Fix (beta): move task filtering to backend fully (#12956)

src/documents/filters.py

@@ -28,6 +28,7 @@ from django.db.models.functions import Cast
 from django.utils.translation import gettext_lazy as _
 from django_filters import DateFilter
 from django_filters.rest_framework import BooleanFilter
+from django_filters.rest_framework import CharFilter
 from django_filters.rest_framework import DateTimeFilter
 from django_filters.rest_framework import Filter
 from django_filters.rest_framework import FilterSet
@@ -900,6 +901,16 @@ class ShareLinkBundleFilterSet(FilterSet):
 
 
 class PaperlessTaskFilterSet(FilterSet):
+    name = CharFilter(
+        method="filter_name",
+        label="Name",
+    )
+
+    result = CharFilter(
+        method="filter_result",
+        label="Result",
+    )
+
     task_type = MultipleChoiceFilter(
         choices=PaperlessTask.TaskType.choices,
         label="Task Type",
@@ -939,7 +950,58 @@ class PaperlessTaskFilterSet(FilterSet):
 
     class Meta:
         model = PaperlessTask
-        fields = ["task_type", "trigger_source", "status", "acknowledged", "owner"]
+        fields = [
+            "task_type",
+            "trigger_source",
+            "status",
+            "acknowledged",
+            "owner",
+            "name",
+            "result",
+        ]
+
+    def filter_name(self, queryset, name, value):
+        if not value:
+            return queryset
+
+        matching_task_types = [
+            task_type
+            for task_type, label in PaperlessTask.TaskType.choices
+            if value.lower() in str(label).lower()
+        ]
+        matching_trigger_sources = [
+            trigger_source
+            for trigger_source, label in PaperlessTask.TriggerSource.choices
+            if value.lower() in str(label).lower()
+        ]
+
+        return queryset.filter(
+            Q(input_data__filename__icontains=value)
+            | Q(task_type__in=matching_task_types)
+            | Q(trigger_source__in=matching_trigger_sources),
+        )
+
+    def filter_result(self, queryset, name, value):
+        if not value:
+            return queryset
+
+        query = Q(result_data__reason__icontains=value) | Q(
+            result_data__error_message__icontains=value,
+        )
+
+        try:
+            numeric_value = int(value)
+        except (TypeError, ValueError):
+            pass
+        else:
+            query |= Q(result_data__document_id=numeric_value) | Q(
+                result_data__duplicate_of=numeric_value,
+            )
+
+        if "duplicate" in value.lower():
+            query |= Q(result_data__duplicate_of__isnull=False)
+
+        return queryset.filter(query)
 
     def filter_is_complete(self, queryset, name, value):
         if value:

src/documents/tests/test_api_tasks.py

@@ -18,6 +18,7 @@ from guardian.shortcuts import assign_perm
 from rest_framework import status
 from rest_framework.test import APIClient
 
+from documents.filters import PaperlessTaskFilterSet
 from documents.models import PaperlessTask
 from documents.tests.factories import DocumentFactory
 from documents.tests.factories import PaperlessTaskFactory
@@ -169,6 +170,165 @@ class TestGetTasksV10:
             PaperlessTask.Status.STARTED,
         }
 
