Fix backups for tasks using a mounted file share

Currently, backing up such a task succeeds, but restoring fails with "No
such file or directory: <root>/share/manifest.jsonl".

IMO, the problem is not really in the restore logic, but in the backup
logic. Share task backups are "heavyweight", in that they include a copy of
the media files. But the `storage` field in `task.json` is still set to
"share". As a result, CVAT tries to import it as a share task, and that
logic is broken.

I think we should handle such backups consistently with heavyweight cloud
storage backups, and set `storage` to "local". The restore logic will then
work perfectly well. In the future, if we happen to implement lightweight
backups for share tasks, we can use `storage: "share"` for those.

Author: SpecLad

changelog.d/20251104_153655_roman_restore_backup_from_share.md

@@ -0,0 +1,4 @@
+### Fixed
+
+- Backups of tasks created from a mounted file share no longer fail to import
+  (<https://github.com/cvat-ai/cvat/pull/9972>)

cvat/apps/engine/backup.py

@@ -619,7 +619,10 @@ def serialize_data():
                 ]
                 data['validation_layout'] = validation_params
 
-            if self._db_data.storage == StorageChoice.CLOUD_STORAGE and not self._lightweight:
+            if (
+                self._db_data.storage == StorageChoice.SHARE
+                or self._db_data.storage == StorageChoice.CLOUD_STORAGE and not self._lightweight
+            ):
                 data["storage"] = StorageChoice.LOCAL
 
             return self._prepare_data_meta(data)

tests/docker-compose.file_share.yml

@@ -1,10 +1,7 @@
 services:
-  cvat_worker_import:
-    volumes:
-      - ./tests/mounted_file_share:/home/django/share:rw
-  cvat_server:
-    volumes:
-      - ./tests/mounted_file_share:/home/django/share:rw
-  cvat_worker_chunks:
+  cvat_server: &share-config
     volumes:
       - ./tests/mounted_file_share:/home/django/share:rw
+  cvat_worker_chunks: *share-config
+  cvat_worker_import: *share-config
+  cvat_worker_export: *share-config

tests/python/rest_api/test_tasks.py

@@ -1298,6 +1298,33 @@ def test_cannot_export_backup_for_task_without_data(self, tasks):
 
         assert "Backup of a task without data is not allowed" in str(capture.value.body)
 
+    @pytest.mark.with_external_services
+    def test_can_export_and_import_backup_task_with_mounted_share(self):
+        share_content = [
+            "images/image_1.jpg",
+            "images/image_2.jpg",
+        ]
+        task_spec = {
+            "name": "Task with files from mounted share",
+            "labels": [{"name": "car"}],
+        }
+        data_spec = {"image_quality": 75, "server_files": share_content}
+        task_id, _ = create_task(self.user, task_spec, data_spec)
+
+        task = self.client.tasks.retrieve(task_id)
+
+        filename = self.tmp_dir / f"share_task_{task.id}_backup.zip"
+        task.download_backup(filename)
+
+        with zipfile.ZipFile(filename, "r") as zf:
+            files_in_data = {
+                name.removeprefix("data/") for name in zf.namelist() if name.startswith("data/")
+            }
+
+        assert files_in_data == {"manifest.jsonl", *share_content}
+
+        self._test_can_restore_task_from_backup(task_id)
+
     @pytest.mark.with_external_services
     @pytest.mark.parametrize("lightweight_backup", [True, False])
     def test_can_export_and_import_backup_task_with_cloud_storage(self, lightweight_backup):
@@ -1404,6 +1431,9 @@ def _test_can_restore_task_from_backup(self, task_id: int, lightweight_backup: b
             assert new_meta["storage"] == ("cloud_storage" if lightweight_backup else "local")
             assert new_meta["cloud_storage_id"] is None
             exclude_regex_paths.extend([r"root\['cloud_storage_id'\]", r"root\['storage'\]"])
+        if old_meta["storage"] == "share":
+            assert new_meta["storage"] == "local"
+            exclude_regex_paths.append(r"root\['storage'\]")
 
         assert (
             DeepDiff(