ultralytics yolo support for tracks (#8883)

<!-- Raise an issue to propose your change
(https://github.com/cvat-ai/cvat/issues).
It helps to avoid duplication of efforts from multiple independent
contributors.
Discuss your ideas with maintainers to be sure that changes will be
approved and merged.
Read the [Contribution guide](https://docs.cvat.ai/docs/contributing/).
-->

<!-- Provide a general summary of your changes in the Title above -->

### Motivation and context
<!-- Why is this change required? What problem does it solve? If it
fixes an open
issue, please link to the issue here. Describe your changes in detail,
add
screenshots. -->
Can now import Ultralytics Yolo formats with track_id
Can now export Ultralytics Yolo Detection with track_id

depends on cvat-ai/datumaro#70


### How has this been tested?
<!-- Please describe in detail how you tested your changes.
Include details of your testing environment, and the tests you ran to
see how your change affects other areas of the code, etc. -->

### Checklist
<!-- Go over all the following points, and put an `x` in all the boxes
that apply.
If an item isn't applicable for some reason, then ~~explicitly
strikethrough~~ the whole
line. If you don't do that, GitHub will show incorrect progress for the
pull request.
If you're unsure about any of these, don't hesitate to ask. We're here
to help! -->
- [ ] I submit my changes into the `develop` branch
- [ ] I have created a changelog fragment <!-- see top comment in
CHANGELOG.md -->
- [ ] I have updated the documentation accordingly
- [ ] I have added tests to cover my changes
- [ ] I have linked related issues (see [GitHub docs](

https://help.github.com/en/github/managing-your-work-on-github/linking-a-pull-request-to-an-issue#linking-a-pull-request-to-an-issue-using-a-keyword))
- [ ] I have increased versions of npm packages if it is necessary

([cvat-canvas](https://github.com/cvat-ai/cvat/tree/develop/cvat-canvas#versioning),

[cvat-core](https://github.com/cvat-ai/cvat/tree/develop/cvat-core#versioning),

[cvat-data](https://github.com/cvat-ai/cvat/tree/develop/cvat-data#versioning)
and

[cvat-ui](https://github.com/cvat-ai/cvat/tree/develop/cvat-ui#versioning))

### License

- [ ] I submit _my code changes_ under the same [MIT License](
https://github.com/cvat-ai/cvat/blob/develop/LICENSE) that covers the
project.
  Feel free to contact the maintainers if that's a concern.


<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

## Release Notes

- **New Features**
  - Expanded dataset import formats to include multiple YOLO formats.
- Introduced new classes for annotation transformations:
`MaskConverter`, `EllipsesToMasks`, `MaskToPolygonTransformation`, and
`SetKeyframeForEveryTrackShape`.
  - Added support for exporting YOLO detection data with track IDs.

- **Bug Fixes**
- Minor formatting adjustments to ensure existing formats remain intact.

- **Tests**
- Added new export and import formats for "Ultralytics YOLO Detection
Track 1.0" in the test suite.
  
- **Chores**
- Updated the `datumaro` package to a newer version in requirements
files.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->

---------

Co-authored-by: Maxim Zhiltsov <[email protected]>

changelog.d/20250107_163424_dmitrii.lavrukhin_yolo_tracks.md

@@ -0,0 +1,4 @@
+### Added
+
+- Ultralytics YOLO formats now support tracks
+  (<https://github.com/cvat-ai/cvat/pull/8883>)

cvat/apps/dataset_manager/bindings.py

@@ -2185,7 +2185,11 @@ def import_dm_annotations(dm_dataset: dm.Dataset, instance_data: Union[ProjectDa
         'coco',
         'coco_instances',
         'coco_person_keypoints',
-        'voc'
+        'voc',
+        'yolo_ultralytics_detection',
+        'yolo_ultralytics_segmentation',
+        'yolo_ultralytics_oriented_boxes',
+        'yolo_ultralytics_pose',
     ]
 
     label_cat = dm_dataset.categories()[dm.AnnotationType.label]

cvat/apps/dataset_manager/formats/transformations.py

@@ -37,6 +37,7 @@ def transform_item(self, item):
 
         return item.wrap(annotations=annotations)
 
