chore: lint

Author: rishisurana-labelbox

libs/labelbox/src/labelbox/data/annotation_types/audio.py

@@ -1,17 +1,23 @@
 from typing import Optional
 
-from labelbox.data.annotation_types.annotation import ClassificationAnnotation, ObjectAnnotation
-from labelbox.data.mixins import ConfidenceNotSupportedMixin, CustomMetricsNotSupportedMixin
+from labelbox.data.annotation_types.annotation import (
+    ClassificationAnnotation,
+    ObjectAnnotation,
+)
+from labelbox.data.mixins import (
+    ConfidenceNotSupportedMixin,
+    CustomMetricsNotSupportedMixin,
+)
 
 
 class AudioClassificationAnnotation(ClassificationAnnotation):
     """Audio classification for specific time range
-    
+
     Examples:
     - Speaker identification from 2500ms to 4100ms
     - Audio quality assessment for a segment
     - Language detection for audio segments
-    
+
     Args:
         name (Optional[str]): Name of the classification
         feature_schema_id (Optional[Cuid]): Feature schema identifier
@@ -27,14 +33,18 @@ class AudioClassificationAnnotation(ClassificationAnnotation):
     segment_index: Optional[int] = None
 
 
-class AudioObjectAnnotation(ObjectAnnotation, ConfidenceNotSupportedMixin, CustomMetricsNotSupportedMixin):
+class AudioObjectAnnotation(
+    ObjectAnnotation,
+    ConfidenceNotSupportedMixin,
+    CustomMetricsNotSupportedMixin,
+):
     """Audio object annotation for specific time range
-    
+
     Examples:
     - Transcription: "Hello world" from 2500ms to 4100ms
     - Sound events: "Dog barking" from 10000ms to 12000ms
     - Audio segments with metadata
-    
+
     Args:
         name (Optional[str]): Name of the annotation
         feature_schema_id (Optional[Cuid]): Feature schema identifier

libs/labelbox/src/labelbox/data/annotation_types/label.py

@@ -90,12 +90,14 @@ def frame_annotations(
 
     def audio_annotations_by_frame(
         self,
-    ) -> Dict[int, List[Union[AudioObjectAnnotation, AudioClassificationAnnotation]]]:
+    ) -> Dict[
+        int, List[Union[AudioObjectAnnotation, AudioClassificationAnnotation]]
+    ]:
         """Get audio annotations organized by frame (millisecond)
-        
+
         Returns:
             Dict[int, List]: Dictionary mapping frame (milliseconds) to list of audio annotations
-            
+
         Example:
             >>> label.audio_annotations_by_frame()
             {2500: [AudioClassificationAnnotation(...)], 10000: [AudioObjectAnnotation(...)]}

libs/labelbox/src/labelbox/data/serialization/ndjson/classification.py

@@ -401,7 +401,11 @@ class NDClassification:
     @staticmethod
     def to_common(
         annotation: "NDClassificationType",
-    ) -> Union[ClassificationAnnotation, VideoClassificationAnnotation]:
+    ) -> Union[
+        ClassificationAnnotation,
+        VideoClassificationAnnotation,
+        AudioClassificationAnnotation,
+    ]:
         common = ClassificationAnnotation(
             value=annotation.to_common(),
             name=annotation.name,
@@ -416,18 +420,35 @@ def to_common(
         results = []
         for frame in annotation.frames:
             for idx in range(frame.start, frame.end + 1, 1):
-                results.append(
-                    VideoClassificationAnnotation(
-                        frame=idx, **common.model_dump(exclude_none=True)
+                # Check if this is an audio annotation by looking at the extra data
+                # Audio annotations will have frame/end_frame in extra, video annotations won't
+                if (
+                    hasattr(annotation, "extra")
+                    and annotation.extra
+                    and "frames" in annotation.extra
+                ):
+                    # This is likely an audio temporal annotation
+                    results.append(
+                        AudioClassificationAnnotation(
+                            frame=idx, **common.model_dump(exclude_none=True)
+                        )
+                    )
+                else:
+                    # This is a video temporal annotation
+                    results.append(
+                        VideoClassificationAnnotation(
+                            frame=idx, **common.model_dump(exclude_none=True)
+                        )
                     )
-                )
         return results
 
