Merge branch 'refs/heads/develop' into dl/yolo-preserve-rotation

# Conflicts:
#	cvat/requirements/base.in
#	cvat/requirements/base.txt

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.txt

@@ -1,4 +1,4 @@
-# SHA1:8d2c3b645ed1f9efb7ca325d5da7eb9e07aced87
+# SHA1:d210065224fc86c1fba6b5d22b77ab38d02bcbb1
 #
 # This file is autogenerated by pip-compile-multi
 # To update, run:
@@ -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

tests/yarn.lock

@@ -1146,9 +1146,9 @@ crc32-stream@^4.0.2:
     readable-stream "^3.4.0"
 
 cross-spawn@^7.0.0, cross-spawn@^7.0.3:
-  version "7.0.3"
-  resolved "https://registry.npmjs.org/cross-spawn/-/cross-spawn-7.0.3.tgz"
-  integrity sha512-iRDPJKUPVEND7dHPO8rkbOnPpyDygcDFtWjpeWNCgy8WP2rXcxXL8TskReQl6OrB2G7+UJrags1q15Fudc7G6w==
+  version "7.0.6"
+  resolved "https://registry.yarnpkg.com/cross-spawn/-/cross-spawn-7.0.6.tgz#8a58fe78f00dcd70c370451759dfbfaf03e8ee9f"
+  integrity sha512-uV2QOWP2nWzsy2aMp8aRibhi9dlzF5Hgh5SHaB9OiTGEyDTiJJyx0uy51QXdyWbtAHNua4XJzUKca3OzKUd3vA==
   dependencies:
     path-key "^3.1.0"
     shebang-command "^2.0.0"

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