+
 class MaskConverter:
     @staticmethod
     def cvat_rle_to_dm_rle(shape, img_h: int, img_w: int) -> dm.RleMask:
@@ -100,6 +101,7 @@ def rle(cls, arr: np.ndarray) -> list[int]:
 
         return cvat_rle
 
+
 class EllipsesToMasks:
     @staticmethod
     def convert_ellipse(ellipse, img_h, img_w):
@@ -115,6 +117,7 @@ def convert_ellipse(ellipse, img_h, img_w):
         return dm.RleMask(rle=rle, label=ellipse.label, z_order=ellipse.z_order,
             attributes=ellipse.attributes, group=ellipse.group)
 
+
 class MaskToPolygonTransformation:
     """
     Manages common logic for mask to polygons conversion in dataset import.
@@ -130,3 +133,13 @@ def convert_dataset(cls, dataset, **kwargs):
         if kwargs.get('conv_mask_to_poly', True):
             dataset.transform('masks_to_polygons')
         return dataset
+
+
+class SetKeyframeForEveryTrackShape(dm.ItemTransform):
+    def transform_item(self, item):
+        annotations = []
+        for ann in item.annotations:
+            if "track_id" in ann.attributes:
+                ann = ann.wrap(attributes=dict(ann.attributes, keyframe=True))
+            annotations.append(ann)
+        return item.wrap(annotations=annotations)

cvat/apps/dataset_manager/formats/yolo.py

@@ -4,14 +4,17 @@
 # SPDX-License-Identifier: MIT
 import os.path as osp
 from glob import glob
+from typing import Callable, Optional
 
 from datumaro.components.annotation import AnnotationType
 from datumaro.components.extractor import DatasetItem
 from datumaro.components.project import Dataset
 from pyunpack import Archive
 
 from cvat.apps.dataset_manager.bindings import (
+    CommonData,
     GetCVATDataExtractor,
+    ProjectData,
     detect_dataset,
     find_dataset_root,
     import_dm_annotations,
@@ -20,12 +23,21 @@
 from cvat.apps.dataset_manager.util import make_zip_archive
 
 from .registry import dm_env, exporter, importer
+from .transformations import SetKeyframeForEveryTrackShape
 
 
-def _export_common(dst_file, temp_dir, instance_data, format_name, *, save_images=False):
+def _export_common(
+    dst_file: str,
+    temp_dir: str,
+    instance_data: ProjectData | CommonData,
+    format_name: str,
+    *,
+    save_images: bool = False,
+    **kwargs
+):
     with GetCVATDataExtractor(instance_data, include_images=save_images) as extractor:
         dataset = Dataset.from_extractors(extractor, env=dm_env)
-        dataset.export(temp_dir, format_name, save_images=save_images)
+        dataset.export(temp_dir, format_name, save_images=save_images, **kwargs)
 
     make_zip_archive(temp_dir, dst_file)
 
@@ -37,12 +49,12 @@ def _export_yolo(*args, **kwargs):
 
 def _import_common(
     src_file,
-    temp_dir,
-    instance_data,
-    format_name,
+    temp_dir: str,
+    instance_data: ProjectData | CommonData,
+    format_name: str,
     *,
-    load_data_callback=None,
-    import_kwargs=None,
+    load_data_callback: Optional[Callable] = None,
+    import_kwargs: dict | None = None,
     **kwargs
 ):
     Archive(src_file.name).extractall(temp_dir)
@@ -67,6 +79,7 @@ def _import_common(
     detect_dataset(temp_dir, format_name=format_name, importer=dm_env.importers.get(format_name))
     dataset = Dataset.import_from(temp_dir, format_name,
         env=dm_env, image_info=image_info, **(import_kwargs or {}))
+    dataset = dataset.transform(SetKeyframeForEveryTrackShape)
     if load_data_callback is not None:
         load_data_callback(dataset, instance_data)
     import_dm_annotations(dataset, instance_data)
@@ -82,6 +95,11 @@ def _export_yolo_ultralytics_detection(*args, **kwargs):
     _export_common(*args, format_name='yolo_ultralytics_detection', **kwargs)
 
