Fix an issue with Imagenet dataset import (#4861)

CHANGELOG.md

@@ -138,6 +138,7 @@ Skeleton (<https://github.com/cvat-ai/cvat/pull/1>), (<https://github.com/opencv
 - Fix build dependencies for Siammask (<https://github.com/openvinotoolkit/cvat/pull/4486>)
 - Bug: Exif orientation information handled incorrectly (<https://github.com/openvinotoolkit/cvat/pull/4529>)
 - Fixed build of retinanet function image (<https://github.com/cvat-ai/cvat/pull/54>)
+- Bug: Import dataset of Imagenet format fail (<https://github.com/opencv/cvat/issues/4850>)
 
 ## \[2.0.0] - 2022-03-04
 ### Added

cvat/apps/dataset_manager/bindings.py

@@ -1885,17 +1885,24 @@ def load_dataset_data(project_annotation, dataset: dm.Dataset, project_data):
             'data_root': dataset.data_path + osp.sep,
         }
 
+        root_paths = set()
         for dataset_item in subset_dataset:
             if dataset_item.image and dataset_item.image.has_data:
                 dataset_files['media'].append(dataset_item.image.path)
+                data_root = dataset_item.image.path.rsplit(dataset_item.id, 1)
+                if len(data_root) == 2:
+                    root_paths.add(data_root[0])
             elif dataset_item.point_cloud:
                 dataset_files['media'].append(dataset_item.point_cloud)
+                data_root = dataset_item.point_cloud.rsplit(dataset_item.id, 1)
+                if len(data_root) == 2:
+                    root_paths.add(data_root[0])
+
             if isinstance(dataset_item.related_images, list):
                 dataset_files['media'] += \
                     list(map(lambda ri: ri.path, dataset_item.related_images))
 
-        shortes_path = min(dataset_files['media'], key=lambda x: len(Path(x).parts), default=None)
-        if shortes_path is not None:
-            dataset_files['data_root'] = str(Path(shortes_path).parent.absolute()) + osp.sep
+        if len(root_paths):
+            dataset_files['data_root'] = osp.commonpath(root_paths) + osp.sep
 
         project_annotation.add_task(task_fields, dataset_files, project_data)

tests/python/rest_api/test_projects.py

@@ -458,6 +458,23 @@ class TestImportExportDatasetProject:
             )
             assert response.status == HTTPStatus.ACCEPTED
 
+    def test_can_import_export_dataset_with_imagenet_format(self):
+        # https://github.com/opencv/cvat/issues/4850
+        username = "admin1"
+        format_name = "ImageNet 1.0"
+        project_id = 4
+
+        response = self._test_export_project(username, project_id, format_name)
+
+        tmp_file = io.BytesIO(response.data)
+        tmp_file.name = "dataset.zip"
+
+        import_data = {
+            "dataset_file": tmp_file,
+        }
+
+        self._test_import_project(username, project_id, format_name, import_data)
+
 
 @pytest.mark.usefixtures("restore_db_per_function")
 class TestPatchProjectLabel:

tests/python/shared/assets/annotations.json

@@ -361,7 +361,24 @@
           "z_order": 0
         }
       ],
-      "tags": [],
+      "tags": [
+        {
+          "attributes": [],
+          "frame": 2,
+          "group": 0,
+          "id": 1,
+          "label_id": 17,
+          "source": "manual"
+        },
+        {
+          "attributes": [],
+          "frame": 3,
+          "group": 0,
+          "id": 2,
+          "label_id": 16,
+          "source": "manual"
+        }
+      ],
       "tracks": [],
       "version": 0
     },
@@ -1003,7 +1020,24 @@
           "z_order": 0
         }
       ],
-      "tags": [],
+      "tags": [
+        {
+          "attributes": [],
+          "frame": 2,
+          "group": 0,
+          "id": 1,
+          "label_id": 17,
+          "source": "manual"
+        },
+        {
+          "attributes": [],
+          "frame": 3,
+          "group": 0,
+          "id": 2,
+          "label_id": 16,
+          "source": "manual"
+        }
+      ],
       "tracks": [],
       "version": 0
     },

