Skip to content

Commit

Permalink
Fix Dataset.get when there are transforms in the dataset (#45)
Browse files Browse the repository at this point in the history
* Fix dataset get when there are transforms

* Extend test

* Update changelog

* Help linter
  • Loading branch information
zhiltsov-max committed Jun 5, 2024
1 parent d0456d1 commit 453c891
Show file tree
Hide file tree
Showing 3 changed files with 76 additions and 2 deletions.
2 changes: 2 additions & 0 deletions CHANGELOG.md
Original file line number Diff line number Diff line change
Expand Up @@ -94,6 +94,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
(<https://github.com/cvat-ai/datumaro/pull/34>)
- Incorrect computation of binary mask bbox (missed 1 pixel of the size)
(<https://github.com/cvat-ai/datumaro/pull/41>)
- `Dataset.get()` could ignore existing transforms in the dataset
(<https://github.com/cvat-ai/datumaro/pull/45>)

### Security
- TBD
Expand Down
5 changes: 3 additions & 2 deletions datumaro/components/dataset.py
Original file line number Diff line number Diff line change
Expand Up @@ -419,8 +419,9 @@ def _update_status(item_id, new_status: ItemStatus):
source = self._source or DatasetItemStorageDatasetView(
self._storage, categories=self._categories, media_type=media_type
)
transform = None

transform = None
old_ids = None
if self._transforms:
transform = _StackedTransform(source, self._transforms)
if transform.is_local:
Expand Down Expand Up @@ -588,7 +589,7 @@ def get(self, id, subset=None) -> Optional[DatasetItem]:

item = self._storage.get(id, subset)
if item is None and not self.is_cache_initialized():
if self._source.get.__func__ == Extractor.get:
if self._source.get.__func__ == Extractor.get or self._transforms:
# can be improved if IDataset is ABC
self.init_cache()
item = self._storage.get(id, subset)
Expand Down
71 changes: 71 additions & 0 deletions tests/test_dataset.py
Original file line number Diff line number Diff line change
Expand Up @@ -1267,6 +1267,77 @@ def __iter__(self):

self.assertEqual(iter_called, 2)

@mark_requirement(Requirements.DATUM_GENERAL_REQ)
def test_can_get_item_after_local_transforms(self):
iter_called = 0

class TestExtractor(Extractor):
def __iter__(self):
nonlocal iter_called
iter_called += 1
yield from [
DatasetItem(1),
DatasetItem(2),
DatasetItem(3),
DatasetItem(4),
]

dataset = Dataset.from_extractors(TestExtractor())

class TestTransform(ItemTransform):
def transform_item(self, item):
return self.wrap_item(item, id=int(item.id) + 1)

dataset.transform(TestTransform)
dataset.transform(TestTransform)

self.assertEqual(iter_called, 0)

self.assertIsNone(dataset.get(1))
self.assertIsNone(dataset.get(2))
self.assertIsNotNone(dataset.get(3))
self.assertIsNotNone(dataset.get(4))
self.assertIsNotNone(dataset.get(5))
self.assertIsNotNone(dataset.get(6))

self.assertEqual(iter_called, 1)

@mark_requirement(Requirements.DATUM_GENERAL_REQ)
def test_can_get_item_after_nonlocal_transforms(self):
iter_called = 0

class TestExtractor(Extractor):
def __iter__(self):
nonlocal iter_called
iter_called += 1
yield from [
DatasetItem(1),
DatasetItem(2),
DatasetItem(3),
DatasetItem(4),
]

dataset = Dataset.from_extractors(TestExtractor())

class TestTransform(Transform):
def __iter__(self):
for item in self._extractor:
yield self.wrap_item(item, id=int(item.id) + 1)

dataset.transform(TestTransform)
dataset.transform(TestTransform)

self.assertEqual(iter_called, 0)

self.assertIsNone(dataset.get(1))
self.assertIsNone(dataset.get(2))
self.assertIsNotNone(dataset.get(3))
self.assertIsNotNone(dataset.get(4))
self.assertIsNotNone(dataset.get(5))
self.assertIsNotNone(dataset.get(6))

self.assertEqual(iter_called, 2)

@mark_requirement(Requirements.DATUM_GENERAL_REQ)
def test_can_get_subsets_after_local_transforms(self):
iter_called = 0
Expand Down

0 comments on commit 453c891

Please sign in to comment.