 
+@exporter(name='Ultralytics YOLO Detection Track', ext='ZIP', version='1.0')
+def _export_yolo_ultralytics_detection_track(*args, **kwargs):
+    _export_common(*args, format_name='yolo_ultralytics_detection', write_track_id=True, **kwargs)
+
+
 @exporter(name='Ultralytics YOLO Oriented Bounding Boxes', ext='ZIP', version='1.0')
 def _export_yolo_ultralytics_oriented_boxes(*args, **kwargs):
     _export_common(*args, format_name='yolo_ultralytics_oriented_boxes', **kwargs)

cvat/apps/dataset_manager/tests/assets/annotations.json

@@ -1008,6 +1008,54 @@
     ],
     "tracks": []
   },
+  "Ultralytics YOLO Detection Track 1.0": {
+    "version": 0,
+    "tags": [],
+    "shapes": [
+      {
+        "type": "rectangle",
+        "occluded": false,
+        "z_order": 0,
+        "points": [0.3, 0.1, 0.2, 0.8],
+        "frame": 0,
+        "label_id": null,
+        "group": 0,
+        "source": "manual",
+        "attributes": []
+      }
+    ],
+    "tracks": [
+      {
+        "frame": 0,
+        "label_id": null,
+        "group": 0,
+        "source": "manual",
+        "shapes": [
+          {
+            "type": "rectangle",
+            "occluded": false,
+            "z_order": 0,
+            "points": [0.2, 0.1, 0.2, 0.8],
+            "frame": 0,
+            "outside": false,
+            "attributes": [],
+            "keyframe": true
+          },
+          {
+            "type": "rectangle",
+            "occluded": false,
+            "z_order": 0,
+            "points": [0.4, 0.1, 0.2, 0.8],
+            "frame": 1,
+            "outside": true,
+            "attributes": [],
+            "keyframe": true
+          }
+        ],
+        "attributes": []
+      }
+    ]
+  },
   "Ultralytics YOLO Oriented Bounding Boxes 1.0": {
     "version": 0,
     "tags": [],

cvat/apps/dataset_manager/tests/test_formats.py

@@ -300,6 +300,7 @@ def test_export_formats_query(self):
             'Ultralytics YOLO Classification 1.0',
             'Ultralytics YOLO Oriented Bounding Boxes 1.0',
             'Ultralytics YOLO Detection 1.0',
+            'Ultralytics YOLO Detection Track 1.0',
             'Ultralytics YOLO Pose 1.0',
             'Ultralytics YOLO Segmentation 1.0',
         })

cvat/apps/dataset_manager/tests/test_rest_api_formats.py

@@ -59,6 +59,7 @@
     "Ultralytics YOLO Classification 1.0",
     "YOLO 1.1",
     "Ultralytics YOLO Detection 1.0",
+    "Ultralytics YOLO Detection Track 1.0",
     "Ultralytics YOLO Segmentation 1.0",
     "Ultralytics YOLO Oriented Bounding Boxes 1.0",
     "Ultralytics YOLO Pose 1.0",
@@ -980,6 +981,8 @@ def test_api_v2_rewriting_annotations(self):
 
                     if dump_format_name == "CVAT for images 1.1" or dump_format_name == "CVAT for video 1.1":
                         dump_format_name = "CVAT 1.1"
+                    elif dump_format_name == "Ultralytics YOLO Detection Track 1.0":
+                        dump_format_name = "Ultralytics YOLO Detection 1.0"
                     url = self._generate_url_upload_tasks_annotations(task_id, dump_format_name)
 
