin,
+    TestCase,
+):
+    def do_failed_task(self, test_file: Path) -> PaperlessTask:
+        failed_file = settings.CONSUMPTION_FAILED_DIR / test_file.name
+        shutil.copy(test_file, failed_file)
+
+        task = PaperlessTask.objects.create(
+            type=PaperlessTask.TaskType.AUTO,
+            task_id=str(uuid.uuid4()),
+            task_file_name=failed_file.name,
+            task_name=PaperlessTask.TaskName.CONSUME_FILE,
+            status=states.FAILURE,
+            date_created=timezone.now(),
+            date_done=timezone.now(),
+        )
+        self.assertIsFile(settings.CONSUMPTION_FAILED_DIR / task.task_file_name)
+        return task
+
+    @mock.patch("documents.tasks.consume_file.delay")
+    @mock.patch("documents.tasks.run_subprocess")
+    def test_retry_consume_clean(self, m_subprocess, m_consume_file) -> None:
+        task = self.do_failed_task(self.SAMPLE_DIR / "corrupted.pdf")
+        m_subprocess.return_value.returncode = 0
+        task_id = tasks.retry_failed_file(task_id=task.task_id, clean=True)
+        self.assertIsNotNone(task_id)
+        m_consume_file.assert_called_once()
+
+    def test_cleanup(self) -> None:
+        task = self.do_failed_task(self.SAMPLE_DIR / "corrupted.pdf")
+        task.acknowledged = True
+        task.save()
+        self.assertIsNotFile(settings.CONSUMPTION_FAILED_DIR / task.task_file_name)
+
+
 class TestUpdateContent(DirectoriesMixin, TestCase):
     def test_update_content_maybe_archive_file(self) -> None:
         """

src/documents/tests/utils.py

@@ -37,6 +37,7 @@ def setup_directories():
     dirs.scratch_dir = Path(tempfile.mkdtemp()).resolve()
     dirs.media_dir = Path(tempfile.mkdtemp()).resolve()
     dirs.consumption_dir = Path(tempfile.mkdtemp()).resolve()
+    dirs.consumption_failed_dir = Path(tempfile.mkdtemp("failed")).resolve()
     dirs.static_dir = Path(tempfile.mkdtemp()).resolve()
     dirs.index_dir = dirs.data_dir / "index"
     dirs.originals_dir = dirs.media_dir / "documents" / "originals"
@@ -58,6 +59,7 @@ def setup_directories():
         THUMBNAIL_DIR=dirs.thumbnail_dir,
         ARCHIVE_DIR=dirs.archive_dir,
         CONSUMPTION_DIR=dirs.consumption_dir,
+        CONSUMPTION_FAILED_DIR=dirs.consumption_failed_dir,
         LOGGING_DIR=dirs.logging_dir,
         INDEX_DIR=dirs.index_dir,
         STATIC_ROOT=dirs.static_dir,
@@ -74,6 +76,7 @@ def remove_dirs(dirs) -> None:
     shutil.rmtree(dirs.data_dir, ignore_errors=True)
     shutil.rmtree(dirs.scratch_dir, ignore_errors=True)
     shutil.rmtree(dirs.consumption_dir, ignore_errors=True)
+    shutil.rmtree(dirs.consumption_failed_dir, ignore_errors=True)
     shutil.rmtree(dirs.static_dir, ignore_errors=True)
     dirs.settings_override.disable()
 

src/documents/views.py

@@ -189,6 +189,7 @@ from documents.serialisers import NotesSerializer
 from documents.serialisers import PostDocumentSerializer
 from documents.serialisers import RemovePasswordDocumentsSerializer
 from documents.serialisers import ReprocessDocumentsSerializer
+from documents.serialisers import RetryTaskSerializer
 from documents.serialisers import RotateDocumentsSerializer
 from documents.serialisers import RunTaskViewSerializer
 from documents.serialisers import SavedViewSerializer
@@ -211,6 +212,7 @@ from documents.tasks import consume_file
 from documents.tasks import empty_trash
 from documents.tasks import index_optimize
 from documents.tasks import llmindex_index
+from documents.tasks import retry_failed_file
 from documents.tasks import sanity_check
 from documents.tasks import train_classifier
 from documents.tasks import update_document_parent_tags
@@ -3467,6 +3469,25 @@ class TasksViewSet(ReadOnlyModelViewSet):
             queryset = PaperlessTask.objects.filter(task_id=task_id)
         return queryset
 
