tfds-nightly 4.9.9.dev202507160045__py3-none-any.whl → 4.9.9.dev202507180045__py3-none-any.whl

tensorflow_datasets/core/dataset_builders/croissant_builder.py

@@ -48,7 +48,9 @@ from tensorflow_datasets.core import dataset_info
 from tensorflow_datasets.core import download
 from tensorflow_datasets.core import split_builder as split_builder_lib
 from tensorflow_datasets.core import splits as splits_lib
+from tensorflow_datasets.core.features import audio_feature
 from tensorflow_datasets.core.features import bounding_boxes
+from tensorflow_datasets.core.features import bounding_boxes_utils as bb_utils
 from tensorflow_datasets.core.features import feature as feature_lib
 from tensorflow_datasets.core.features import features_dict
 from tensorflow_datasets.core.features import image_feature
@@ -174,8 +176,21 @@ def datatype_converter(
     feature = image_feature.Image(doc=field.description)
   elif field_data_type == mlc.DataType.BOUNDING_BOX:
     # TFDS uses REL_YXYX by default, but Hugging Face doesn't enforce a format.
+    if bbox_format := field.source.format:
+      try:
+        bbox_format = bb_utils.BBoxFormat(bbox_format)
+      except ValueError as e:
+        raise ValueError(
+            f'Unsupported bounding box format: {bbox_format}. Currently'
+            ' supported bounding box formats are: '
+            f'{[format.value for format in bb_utils.BBoxFormat]}'
+        ) from e
     feature = bounding_boxes.BBoxFeature(
-        doc=field.description, bbox_format=None
+        doc=field.description, bbox_format=bbox_format
+    )
+  elif field_data_type == mlc.DataType.AUDIO_OBJECT:
+    feature = audio_feature.Audio(
+        doc=field.description, sample_rate=field.source.sampling_rate
     )
   else:
     raise ValueError(f'Unknown data type: {field_data_type}.')

tensorflow_datasets/core/dataset_builders/croissant_builder_test.py

@@ -20,7 +20,9 @@ import pytest
 from tensorflow_datasets import testing
 from tensorflow_datasets.core import file_adapters
 from tensorflow_datasets.core.dataset_builders import croissant_builder
+from tensorflow_datasets.core.features import audio_feature
 from tensorflow_datasets.core.features import bounding_boxes
+from tensorflow_datasets.core.features import bounding_boxes_utils as bb_utils
 from tensorflow_datasets.core.features import features_dict
 from tensorflow_datasets.core.features import image_feature
 from tensorflow_datasets.core.features import sequence_feature
@@ -129,6 +131,27 @@ def test_simple_datatype_converter(
   assert actual_feature == expected_feature
 
 
+def test_bbox_datatype_converter():
+  field = mlc.Field(
+      data_types=mlc.DataType.BOUNDING_BOX,
+      description="Bounding box feature",
+      source=mlc.Source(format="XYWH"),
+  )
+  actual_feature = croissant_builder.datatype_converter(field)
+  assert isinstance(actual_feature, bounding_boxes.BBoxFeature)
+  assert actual_feature.bbox_format == bb_utils.BBoxFormat.XYWH
+
+
+def test_bbox_datatype_converter_with_invalid_format():
+  field = mlc.Field(
+      data_types=mlc.DataType.BOUNDING_BOX,
+      description="Bounding box feature",
+      source=mlc.Source(format="InvalidFormat"),
+  )
+  with pytest.raises(ValueError, match="Unsupported bounding box format"):
+    croissant_builder.datatype_converter(field)
+
+
 @pytest.mark.parametrize(
     ["field", "feature_type", "subfield_types"],
     [
@@ -150,6 +173,14 @@ def test_simple_datatype_converter(
             image_feature.Image,
             None,
         ),
+        (
+            mlc.Field(
+                data_types=mlc.DataType.AUDIO_OBJECT,
+                description="Audio feature",
+            ),
+            audio_feature.Audio,
+            None,
+        ),
         (
             mlc.Field(
                 data_types=mlc.DataType.BOUNDING_BOX,

{tfds_nightly-4.9.9.dev202507160045.dist-info → tfds_nightly-4.9.9.dev202507180045.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: tfds-nightly
-Version: 4.9.9.dev202507160045
+Version: 4.9.9.dev202507180045
 Summary: tensorflow/datasets is a library of datasets ready to use with TensorFlow.
 Home-page: https://github.com/tensorflow/datasets
 Download-URL: https://github.com/tensorflow/datasets/tags

{tfds_nightly-4.9.9.dev202507160045.dist-info → tfds_nightly-4.9.9.dev202507180045.dist-info}/RECORD