+    def test_filter_by_task_name(self, admin_client: APIClient) -> None:
+        """?name= searches task filenames, task types, and trigger sources."""
+        filename_task = PaperlessTaskFactory(input_data={"filename": "invoice-123.pdf"})
+        type_task = PaperlessTaskFactory(task_type=PaperlessTask.TaskType.SANITY_CHECK)
+        source_task = PaperlessTaskFactory(
+            trigger_source=PaperlessTask.TriggerSource.EMAIL_CONSUME,
+        )
+        PaperlessTaskFactory(input_data={"filename": "unrelated.pdf"})
+
+        response = admin_client.get(ENDPOINT, {"name": "invoice"})
+
+        assert response.status_code == status.HTTP_200_OK
+        assert response.data["count"] == 1
+        assert response.data["results"][0]["task_id"] == filename_task.task_id
+
+        response = admin_client.get(ENDPOINT, {"name": "sanity"})
+
+        assert response.status_code == status.HTTP_200_OK
+        assert response.data["count"] == 1
+        assert response.data["results"][0]["task_id"] == type_task.task_id
+
+        response = admin_client.get(ENDPOINT, {"name": "email"})
+
+        assert response.status_code == status.HTTP_200_OK
+        assert response.data["count"] == 1
+        assert response.data["results"][0]["task_id"] == source_task.task_id
+
+    def test_filter_by_task_result(self, admin_client: APIClient) -> None:
+        """?result= searches common structured task result messages."""
+        reason_task = PaperlessTaskFactory(result_data={"reason": "Manual review"})
+        error_task = PaperlessTaskFactory(
+            result_data={"error_message": "Duplicate detected"},
+        )
+        document_task = PaperlessTaskFactory(result_data={"document_id": 321})
+        duplicate_task = PaperlessTaskFactory(result_data={"duplicate_of": 123})
+        PaperlessTaskFactory(result_data={"reason": "unrelated"})
+
+        response = admin_client.get(ENDPOINT, {"result": "manual"})
+
+        assert response.status_code == status.HTTP_200_OK
+        assert response.data["count"] == 1
+        assert response.data["results"][0]["task_id"] == reason_task.task_id
+
+        response = admin_client.get(ENDPOINT, {"result": "duplicate"})
+
+        assert response.status_code == status.HTTP_200_OK
+        returned_ids = {task["task_id"] for task in response.data["results"]}
+        assert returned_ids == {error_task.task_id, duplicate_task.task_id}
+
+        response = admin_client.get(ENDPOINT, {"result": "321"})
+
+        assert response.status_code == status.HTTP_200_OK
+        assert response.data["count"] == 1
+        assert response.data["results"][0]["task_id"] == document_task.task_id
+
+    def test_empty_task_name_and_result_filters(self) -> None:
+        """Empty name/result values leave the queryset unchanged."""
+        PaperlessTaskFactory.create_batch(2)
+        queryset = PaperlessTask.objects.all()
+        filterset = PaperlessTaskFilterSet()
+
+        assert filterset.filter_name(queryset, "name", "").count() == 2
+        assert filterset.filter_result(queryset, "result", "").count() == 2
+
+    def test_status_counts_respects_filters(self, admin_client: APIClient) -> None:
+        """status_counts/ returns section counts for the filtered task queryset."""
+        PaperlessTaskFactory(
+            acknowledged=False,
+            status=PaperlessTask.Status.FAILURE,
+            input_data={"filename": "invoice-a.pdf"},
+        )
+        PaperlessTaskFactory(
+            acknowledged=False,
+            status=PaperlessTask.Status.REVOKED,
+            input_data={"filename": "invoice-b.pdf"},
+        )
+        PaperlessTaskFactory(
+            acknowledged=False,
+            status=PaperlessTask.Status.PENDING,
+            input_data={"filename": "invoice-c.pdf"},
+        )
+        PaperlessTaskFactory(
+            acknowledged=False,
+            status=PaperlessTask.Status.STARTED,
+            input_data={"filename": "invoice-d.pdf"},
+        )
+        PaperlessTaskFactory(
+            acknowledged=False,
+            status=PaperlessTask.Status.SUCCESS,
+            input_data={"filename": "invoice-e.pdf"},
+        )
+        PaperlessTaskFactory(
+            acknowledged=True,
+            status=PaperlessTask.Status.SUCCESS,
+            input_data={"filename": "invoice-acknowledged.pdf"},
+        )
+        PaperlessTaskFactory(
+            acknowledged=False,
+            status=PaperlessTask.Status.SUCCESS,
+            input_data={"filename": "unrelated.pdf"},
+        )
+
+        response = admin_client.get(
+            f"{ENDPOINT}status_counts/",
+            {"acknowledged": "false", "name": "invoice"},
+        )
+
+        assert response.status_code == status.HTTP_200_OK
+        assert response.data == {
+            "all": 5,
+            "needs_attention": 2,
+            "in_progress": 2,
+            "completed": 1,
+        }
+
+    def test_status_counts_ignores_section_filters(
+        self,
+        admin_client: APIClient,
+    ) -> None:
+        """status_counts/ ignores status-like filters for the sections it counts."""
+        PaperlessTaskFactory(
+            acknowledged=False,
+            status=PaperlessTask.Status.FAILURE,
+            input_data={"filename": "invoice-a.pdf"},
+        )
+        PaperlessTaskFactory(
+            acknowledged=False,
+            status=PaperlessTask.Status.PENDING,
+            input_data={"filename": "invoice-b.pdf"},
+        )
+        PaperlessTaskFactory(
+            acknowledged=False,
+            status=PaperlessTask.Status.SUCCESS,
+            input_data={"filename": "invoice-c.pdf"},
+        )
+        PaperlessTaskFactory(
+            acknowledged=False,
+            status=PaperlessTask.Status.FAILURE,
+            input_data={"filename": "unrelated.pdf"},
+        )
+
+        response = admin_client.get(
+            f"{ENDPOINT}status_counts/",
+            {
+                "acknowledged": "false",
+                "name": "invoice",
+                "status": PaperlessTask.Status.FAILURE,
+                "is_complete": "false",
+            },
+        )
+
+        assert response.status_code == status.HTTP_200_OK
+        assert response.data == {
+            "all": 3,
+            "needs_attention": 1,
+            "in_progress": 1,
+            "completed": 1,
+        }
+
     def test_default_ordering_is_newest_first(self, admin_client: APIClient) -> None:
         """Tasks are returned in descending date_created order (newest first)."""
         base = timezone.now()