     @classmethod
     def from_common(
         cls,
         annotation: Union[
-            ClassificationAnnotation, VideoClassificationAnnotation, AudioClassificationAnnotation
+            ClassificationAnnotation,
+            VideoClassificationAnnotation,
+            AudioClassificationAnnotation,
         ],
         data: GenericDataRowData,
     ) -> Union[NDTextSubclass, NDChecklistSubclass, NDRadioSubclass]:
@@ -450,7 +471,9 @@ def from_common(
     @staticmethod
     def lookup_classification(
         annotation: Union[
-            ClassificationAnnotation, VideoClassificationAnnotation, AudioClassificationAnnotation
+            ClassificationAnnotation,
+            VideoClassificationAnnotation,
+            AudioClassificationAnnotation,
         ],
     ) -> Union[NDText, NDChecklist, NDRadio]:
         return {Text: NDText, Checklist: NDChecklist, Radio: NDRadio}.get(

libs/labelbox/src/labelbox/data/serialization/ndjson/label.py

@@ -171,17 +171,17 @@ def _create_audio_annotations(
         cls, label: Label
     ) -> Generator[Union[NDChecklistSubclass, NDRadioSubclass], None, None]:
         """Create audio annotations using generic temporal processor
-        
+
         Args:
             label: Label containing audio annotations to be processed
-            
+
         Yields:
             NDClassification or NDObject: Audio annotations in NDJSON format
         """
         # Use processor with configurable behavior
         processor = AudioTemporalProcessor(
             group_text_annotations=True,  # Group multiple TEXT annotations into one feature
-            enable_token_mapping=True     # Enable per-keyframe token content
+            enable_token_mapping=True,  # Enable per-keyframe token content
         )
         yield from processor.process_annotations(label)
 
@@ -215,7 +215,7 @@ def _create_non_video_annotations(cls, label: Label):
                 yield NDMessageTask.from_common(annotation, label.data)
             else:
                 raise TypeError(
-                    f"Unable to convert object to MAL format. `{type(getattr(annotation, 'value',annotation))}`"
+                    f"Unable to convert object to MAL format. `{type(getattr(annotation, 'value', annotation))}`"
                 )
 
     @classmethod

libs/labelbox/src/labelbox/data/serialization/ndjson/objects.py

@@ -773,29 +773,31 @@ def from_common(
         )
 
     @classmethod
-    def _serialize_audio_object_annotation(cls, annotation: AudioObjectAnnotation, data: GenericDataRowData):
+    def _serialize_audio_object_annotation(
+        cls, annotation: AudioObjectAnnotation, data: GenericDataRowData
+    ):
         """Serialize audio object annotation with temporal information
-        
+
         Args:
             annotation: Audio object annotation to process
             data: Data row data
-            
+
         Returns:
             NDObject: Serialized audio object annotation
         """
         # Get the appropriate NDObject subclass based on the annotation value type
         obj = cls.lookup_object(annotation)
-        
+
         # Process sub-classifications if any
         subclasses = [
             NDSubclassification.from_common(annot)
             for annot in annotation.classifications
         ]
-        
+
         # Add frame information to extra (milliseconds)
         extra = annotation.extra.copy() if annotation.extra else {}
         extra.update({"frame": annotation.frame})
-        
+
         # Create the NDObject with frame information
         return obj.from_common(
             str(annotation._uuid),

libs/labelbox/src/labelbox/data/serialization/ndjson/utils/__init__.py

@@ -1 +1 @@
-# Utils package for NDJSON serialization helpers
+# Utils package for NDJSON serialization helpers