                     with open(file_zip_name, 'rb') as binary_file:
@@ -1093,6 +1096,8 @@ def test_api_v2_tasks_annotations_dump_and_upload_with_datumaro(self):
                         # upload annotations
                         if dump_format_name in ["CVAT for images 1.1", "CVAT for video 1.1"]:
                             upload_format_name = "CVAT 1.1"
+                        elif dump_format_name in ['Ultralytics YOLO Detection Track 1.0']:
+                            upload_format_name = 'Ultralytics YOLO Detection 1.0'
                         else:
                             upload_format_name = dump_format_name
                         url = self._generate_url_upload_tasks_annotations(task_id, upload_format_name)

cvat/apps/engine/tests/test_rest_api.py

@@ -6398,6 +6398,9 @@ def _get_initial_annotation(annotation_format):
                 formats['CVAT for video 1.1'] = 'CVAT 1.1'
             if 'CVAT for images 1.1' in export_formats:
                 formats['CVAT for images 1.1'] = 'CVAT 1.1'
+        if 'Ultralytics YOLO Detection 1.0' in import_formats:
+            if 'Ultralytics YOLO Detection Track 1.0' in export_formats:
+                formats['Ultralytics YOLO Detection Track 1.0'] = 'Ultralytics YOLO Detection 1.0'
         if set(import_formats) ^ set(export_formats):
             # NOTE: this may not be an error, so we should not fail
             print("The following import formats have no pair:",

cvat/requirements/base.in

@@ -12,7 +12,7 @@ azure-storage-blob==12.13.0
 boto3==1.17.61
 clickhouse-connect==0.6.8
 coreapi==2.3.3
-datumaro @ git+https://github.com/cvat-ai/datumaro.git@02507a9d41413327ca2d67a4349e5553ce763a7c
+datumaro @ git+https://github.com/cvat-ai/datumaro.git@fad7636b79889f0c7b8fb0c3010b894324c4c18a
 dj-pagination==2.5.0
 # Despite direct indication allauth in requirements we should keep 'with_social' for dj-rest-auth
 # to avoid possible further versions conflicts (we use registration functionality)

cvat/requirements/base.txt

@@ -1,4 +1,4 @@
-# SHA1:8d2c3b645ed1f9efb7ca325d5da7eb9e07aced87
+# SHA1:d210065224fc86c1fba6b5d22b77ab38d02bcbb1
 #
 # This file is autogenerated by pip-compile-multi
 # To update, run:
@@ -56,7 +56,7 @@ cryptography==44.0.0
     #   pyjwt
 cycler==0.12.1
     # via matplotlib
-datumaro @ git+https://github.com/cvat-ai/datumaro.git@02507a9d41413327ca2d67a4349e5553ce763a7c
+datumaro @ git+https://github.com/cvat-ai/datumaro.git@fad7636b79889f0c7b8fb0c3010b894324c4c18a
     # via -r cvat/requirements/base.in
 defusedxml==0.7.1
     # via
@@ -242,7 +242,7 @@ pyjwt[crypto]==2.10.1
     # via django-allauth
 pylogbeat==2.0.1
     # via python-logstash-async
-pyparsing==3.2.0
+pyparsing==3.2.1
     # via matplotlib
 pyrsistent==0.20.0
     # via jsonschema
@@ -308,7 +308,7 @@ rq-scheduler==0.13.1
     # via -r cvat/requirements/base.in
 rsa==4.9
     # via google-auth
-ruamel-yaml==0.18.7
+ruamel-yaml==0.18.8
     # via datumaro
 ruamel-yaml-clib==0.2.12
     # via ruamel-yaml

utils/dataset_manifest/requirements.txt

@@ -13,7 +13,7 @@ numpy==1.22.4
     # via opencv-python-headless
 opencv-python-headless==4.10.0.84
     # via -r utils/dataset_manifest/requirements.in
-pillow==11.0.0
+pillow==11.1.0
     # via -r utils/dataset_manifest/requirements.in
 tqdm==4.67.1
     # via -r utils/dataset_manifest/requirements.in