Ignore tasks without data during project export (#6658)

### Motivation and context
- Allow backup and export annotations for a project even if it has tasks
without data
- Add an error message on attempting to create backup of a task that has
no data

CHANGELOG.md

@@ -27,6 +27,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Fixed
 
+- Exporting project when its tasks has not data (<https://github.com/opencv/cvat/pull/6658>)
 - Removing job assignee (<https://github.com/opencv/cvat/pull/6712>)
 - Fixed switching from organization to sandbox while getting a resource (<https://github.com/opencv/cvat/pull/6689>)
 - \[SDK\]: `FileExistsError` exception raised on Windows when a dataset is loaded from cache

cvat/apps/dataset_manager/bindings.py

@@ -949,7 +949,10 @@ class ProjectData(InstanceLabelData):
 
     def _init_tasks(self):
         self._db_tasks: OrderedDict[int, Task] = OrderedDict(
-            ((db_task.id, db_task) for db_task in self._db_project.tasks.order_by("subset","id").all())
+            (
+                (db_task.id, db_task)
+                for db_task in self._db_project.tasks.exclude(data=None).order_by("subset","id").all()
+            )
         )
 
         subsets = set()

cvat/apps/dataset_manager/project.py

@@ -38,7 +38,7 @@ def export_project(project_id, dst_file, format_name,
 class ProjectAnnotationAndData:
     def __init__(self, pk: int):
         self.db_project = models.Project.objects.get(id=pk)
-        self.db_tasks = models.Task.objects.filter(project__id=pk).order_by('id')
+        self.db_tasks = models.Task.objects.filter(project__id=pk).exclude(data=None).order_by('id')
 
         self.task_annotations: dict[int, TaskAnnotation] = dict()
         self.annotation_irs: dict[int, AnnotationIR] = dict()
@@ -98,7 +98,7 @@ class ProjectAnnotationAndData:
         data['server_files'] = list(map(split_name, data['server_files']))
 
         create_task(db_task, data, isDatasetImport=True)
-        self.db_tasks = models.Task.objects.filter(project__id=self.db_project.id).order_by('id')
+        self.db_tasks = models.Task.objects.filter(project__id=self.db_project.id).exclude(data=None).order_by('id')
         self.init_from_db()
         if project_data is not None:
             project_data.new_tasks.add(db_task.id)

cvat/apps/engine/backup.py

@@ -780,7 +780,8 @@ class ProjectExporter(_ExporterBase, _ProjectBackupBase):
 
     def _write_tasks(self, zip_object):
         for idx, db_task in enumerate(self._db_project.tasks.all().order_by('id')):
-            TaskExporter(db_task.id, self._version).export_to(zip_object, self.TASKNAME_TEMPLATE.format(idx))
+            if db_task.data is not None:
+                TaskExporter(db_task.id, self._version).export_to(zip_object, self.TASKNAME_TEMPLATE.format(idx))
 
     def _write_manifest(self, zip_object):
         def serialize_project():

cvat/apps/engine/views.py

@@ -870,9 +870,15 @@ class TaskViewSet(viewsets.GenericViewSet, mixins.ListModelMixin,
             '200': OpenApiResponse(description='Download of file started'),
             '201': OpenApiResponse(description='Output backup file is ready for downloading'),
             '202': OpenApiResponse(description='Creating a backup file has been started'),
+            '400': OpenApiResponse(description='Backup of a task without data is not allowed'),
         })
     @action(methods=['GET'], detail=True, url_path='backup')
     def export_backup(self, request, pk=None):
+        if self.get_object().data is None:
+            return Response(
+                data='Backup of a task without data is not allowed',
+                status=status.HTTP_400_BAD_REQUEST
+            )
         return self.serialize(request, backup.export)
 
     @transaction.atomic

cvat/schema.yml

@@ -4917,6 +4917,8 @@ paths:
           description: Output backup file is ready for downloading
         '202':
           description: Creating a backup file has been started
+        '400':
+          description: Backup of a task without data is not allowed
   /api/tasks/{id}/data/:
     get:
       operationId: tasks_retrieve_data

tests/python/rest_api/test_projects.py

@@ -310,6 +310,37 @@ class TestGetProjectBackup:
 
         self._test_can_get_project_backup(user["username"], project["id"])
 