+    @action(methods=["post"], detail=True)
+    def retry(self, request, pk=None):
+        task = self.get_object()
+
+        serializer = RetryTaskSerializer(data=request.data)
+        serializer.is_valid(raise_exception=True)
+        clean = serializer.validated_data.get("clean")
+
+        try:
+            new_task_id = retry_failed_file(task.task_id, clean)
+            return Response({"task_id": new_task_id})
+        except FileNotFoundError:
+            return HttpResponseBadRequest("Original file not found")
+        except Exception as e:
+            logger.warning(f"An error occurred retrying task: {e!s}")
+            return HttpResponseBadRequest(
+                "Error retrying task, check logs for more detail.",
+            )
+
     @action(
         methods=["post"],
         detail=False,

src/paperless/checks.py

@@ -68,6 +68,10 @@ def paths_check(app_configs, **kwargs) -> list[Error]:
         + path_check("PAPERLESS_EMPTY_TRASH_DIR", settings.EMPTY_TRASH_DIR)
         + path_check("PAPERLESS_MEDIA_ROOT", settings.MEDIA_ROOT)
         + path_check("PAPERLESS_CONSUMPTION_DIR", settings.CONSUMPTION_DIR)
+        + path_check(
+            "PAPERLESS_CONSUMPTION_FAILED_DIR",
+            settings.CONSUMPTION_FAILED_DIR,
+        )
     )
 
 

src/paperless/settings/__init__.py

@@ -98,6 +98,11 @@ CONSUMPTION_DIR = get_path_from_env(
     BASE_DIR.parent / "consume",
 )
 
+CONSUMPTION_FAILED_DIR = get_path_from_env(
+    "PAPERLESS_CONSUMPTION_FAILED_DIR",
+    CONSUMPTION_DIR / "failed",
+)
+
 # This will be created if it doesn't exist
 SCRATCH_DIR = get_path_from_env(
     "PAPERLESS_SCRATCH_DIR",
@@ -782,6 +787,8 @@ CONSUMER_IGNORE_PATTERNS = list(
         ),
     ),
 )
+if CONSUMPTION_DIR in CONSUMPTION_FAILED_DIR.parents:
+    CONSUMER_IGNORE_PATTERNS.append(CONSUMPTION_FAILED_DIR.name)
 
 # Directories to always ignore.  These are matched by directory name, not full path
 CONSUMER_IGNORE_DIRS = list(

src/paperless/tests/test_checks.py

@@ -24,6 +24,7 @@ class PaperlessTestDirs:
     data_dir: Path
     media_dir: Path
     consumption_dir: Path
+    consumption_failed_dir: Path
 
 
 # TODO: consolidate with documents/tests/conftest.py PaperlessDirs/paperless_dirs
@@ -33,18 +34,21 @@ def directories(tmp_path: Path, settings: SettingsWrapper) -> PaperlessTestDirs:
     data_dir = tmp_path / "data"
     media_dir = tmp_path / "media"
     consumption_dir = tmp_path / "consumption"
+    consumption_failed_dir = tmp_path / "consumption_failed"
 
-    for d in (data_dir, media_dir, consumption_dir):
+    for d in (data_dir, media_dir, consumption_dir, consumption_failed_dir):
         d.mkdir()
 
     settings.DATA_DIR = data_dir
     settings.MEDIA_ROOT = media_dir
     settings.CONSUMPTION_DIR = consumption_dir
+    settings.CONSUMPTION_FAILED_DIR = consumption_failed_dir
 
     return PaperlessTestDirs(
         data_dir=data_dir,
         media_dir=media_dir,
         consumption_dir=consumption_dir,
+        consumption_failed_dir=consumption_failed_dir,
     )
 
 
@@ -64,10 +68,11 @@ class TestChecks:
         settings.MEDIA_ROOT = Path("uuh")
         settings.DATA_DIR = Path("whatever")
         settings.CONSUMPTION_DIR = Path("idontcare")
+        settings.CONSUMPTION_FAILED_DIR = Path("nope")
 
         msgs = paths_check(None)
 
-        assert len(msgs) == 3, str(msgs)
+        assert len(msgs) == 4, str(msgs)
         for msg in msgs:
             assert msg.msg.endswith("is set but doesn't exist.")
 
@@ -75,6 +80,7 @@ class TestChecks:
         directories.data_dir.chmod(0o000)
         directories.media_dir.chmod(0o000)
         directories.consumption_dir.chmod(0o000)
+        directories.consumption_failed_dir.chmod(0o000)
 
         try:
             msgs = paths_check(None)
@@ -82,8 +88,9 @@ class TestChecks:
             directories.data_dir.chmod(0o777)
             directories.media_dir.chmod(0o777)
             directories.consumption_dir.chmod(0o777)
+            directories.consumption_failed_dir.chmod(0o777)
 
-        assert len(msgs) == 3
+        assert len(msgs) == 4
         for msg in msgs:
             assert msg.msg.endswith("is not writeable")