Fix: queue single document zone OCR runs instead of running them in the request

Rendering and OCR ran inside the HTTP request and could block other
requests for minutes. The action now validates, then queues the bulk
zone OCR operation for the selected file and returns its task id.
This commit is contained in:
stumpylog committed 2026-11-19 15:25:48 -08:00
1 parent a6b81b2d64
commit c5fdba73a1
11 files changed
+379 -483

No files matched your search

+69 -48
View File
@@ -248,7 +248,6 @@ from documents.serialisers import UiSettingsViewSerializer
from documents.serialisers import WorkflowActionSerializer
from documents.serialisers import WorkflowSerializer
from documents.serialisers import WorkflowTriggerSerializer
from documents.serialisers import ZoneOcrResultSerializer
from documents.signals import document_updated
from documents.tasks import build_share_link_bundle
from documents.tasks import consume_file
@@ -2300,8 +2299,11 @@ class DocumentViewSet(
)
def run_zone_ocr(self, request: Request, pk: str | None = None) -> HttpResponse:
"""
Run zone-based OCR extraction on this document, with the given template
or with every template that matches it.
Queue zone-based OCR extraction on this document, with the given
template or with every template that matches it. The file version
given by the version parameter is run when it is not the root
document. Responds with the task id, or skipped when no template
applies.
"""
serializer = RunZoneOcrSerializer(
data=request.data,
@@ -2320,11 +2322,8 @@ class DocumentViewSet(
) or not has_perms_owner_aware(user, "change_document", document):
return HttpResponseForbidden("Insufficient permissions")
from documents.zone_ocr import ZoneOcrRunStatus
from documents.zone_ocr import ZoneOcrSkipReason
from documents.zone_ocr import get_templates_for_document
from documents.zone_ocr import run_zone_extraction
from documents.zone_ocr import zone_results_written
resolution = resolve_effective_document(document, request)
if resolution.error == VersionResolutionError.INVALID:
@@ -2333,17 +2332,14 @@ class DocumentViewSet(
if file_document is None:
raise Http404
templates = (
[template]
if template is not None
else get_templates_for_document(file_document, requester=user)
)
if not templates:
if template is None and not get_templates_for_document(
file_document,
requester=user,
):
return Response(
{
"status": ZoneOcrRunStatus.SKIPPED,
"result": "skipped",
"reason": ZoneOcrSkipReason.NO_TEMPLATE,
"results": [],
},
)
@@ -2354,18 +2350,22 @@ class DocumentViewSet(
status=status.HTTP_404_NOT_FOUND,
)
results = run_zone_extraction(file_document, templates=templates)
if zone_results_written(results):
bulk_edit.bulk_update_documents.apply_async(
kwargs={"document_ids": [get_root_document(file_document).pk]},
headers={"trigger_source": PaperlessTask.TriggerSource.SYSTEM},
parameters: dict[str, Any] = {"resolve_latest": False}
if template is not None:
parameters["template"] = template.pk
try:
return queue_bulk_operation(
user,
bulk_edit.run_zone_ocr,
[file_document.pk],
parameters,
from_webui=serializer.validated_data["from_webui"],
)
except Exception as e:
logger.warning(f"An error occurred performing document zone OCR: {e!s}")
return HttpResponseBadRequest(
"Error performing document zone OCR, check logs for more detail.",
)
return Response(
{
"status": ZoneOcrRunStatus.COMPLETED,
"results": ZoneOcrResultSerializer(results, many=True).data,
},
)
@extend_schema(
operation_id="documents_delete_version",
@@ -3101,35 +3101,56 @@ class DocumentSelectionMixin:
return list(filtered_documents.values_list("pk", flat=True))
def queue_bulk_operation(
user: User,
method: Callable[..., Any],
documents: list[int],
parameters: dict[str, Any],
*,
from_webui: bool,
) -> Response:
"""
Validate the parameters against the bulk edit method, queue it as a tracked
bulk operation task and return the task id.
"""
if "update_callback" in parameters:
raise ValueError("Unsupported bulk operation parameter: update_callback")
trigger_source = (
PaperlessTask.TriggerSource.WEB_UI
if from_webui
else PaperlessTask.TriggerSource.API_UPLOAD
)
final_parameters = parameters.copy()
if method.__name__ in USER_OPERATIONS:
final_parameters["user"] = user
if method.__name__ in TRIGGER_SOURCE_OPERATIONS:
final_parameters["trigger_source"] = trigger_source
signature(method).bind(documents, **final_parameters)
task = execute_bulk_operation.apply_async(
kwargs={
"operation": method.__name__,
"document_ids": documents,
"parameters": parameters,
"user_id": user.pk,
"trigger_source": trigger_source,
},
headers={"trigger_source": trigger_source},
)
return Response({"result": "OK", "task_id": task.id})
class DocumentOperationPermissionMixin(PassUserMixin, DocumentSelectionMixin):
permission_classes = (IsAuthenticated,)
parser_classes = (parsers.JSONParser,)
def _queue_bulk_operation(self, method, documents, parameters, from_webui):
if "update_callback" in parameters:
raise ValueError("Unsupported bulk operation parameter: update_callback")
trigger_source = (
PaperlessTask.TriggerSource.WEB_UI
if from_webui
else PaperlessTask.TriggerSource.API_UPLOAD
return queue_bulk_operation(
cast("User", self.request.user),
method,
documents,
parameters,
from_webui=from_webui,
)
final_parameters = parameters.copy()
if method.__name__ in USER_OPERATIONS:
final_parameters["user"] = self.request.user
if method.__name__ in TRIGGER_SOURCE_OPERATIONS:
final_parameters["trigger_source"] = trigger_source
signature(method).bind(documents, **final_parameters)
task = execute_bulk_operation.apply_async(
kwargs={
"operation": method.__name__,
"document_ids": documents,
"parameters": parameters,
"user_id": self.request.user.pk,
"trigger_source": trigger_source,
},
headers={"trigger_source": trigger_source},
)
return Response({"result": "OK", "task_id": task.id})
def _has_document_permissions(
self,