Regression test for missing frames after exporting a CVAT dataset

.vscode/launch.json

@@ -444,6 +444,8 @@
             "python": "${command:python.interpreterPath}",
             "module": "pytest",
             "args": [
+                "--verbose",
+                "--no-cov", // vscode debugger might not work otherwise
                 "tests/python/rest_api/"
             ],
             "cwd": "${workspaceFolder}",

tests/python/requirements.txt

@@ -10,3 +10,5 @@ Pillow==10.3.0
 python-dateutil==2.8.2
 pyyaml==6.0.0
 numpy==2.0.0
+
+# TODO: update pytest to 7.0.0 and pytest-timeout to 2.3.1 (better debug in vscode)

tests/python/rest_api/test_tasks.py

@@ -24,7 +24,7 @@
 from itertools import chain, groupby, product
 from math import ceil
 from operator import itemgetter
-from pathlib import Path
+from pathlib import Path, PurePosixPath
 from tempfile import NamedTemporaryFile, TemporaryDirectory
 from time import sleep, time
 from typing import Any, Callable, ClassVar, Optional, Union
@@ -1012,6 +1012,71 @@ def test_datumaro_export_without_annotations_includes_image_info(
                     if "size" in related_image:
                         assert tuple(related_image["size"]) > (0, 0)
 
+    @pytest.mark.parametrize(
+            "format_name, num_frames, frame_step, expected_frames",
+            [
+                ("Datumaro 1.0", 100, 5, 20),
+                ("COCO 1.0", 100, 5, 20),
+                ("CVAT for video 1.1", 100, 5, 20), # no remainder
+                ("CVAT for video 1.1", 97, 5, 20), # prime
+                ("CVAT for video 1.1", 97, 2, 49),
+                ("CVAT for video 1.1", 100, 3, 34), # three
+                    # we assert that expected frames are ceil(frames / step)
+            ]
+    )
+    def test_export_with_non_default_frame_step(
+        self,
+        # fixtures
+        tmp_path: Path,
+        admin_user: str,
+        # params
+        format_name: str,
+        num_frames: int,
+        frame_step: int,
+        expected_frames: int,
+    ):
+        # parameter validation
+        assert expected_frames == math.ceil(num_frames / frame_step), 'Test params are wrong'
+
+        spec = {
+            'name': f"test_video_frames_in_{format_name}_after_export",
+            "labels": [{"name": "goofy ahh car"}]
+        }
+
+        data = {
+            'image_quality': 70,
+            'client_files': [generate_video_file(num_frames)],
+            'frame_filter': f"step={frame_step}"
+        }
+
+        # create a task and get its instance
+        (task_id, _) = create_task(admin_user, spec, data)
+        with make_sdk_client(admin_user) as client:
+            task_obj: Task  = client.tasks.retrieve(task_id)
+
+        # export the video
+        dataset_file = tmp_path / "dataset.zip"
+        task_obj.export_dataset(format_name, dataset_file, include_images=True)
+
+        def get_png_index(zinfo: zipfile.ZipInfo) -> int:
+            name = PurePosixPath(zinfo.filename)
+            if name.suffix.lower() != '.png':
+                return -1
+            name = os.path.basename(name).removesuffix(name.suffix)
+            idx = name[name.rfind('_') + 1:]
+            assert idx.isnumeric()
+            return int(idx)
+
+        # get frames and sort them
+        with zipfile.ZipFile(dataset_file) as dataset:
+            frames = [png_idx for png_idx in map(get_png_index, dataset.filelist) if png_idx != -1]
+            frames.sort()
+
+
+        assert len(frames) == expected_frames, 'Some frames were lost'
+        assert frames == list(range(0, num_frames, frame_step)), 'Some frames are wrong'
+
+
 
 @pytest.mark.usefixtures("restore_db_per_function")
 @pytest.mark.usefixtures("restore_cvat_data_per_function")