[GSoC2024] Fix missing related images 3d dataset exports

changelog.d/20240412_013221_jbargav025_related_imgs_in_export.md

@@ -0,0 +1,4 @@
+### Fixed
+
+- Fixed missing related images in Datumaro 3D exports
+  (<https://github.com/cvat-ai/cvat/pull/7699>)

cvat/apps/dataset_manager/bindings.py

@@ -1534,9 +1534,13 @@ def __init__(
                         attributes["labels"].append({"label_id": idx, "name": label["name"], "color": label["color"], "type": label["type"]})
                         attributes["track_id"] = -1
 
+                extra_images_paths = dm_image[1]
+                extra_images_objects = [dm.Image(path=image_path) for image_path in extra_images_paths]
+
                 dm_item = dm.DatasetItem(
                     id=osp.splitext(osp.split(frame_data.name)[-1])[0],
-                    annotations=dm_anno, media=PointCloud(dm_image[0]), related_images=dm_image[1],
+                    annotations=dm_anno,
+                    media=PointCloud(dm_image[0], extra_images=extra_images_objects),
                     attributes=attributes
                 )
 
@@ -1624,9 +1628,13 @@ def __init__(
                         attributes["labels"].append({"label_id": idx, "name": label["name"], "color": label["color"], "type": label["type"]})
                         attributes["track_id"] = -1
 
+                extra_images_paths = dm_image[1]
+                extra_images_objects = [dm.Image(path=image_path) for image_path in extra_images_paths]
+
                 dm_item = dm.DatasetItem(
                     id=osp.splitext(osp.split(frame_data.name)[-1])[0],
-                    annotations=dm_anno, media=PointCloud(dm_image[0]), related_images=dm_image[1],
+                    annotations=dm_anno,
+                    media=PointCloud(dm_image[0], extra_images=extra_images_objects),
                     attributes=attributes, subset=frame_data.subset
                 )
             dm_items.append(dm_item)

cvat/apps/engine/tests/test_rest_api_3D.py

@@ -250,26 +250,56 @@ def etree_to_dict(t):
                     self.assertEqual(len(items), len(task_data["shapes"]))
         elif format_name == "Sly Point Cloud Format 1.0":
             with tempfile.TemporaryDirectory() as tmp_dir:
-                checking_files = [osp.join(tmp_dir, "key_id_map.json"),
-                                  osp.join(tmp_dir, "meta.json"),
-                                  osp.join(tmp_dir, "ds0", "ann", "000001.pcd.json"),
-                                  osp.join(tmp_dir, "ds0", "ann", "000002.pcd.json"),
-                                  osp.join(tmp_dir, "ds0", "ann","000003.pcd.json")]
-                if related_files:
-                    checking_files.extend([osp.join(tmp_dir, "ds0", "related_images", "000001.pcd_pcd", "000001.png.json"),
-                                  osp.join(tmp_dir, "ds0", "related_images", "000002.pcd_pcd", "000002.png.json"),
-                                  osp.join(tmp_dir, "ds0", "related_images", "000003.pcd_pcd", "000003.png.json")])
                 zipfile.ZipFile(content).extractall(tmp_dir)
+                checking_files = [
+                    osp.join(tmp_dir, "key_id_map.json"),
+                    osp.join(tmp_dir, "meta.json"),
+                    osp.join(tmp_dir, "ds0", "ann", "000001.pcd.json"),
+                    osp.join(tmp_dir, "ds0", "ann", "000002.pcd.json"),
+                    osp.join(tmp_dir, "ds0", "ann", "000003.pcd.json")
+                ]
+                point_cloud_files = [
+                    osp.join(tmp_dir, "pointcloud", "ds0", "000001.pcd"),
+                    osp.join(tmp_dir, "pointcloud",  "ds0", "000002.pcd"),
+                    osp.join(tmp_dir, "pointcloud", "ds0", "000003.pcd")
+                ]
+                point_cloud_files_present = any(osp.exists(f) for f in point_cloud_files)
+                if related_files and point_cloud_files_present:
+                    checking_files.extend([
+                        osp.join(tmp_dir, "ds0", "related_images", "000001_pcd", "000001.png.json"),
+                        osp.join(tmp_dir, "ds0", "related_images", "000002_pcd", "000002.png.json"),
+                        osp.join(tmp_dir, "ds0", "related_images", "000003_pcd", "000003.png.json")
+                    ])
                 jsons = glob(osp.join(tmp_dir, '**', '*.json'), recursive=True)
                 self.assertTrue(jsons)
                 self.assertTrue(set(checking_files).issubset(set(jsons)))
+        elif format_name == "Datumaro 3D 1.0":
+            with tempfile.TemporaryDirectory() as tmp_dir:
+                zipfile.ZipFile(content).extractall(tmp_dir)
+                extracted_files = set(
+                    os.path.relpath(os.path.join(root, file), tmp_dir)
+                    for root, dirs, files in os.walk(tmp_dir)
+                    for file in files
+                )
+                expected_files = {'annotations/default.json'}
+                point_cloud_files = {'point_clouds/default/000001.pcd', 'point_clouds/default/000002.pcd', 'point_clouds/default/000003.pcd'}
+                # Check for point cloud files
+                if not point_cloud_files.isdisjoint(extracted_files):
+                    expected_files.update(point_cloud_files)
+                    if related_files:
+                        expected_files.update([
+                            'related_images/default/000001/image_0.png',
+                            'related_images/default/000002/image_0.png',
+                            'related_images/default/000003/image_0.png'
+                        ])
+                self.assertEqual(extracted_files, expected_files)
 
 
 class Task3DTest(_DbTestBase):
     @classmethod
     def setUpClass(cls):
         super().setUpClass()
-        cls.format_names = ["Sly Point Cloud Format 1.0", "Kitti Raw Format 1.0"]
+        cls.format_names = ["Sly Point Cloud Format 1.0", "Kitti Raw Format 1.0", "Datumaro 3D 1.0"]
         cls._image_sizes = {}
         cls.pointcloud_pcd_filename = "test_canvas3d.zip"
         cls.pointcloud_pcd_path = osp.join(os.path.dirname(__file__), 'assets', cls.pointcloud_pcd_filename)
@@ -761,5 +791,5 @@ def test_api_v2_export_dataset(self):
                             with open(file_name, "wb") as f:
                                 f.write(content.getvalue())
                         self.assertEqual(osp.exists(file_name), edata['file_exists'])
-                        self._check_dump_content(content, task_ann_prev.data, format_name,related_files=False)
+                        self._check_dump_content(content, task_ann_prev.data, format_name, related_files=True)