+    def test_can_get_backup_project_when_some_tasks_have_no_data(self, projects):
+        project = next((p for p in projects if 0 < p["tasks"]["count"]))
+
+        # add empty task to project
+        response = post_method(
+            "admin1", "tasks", {"name": "empty_task", "project_id": project["id"]}
+        )
+        assert response.status_code == HTTPStatus.CREATED
+
+        self._test_can_get_project_backup("admin1", project["id"])
+
+    def test_can_get_backup_project_when_all_tasks_have_no_data(self, projects):
+        project = next((p for p in projects if 0 == p["tasks"]["count"]))
+
+        # add empty tasks to empty project
+        response = post_method(
+            "admin1", "tasks", {"name": "empty_task1", "project_id": project["id"]}
+        )
+        assert response.status_code == HTTPStatus.CREATED
+
+        response = post_method(
+            "admin1", "tasks", {"name": "empty_task2", "project_id": project["id"]}
+        )
+        assert response.status_code == HTTPStatus.CREATED
+
+        self._test_can_get_project_backup("admin1", project["id"])
+
+    def test_can_get_backup_for_empty_project(self, projects):
+        empty_project = next((p for p in projects if 0 == p["tasks"]["count"]))
+        self._test_can_get_project_backup("admin1", empty_project["id"])
+
 
 @pytest.mark.usefixtures("restore_db_per_function")
 class TestPostProjects:
@@ -703,6 +734,37 @@ class TestImportExportDatasetProject:
 
         self._test_export_project(username, project_id, "COCO Keypoints 1.0")
 
+    def test_can_export_dataset_for_empty_project(self, projects):
+        empty_project = next((p for p in projects if 0 == p["tasks"]["count"]))
+        self._test_export_project("admin1", empty_project["id"], "COCO 1.0")
+
+    def test_can_export_project_dataset_when_some_tasks_have_no_data(self, projects):
+        project = next((p for p in projects if 0 < p["tasks"]["count"]))
+
+        # add empty task to project
+        response = post_method(
+            "admin1", "tasks", {"name": "empty_task", "project_id": project["id"]}
+        )
+        assert response.status_code == HTTPStatus.CREATED
+
+        self._test_export_project("admin1", project["id"], "COCO 1.0")
+
+    def test_can_export_project_dataset_when_all_tasks_have_no_data(self, projects):
+        project = next((p for p in projects if 0 == p["tasks"]["count"]))
+
+        # add empty tasks to empty project
+        response = post_method(
+            "admin1", "tasks", {"name": "empty_task1", "project_id": project["id"]}
+        )
+        assert response.status_code == HTTPStatus.CREATED
+
+        response = post_method(
+            "admin1", "tasks", {"name": "empty_task2", "project_id": project["id"]}
+        )
+        assert response.status_code == HTTPStatus.CREATED
+
+        self._test_export_project("admin1", project["id"], "COCO 1.0")
+
 
 @pytest.mark.usefixtures("restore_db_per_function")
 class TestPatchProjectLabel:

tests/python/rest_api/test_tasks.py

@@ -21,7 +21,7 @@ from typing import List, Optional
 import pytest
 from cvat_sdk import Client, Config, exceptions
 from cvat_sdk.api_client import models
-from cvat_sdk.api_client.api_client import ApiClient, Endpoint
+from cvat_sdk.api_client.api_client import ApiClient, ApiException, Endpoint
 from cvat_sdk.core.helpers import get_paginated_collection
 from cvat_sdk.core.proxies.tasks import ResourceType, Task
 from cvat_sdk.core.uploading import Uploader
@@ -1638,6 +1638,18 @@ class TestTaskBackups:
         assert filename.is_file()
         assert filename.stat().st_size > 0
 
+    def test_cannot_export_backup_for_task_without_data(self, tasks):
+        task_id = next(t for t in tasks if t["jobs"]["count"] == 0)["id"]
+        task = self.client.tasks.retrieve(task_id)
+
+        filename = self.tmp_dir / f"task_{task.id}_backup.zip"
+
+        with pytest.raises(ApiException) as exc:
+            task.download_backup(filename)
+
+            assert exc.status == HTTPStatus.BAD_REQUEST
+            assert "Backup of a task without data is not allowed" == exc.body.encode()
+
     @pytest.mark.parametrize("mode", ["annotation", "interpolation"])
     def test_can_import_backup(self, tasks, mode):
         task_json = next(t for t in tasks if t["mode"] == mode)