tests/python/shared/assets/cvat_db/data.json

@@ -36,7 +36,7 @@
   "pk": 1,
   "fields": {
     "password": "pbkdf2_sha256$260000$DevmxlmLwciP1P6sZs2Qag$U9DFtjTWx96Sk95qY6UXVcvpdQEP2LcoFBftk5D2RKY=",
-    "last_login": "2022-09-29T08:00:09.733Z",
+    "last_login": "2022-10-17T17:09:16.903Z",
     "is_superuser": true,
     "username": "admin1",
     "first_name": "Admin",
@@ -487,6 +487,14 @@
     "expire_date": "2022-03-02T06:24:53.914Z"
   }
 },
+{
+  "model": "sessions.session",
+  "pk": "d3rhbsiax4etv2gtdkl6j883hvlz21rc",
+  "fields": {
+    "session_data": ".eJxVjMsOwiAQRf-FtSEMDAIu3fcbyPBQqgaS0q6M_y5NutDtOefeN_O0rcVvPS9-TuzCgJ1-WaD4zHUX6UH13nhsdV3mwPeEH7bzqaX8uh7t30GhXsYaQYNDpYMULqECS9IY0uIGxmYUQdoYMCpIeRgYiXVaDXoWgkImZJ8vqTk2zg:1okTc4:kq0YGX6_uyDwCa3VLl1bcp38oBfM6XfHIk0oE4XzpPw",
+    "expire_date": "2022-10-31T17:09:16.915Z"
+  }
+},
 {
   "model": "sessions.session",
   "pk": "dpaw6pntyqwr5l6qjv6zq3yoajp301be",
@@ -2242,7 +2250,7 @@
     "assignee": null,
     "bug_tracker": "",
     "created_date": "2022-06-08T08:32:45.521Z",
-    "updated_date": "2022-09-28T12:26:54.279Z",
+    "updated_date": "2022-10-17T17:09:36.526Z",
     "status": "annotation",
     "organization": 2,
     "source_storage": null,
@@ -2532,7 +2540,7 @@
     "assignee": null,
     "bug_tracker": "",
     "created_date": "2022-06-08T08:33:06.505Z",
-    "updated_date": "2022-06-08T08:33:20.808Z",
+    "updated_date": "2022-10-17T17:09:36.570Z",
     "overlap": 0,
     "segment_size": 5,
     "status": "annotation",
@@ -3574,7 +3582,7 @@
   "fields": {
     "segment": 17,
     "assignee": null,
-    "updated_date": "2022-06-22T09:18:45.296Z",
+    "updated_date": "2022-10-17T17:09:36.571Z",
     "status": "annotation",
     "stage": "annotation",
     "state": "in progress"
@@ -4726,6 +4734,45 @@
     "job": 18
   }
 },
+{
+  "model": "engine.jobcommit",
+  "pk": 73,
+  "fields": {
+    "scope": "create",
+    "owner": [
+      "admin1"
+    ],
+    "timestamp": "2022-10-17T17:09:36.573Z",
+    "data": {
+      "stage": "annotation",
+      "state": "in progress",
+      "assignee": null
+    },
+    "job": 17
+  }
+},
+{
+  "model": "engine.labeledimage",
+  "pk": 1,
+  "fields": {
+    "job": 17,
+    "label": 17,
+    "frame": 2,
+    "group": 0,
+    "source": "manual"
+  }
+},
+{
+  "model": "engine.labeledimage",
+  "pk": 2,
+  "fields": {
+    "job": 17,
+    "label": 16,
+    "frame": 3,
+    "group": 0,
+    "source": "manual"
+  }
+},
 {
   "model": "engine.labeledshape",
   "pk": 1,

tests/python/shared/assets/jobs.json

@@ -235,7 +235,7 @@
       "status": "annotation",
       "stop_frame": 4,
       "task_id": 13,
-      "updated_date": "2022-06-22T09:18:45.296000Z",
+      "updated_date": "2022-10-17T17:09:36.571000Z",
       "url": "http://localhost:8080/api/jobs/17"
     },
     {

tests/python/shared/assets/projects.json

@@ -353,7 +353,7 @@
       "tasks": [
         13
       ],
-      "updated_date": "2022-09-28T12:26:54.279000Z",
+      "updated_date": "2022-10-17T17:09:36.526000Z",
       "url": "http://localhost:8080/api/projects/4"
     },
     {

tests/python/shared/assets/tasks.json

@@ -303,7 +303,7 @@
       "status": "annotation",
       "subset": "",
       "target_storage": null,
-      "updated_date": "2022-06-08T08:33:20.808000Z",
+      "updated_date": "2022-10-17T17:09:36.570000Z",
       "url": "http://localhost:8080/api/tasks/13"
     },
     {

tests/python/shared/assets/users.json

@@ -310,7 +310,7 @@
       "is_active": true,
       "is_staff": true,
       "is_superuser": true,
-      "last_login": "2022-09-29T08:00:09.733000Z",
+      "last_login": "2022-10-17T17:09:16.903140Z",
       "last_name": "First",
       "url": "http://localhost:8080/api/users/1",
       "username": "admin1"