@@ -141,8 +141,8 @@ tensorflow_datasets/core/data_sources/python_test.py,sha256=O3yqMPx40JlHN0uFfZPN
 tensorflow_datasets/core/dataset_builders/__init__.py,sha256=StTA3euephqDZdpTzJQgfWNqB5inZosrAhaWg2BOeio,1945
 tensorflow_datasets/core/dataset_builders/adhoc_builder.py,sha256=QVE8wWGPOgILPTC27Q28QZ3KIi5N64OGOfKpTq4W4_0,9216
 tensorflow_datasets/core/dataset_builders/adhoc_builder_test.py,sha256=yhRwrznK78MvHeWGRggnMTiyx_SlR1z30iD5VU3Gweo,13096
-tensorflow_datasets/core/dataset_builders/croissant_builder.py,sha256=Pwo0is1rFJmuW6kT_k6OCXmBFTCBlqcZI-0eLSx9GEg,15913
-tensorflow_datasets/core/dataset_builders/croissant_builder_test.py,sha256=o6DSxXkecYYyNXqWimbkoy24UwNkvgWbL-U1_col_tg,10538
+tensorflow_datasets/core/dataset_builders/croissant_builder.py,sha256=Ef9fSTxFIvUKf_zfCqL3JyUje6f9buIQXvT1iuHgB20,16596
+tensorflow_datasets/core/dataset_builders/croissant_builder_test.py,sha256=1KPrGPFYHQvo33TwnG5LmPpdlyRTudPfni4ipFsj0ao,11607
 tensorflow_datasets/core/dataset_builders/huggingface_dataset_builder.py,sha256=Loq3qeGk1Ias-d2oT_dK47BRNgTA4LKJchNGh7aA4a0,18313
 tensorflow_datasets/core/dataset_builders/huggingface_dataset_builder_test.py,sha256=6N3DLsry9LhDqhpleaoXrrhaGiLJMBgUlwDnAji-1fI,4389
 tensorflow_datasets/core/dataset_builders/view_builder.py,sha256=eaCtjN5Vg4rK8JD3auA4PhF9mjH5HvQ9dslDX8LbwyM,11907
@@ -2468,10 +2468,10 @@ tensorflow_datasets/vision_language/wit/wit_test.py,sha256=PXS8DMNW-MDrT2p5oy4Ic
 tensorflow_datasets/vision_language/wit_kaggle/__init__.py,sha256=vGwSGeM8WE4Q-l0-eEE1sBojmk6YT0l1OO60AWa4Q40,719
 tensorflow_datasets/vision_language/wit_kaggle/wit_kaggle.py,sha256=q-vX_FBzIwsFxL4sY9vuyQ3UQD2PLM4yhUR4U6l-qao,16903
 tensorflow_datasets/vision_language/wit_kaggle/wit_kaggle_test.py,sha256=ZymHT1NkmD-pUnh3BmM3_g30c5afsWYnmqDD9dVyDSA,1778
-tfds_nightly-4.9.9.dev202507160045.dist-info/licenses/AUTHORS,sha256=nvBG4WwfgjuOu1oZkuQKw9kg7X6rve679ObS-YDDmXg,309
-tfds_nightly-4.9.9.dev202507160045.dist-info/licenses/LICENSE,sha256=z8d0m5b2O9McPEK1xHG_dWgUBT6EfBDz6wA0F7xSPTA,11358
-tfds_nightly-4.9.9.dev202507160045.dist-info/METADATA,sha256=f1qRJdbrEjRnhS5iNf10pK3bG6V3KV21ZusYp9cEzCc,11694
-tfds_nightly-4.9.9.dev202507160045.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-tfds_nightly-4.9.9.dev202507160045.dist-info/entry_points.txt,sha256=eHEL7nF5y1uCY2FgkuYIdE062epJXlAQTSdq89px4p4,73
-tfds_nightly-4.9.9.dev202507160045.dist-info/top_level.txt,sha256=bAevmk9209s_oxVZVlN6hSDIVS423qrMQvmcWSvW4do,20
-tfds_nightly-4.9.9.dev202507160045.dist-info/RECORD,,
+tfds_nightly-4.9.9.dev202507180045.dist-info/licenses/AUTHORS,sha256=nvBG4WwfgjuOu1oZkuQKw9kg7X6rve679ObS-YDDmXg,309
+tfds_nightly-4.9.9.dev202507180045.dist-info/licenses/LICENSE,sha256=z8d0m5b2O9McPEK1xHG_dWgUBT6EfBDz6wA0F7xSPTA,11358
+tfds_nightly-4.9.9.dev202507180045.dist-info/METADATA,sha256=nJXUk8MApA1PG6w8CkH8by_d1XcPHzen2tW9qs_QzRA,11694
+tfds_nightly-4.9.9.dev202507180045.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+tfds_nightly-4.9.9.dev202507180045.dist-info/entry_points.txt,sha256=eHEL7nF5y1uCY2FgkuYIdE062epJXlAQTSdq89px4p4,73
+tfds_nightly-4.9.9.dev202507180045.dist-info/top_level.txt,sha256=bAevmk9209s_oxVZVlN6hSDIVS423qrMQvmcWSvW4do,20
+tfds_nightly-4.9.9.dev202507180045.dist-info/RECORD,,