Don't export outside annotations (#1729)

* Add option to omit outside annotations

* update changelog

* Fix mot format and test

* Fix outside in mot

* fix repo problem

* t

* Update CHANGELOG.md

Co-authored-by: Nikita Manovich <40690625+nmanovic@users.noreply.github.com>

CHANGELOG.md

@@ -19,6 +19,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Fixed
 - Some objects aren't shown on canvas sometimes. For example after propagation on of objects is invisible (<https://github.com/opencv/cvat/pull/1834>)
+- `outside` annotations should not be in exported images (<https://github.com/opencv/cvat/issues/1620>)
 
 ### Security
 -

cvat/apps/dataset_manager/bindings.py

@@ -442,9 +442,10 @@ class TaskData:
         return None
 
 class CvatTaskDataExtractor(datumaro.SourceExtractor):
-    def __init__(self, task_data, include_images=False):
+    def __init__(self, task_data, include_images=False, include_outside=False):
         super().__init__()
         self._categories = self._load_categories(task_data)
+        self._include_outside = include_outside
 
         dm_items = []
 
@@ -541,6 +542,9 @@ class CvatTaskDataExtractor(datumaro.SourceExtractor):
                 anno_attr['track_id'] = shape_obj.track_id
                 anno_attr['keyframe'] = shape_obj.keyframe
 
+                if not self._include_outside and shape_obj.outside:
+                    continue
+
             anno_points = shape_obj.points
             if shape_obj.type == ShapeType.POINTS:
                 anno = datumaro.Points(anno_points,

cvat/apps/dataset_manager/formats/mot.py

@@ -63,7 +63,7 @@ def _import(src_file, task_data):
                     points=ann.points,
                     occluded=ann.attributes.get('occluded') == True,
                     outside=False,
-                    keyframe=False,
+                    keyframe=True,
                     z_order=ann.z_order,
                     frame=frame_number,
                     attributes=[],
@@ -78,6 +78,11 @@ def _import(src_file, task_data):
         for track in tracks.values():
             # MOT annotations do not require frames to be ordered
             track.shapes.sort(key=lambda t: t.frame)
-            # Set outside=True for the last shape in a track to finish the track
-            track.shapes[-1] = track.shapes[-1]._replace(outside=True)
+            # Append a shape with outside=True to finish the track
+            last_shape = track.shapes[-1]
+            if last_shape.frame + task_data.frame_step <= \
+                    int(task_data.meta['task']['stop_frame']):
+                track.shapes.append(last_shape._replace(outside=True,
+                    frame=last_shape.frame + task_data.frame_step)
+                )
             task_data.add_track(track)

cvat/apps/dataset_manager/tests/_test_formats.py

@@ -77,7 +77,8 @@ from cvat.apps.engine.models import Task
 _setUpModule()
 
 from cvat.apps.dataset_manager.annotation import AnnotationIR
-from cvat.apps.dataset_manager.bindings import TaskData
+from cvat.apps.dataset_manager.bindings import TaskData, CvatTaskDataExtractor
+from cvat.apps.dataset_manager.task import TaskAnnotation
 from cvat.apps.engine.models import Task
 
 
@@ -406,6 +407,22 @@ class TaskExportTest(_DbTestBase):
                     self.assertEqual(len(dataset), task["size"])
                 self._test_export(check, task, format_name, save_images=False)
 
+    def test_can_skip_outside(self):
+        images = self._generate_task_images(3)
+        task = self._generate_task(images)
+        self._generate_annotations(task)
+        task_ann = TaskAnnotation(task["id"])
+        task_ann.init_from_db()
+        task_data = TaskData(task_ann.ir_data, Task.objects.get(pk=task["id"]))
+
+        extractor = CvatTaskDataExtractor(task_data, include_outside=False)
+        dm_dataset = datumaro.components.project.Dataset.from_extractors(extractor)
+        self.assertEqual(4, len(dm_dataset.get("image_1").annotations))
+
+        extractor = CvatTaskDataExtractor(task_data, include_outside=True)
+        dm_dataset = datumaro.components.project.Dataset.from_extractors(extractor)
+        self.assertEqual(5, len(dm_dataset.get("image_1").annotations))
+
     def test_cant_make_rel_frame_id_from_unknown(self):
         images = self._generate_task_images(3)
         images['frame_filter'] = 'step=2'