src/documents/views.py

@@ -4011,7 +4011,7 @@ class RemoteVersionView(GenericAPIView[Any]):
 
 
 class _TasksViewSetSchema(AutoSchema):
-    _UNPAGINATED_ACTIONS = frozenset({"summary", "active"})
+    _UNPAGINATED_ACTIONS = frozenset({"summary", "active", "status_counts"})
 
     def _get_paginator(self):
         if getattr(self.view, "action", None) in self._UNPAGINATED_ACTIONS:
@@ -4071,6 +4071,19 @@ class _TasksViewSetSchema(AutoSchema):
             ),
         ],
     ),
+    status_counts=extend_schema(
+        responses={
+            200: inline_serializer(
+                name="TaskStatusCounts",
+                fields={
+                    "all": serializers.IntegerField(),
+                    "needs_attention": serializers.IntegerField(),
+                    "in_progress": serializers.IntegerField(),
+                    "completed": serializers.IntegerField(),
+                },
+            ),
+        },
+    ),
     active=extend_schema(
         description="Currently pending and running tasks (capped at 50).",
         responses={200: TaskSerializerV10(many=True)},
@@ -4124,6 +4137,7 @@ class TasksViewSet(ReadOnlyModelViewSet[PaperlessTask]):
         PaperlessTask.TaskType.SANITY_CHECK: (sanity_check, {"raise_on_error": False}),
         PaperlessTask.TaskType.LLM_INDEX: (llmindex_index, {"rebuild": False}),
     }
+    _STATUS_COUNT_EXCLUDED_FILTERS = frozenset({"status", "is_complete"})
 
     def get_serializer_class(self):
         # v9: use backwards-compatible serializer with old field names
@@ -4164,6 +4178,21 @@ class TasksViewSet(ReadOnlyModelViewSet[PaperlessTask]):
             queryset = queryset.filter(task_id=task_id)
         return queryset
 
+    def get_status_count_queryset(self):
+        """Apply task filters except the status dimensions represented by the counts."""
+        query_params = self.request.query_params.copy()
+        for param in self._STATUS_COUNT_EXCLUDED_FILTERS:
+            query_params.pop(param, None)
+
+        filterset = self.filterset_class(
+            data=query_params,
+            queryset=self.get_queryset(),
+            request=self.request,
+        )
+        if not filterset.is_valid():
+            raise ValidationError(filterset.errors)
+        return filterset.qs
+
     @action(
         methods=["post"],
         detail=False,
@@ -4233,6 +4262,34 @@ class TasksViewSet(ReadOnlyModelViewSet[PaperlessTask]):
         serializer = TaskSummarySerializer(data, many=True)
         return Response(serializer.data)
 
+    @action(methods=["get"], detail=False)
+    def status_counts(self, request):
+        """Aggregated task counts for task UI sections."""
+        queryset = self.get_status_count_queryset()
+        counts = queryset.aggregate(
+            all=Count("id"),
+            needs_attention=Count(
+                "id",
+                filter=Q(
+                    status__in=[
+                        PaperlessTask.Status.FAILURE,
+                        PaperlessTask.Status.REVOKED,
+                    ],
+                ),
+            ),
+            in_progress=Count(
+                "id",
+                filter=Q(
+                    status__in=[
+                        PaperlessTask.Status.PENDING,
+                        PaperlessTask.Status.STARTED,
+                    ],
+                ),
+            ),
+            completed=Count("id", filter=Q(status=PaperlessTask.Status.SUCCESS)),
+        )
+        return Response(counts)
+
     @action(methods=["get"], detail=False)
     def active(self, request):
         """Currently pending and running tasks (capped at 50)."""