cvat/apps/engine/tests/_test_rest_api.py

@@ -2965,6 +2965,19 @@ class TaskAnnotationAPITestCase(JobAnnotationAPITestCase):
                         "points": [2.0, 2.1, 77.2, 36.22],
                         "type": "rectangle",
                         "occluded": True,
+                        "outside": False,
+                        "attributes": [
+                            {
+                                "spec_id": task["labels"][0]["attributes"][1]["id"],
+                                "value": task["labels"][0]["attributes"][1]["default_value"]
+                            }
+                        ]
+                    },
+                    {
+                        "frame": 2,
+                        "points": [2.0, 2.1, 77.2, 36.22],
+                        "type": "rectangle",
+                        "occluded": True,
                         "outside": True,
                         "attributes": [
                             {
@@ -2976,19 +2989,27 @@ class TaskAnnotationAPITestCase(JobAnnotationAPITestCase):
                 ]
             }]
             rectangle_tracks_wo_attrs = [{
-                "frame": 1,
+                "frame": 0,
                 "label_id": task["labels"][1]["id"],
                 "group": 0,
                 "attributes": [],
                 "shapes": [
                     {
-                        "frame": 1,
+                        "frame": 0,
                         "attributes": [],
                         "points": [1.0, 2.1, 50.2, 36.6],
                         "type": "rectangle",
                         "occluded": False,
                         "outside": False
                     },
+                    {
+                        "frame": 1,
+                        "attributes": [],
+                        "points": [1.0, 2.1, 51, 36.6],
+                        "type": "rectangle",
+                        "occluded": False,
+                        "outside": False
+                    },
                     {
                         "frame": 2,
                         "attributes": [],

datumaro/datumaro/components/project.py

@@ -104,7 +104,7 @@ class GitWrapper:
     def __init__(self, config=None):
         self.repo = None
 
-        if config is not None and osp.isdir(config.project_dir):
+        if config is not None and config.project_dir:
             self.init(config.project_dir)
 
     @staticmethod
@@ -116,8 +116,12 @@ class GitWrapper:
         spawn = not osp.isdir(cls._git_dir(path))
         repo = git.Repo.init(path=path)
         if spawn:
-            author = git.Actor("Nobody", "nobody@example.com")
-            repo.index.commit('Initial commit', author=author)
+            repo.config_writer().set_value("user", "name", "User") \
+                .set_value("user", "email", "user@nowhere.com") \
+                .release()
+            # gitpython does not support init, use git directly
+            repo.git.init()
+            repo.git.commit('-m', 'Initial commit', '--allow-empty')
         return repo
 
     def init(self, path):
@@ -377,9 +381,10 @@ class Dataset(Extractor):
     def get(self, item_id, subset=None, path=None):
         if path:
             raise KeyError("Requested dataset item path is not found")
-        if subset is None:
-            subset = ''
-        return self._subsets[subset].items[item_id]
+        item_id = str(item_id)
+        subset = subset or ''
+        subset = self._subsets[subset]
+        return subset.items[item_id]
 
     def put(self, item, item_id=None, subset=None, path=None):
         if path:
@@ -567,7 +572,7 @@ class ProjectDataset(Dataset):
             rest_path = path[1:]
             return self._sources[source].get(
                 item_id=item_id, subset=subset, path=rest_path)
-        return self._subsets[subset].items[item_id]
+        return super().get(item_id, subset)
 
     def put(self, item, item_id=None, subset=None, path=None):
         if path is None:
@@ -754,9 +759,10 @@ class Project:
 
     @staticmethod
     def generate(save_dir, config=None):
+        config = Config(config)
+        config.project_dir = save_dir
         project = Project(config)
         project.save(save_dir)
-        project.config.project_dir = save_dir
         return project
 
     @staticmethod