First commit for supporting globalKeys in annotation import SDK

Author: Kevin Kim

labelbox/__init__.py

@@ -5,7 +5,7 @@
 from labelbox.schema.project import Project
 from labelbox.schema.model import Model
 from labelbox.schema.bulk_import_request import BulkImportRequest
-from labelbox.schema.annotation_import import MALPredictionImport, MEAPredictionImport, LabelImport
+from labelbox.schema.annotation_import import MALPredictionImport, MEAPredictionImport, LabelImport, MEAToMALPredictionImport
 from labelbox.schema.dataset import Dataset
 from labelbox.schema.data_row import DataRow
 from labelbox.schema.label import Label

labelbox/data/annotation_types/data/base_data.py

@@ -11,5 +11,6 @@ class BaseData(BaseModel, ABC):
     """
     external_id: Optional[str] = None
     uid: Optional[str] = None
+    global_key: Optional[str] = None
     media_attributes: Optional[Dict[str, Any]] = None
     metadata: Optional[List[Dict[str, Any]]] = None

labelbox/schema/annotation_import.py

@@ -155,6 +155,8 @@ def _get_ndjson_from_objects(cls, objects: Union[List[Dict[str, Any]],
             )
 
         objects = serialize_labels(objects)
+        # cls.validate_data_rows(objects)
+
         data_str = ndjson.dumps(objects)
         if not data_str:
             raise ValueError(f"{object_name} cannot be empty")
@@ -171,6 +173,36 @@ def refresh(self) -> None:
                             as_json=True)
         self._set_field_values(res)
 
+    @classmethod
+    def validate_data_rows(cls, objects: List[Dict[str, Any]]):
+        """
+        Validates annotations by checking 'dataRow' is provided
+        and only one of 'id' or 'globalKey' is provided.
+
+        Shows up to `max_num_errors` errors if invalidated, to prevent
+        large number of error messages from being printed out 
+        """
+        errors = []
+        max_num_errors = 100
+        for object in objects:
+            if 'dataRow' not in object:
+                errors.append(f"'dataRow' is missing in {object}")
+            elif 'id' in object['dataRow'] and 'globalKey' in object['dataRow']:
+                errors.append(
+                    f"Must provide only one of 'id' or 'globalKey' for 'dataRow' in {object}"
+                )
+
+        if errors:
+            errors_length = len(errors)
+            formatted_errors = '\n'.join(errors[:max_num_errors])
+            if errors_length > max_num_errors:
+                logger.warning(
+                    f"Found more than {max_num_errors} errors. Showing first {max_num_errors} error messages..."
+                )
+            raise ValueError(
+                f"Error while validating annotations. Found {errors_length} annotations with errors. Errors:\n{formatted_errors}"
+            )
+
     @classmethod
     def from_name(cls,
                   client: "labelbox.Client",

tests/integration/annotation_import/test_label_import.py

@@ -38,6 +38,41 @@ def test_create_from_objects(client, configured_project, object_predictions,
         label_import.input_file_url, object_predictions)
 
 
+def test_data_row_validation_errors(client, configured_project,
+                                    object_predictions):
+    name = str(uuid.uuid4())
+    # Set up data for validation errors
+    # Invalid: Remove 'dataRow' part entirely
+    del object_predictions[0]['dataRow']
+
+    # Invalid: Set both id and globalKey
+    object_predictions[1]['dataRow'] = {
+        'id': 'some id',
+        'globalKey': 'some global key'
+    }
+
+    # Valid
+    object_predictions[2]['dataRow'] = {
+        'id': 'some id',
+    }
+
+    # Valid
+    object_predictions[3]['dataRow'] = {
+        'globalKey': 'some global key',
+    }
+
+    with pytest.raises(ValueError) as exc_info:
+        label_import = LabelImport.create_from_objects(
+            client=client,
+            project_id=configured_project.uid,
+            name=name,
+            labels=object_predictions)
+    exception_str = str(exc_info.value)
+    assert "Found 2 annotations with errors" in exception_str
+    assert "'dataRow' is missing in" in exception_str
+    assert "Must provide only one of 'id' or 'globalKey' for 'dataRow'" in exception_str
+
+
 def test_create_from_label_objects(client, configured_project,
                                    object_predictions,
                                    annotation_import_test_helpers):

tests/integration/annotation_import/test_mea_prediction_import.py

@@ -11,6 +11,38 @@
 """
 
 
+def test_data_row_validation_errors(model_run_with_all_project_labels,
+                                    object_predictions):
+    name = str(uuid.uuid4())
+    # Set up data for validation errors
+    # Invalid: Remove 'dataRow' part entirely
+    del object_predictions[0]['dataRow']
+
+    # Invalid: Set both id and globalKey
+    object_predictions[1]['dataRow'] = {
+        'id': 'some id',
+        'globalKey': 'some global key'
+    }
+
+    # Valid
+    object_predictions[2]['dataRow'] = {
+        'id': 'some id',
+    }
+
+    # Valid
+    object_predictions[3]['dataRow'] = {
+        'globalKey': 'some global key',
+    }
+
+    with pytest.raises(ValueError) as exc_info:
+        model_run_with_all_project_labels.add_predictions(
+            name=name, predictions=object_predictions)
+    exception_str = str(exc_info.value)
+    assert "Found 2 annotations with errors" in exception_str
+    assert "'dataRow' is missing in" in exception_str
+    assert "Must provide only one of 'id' or 'globalKey' for 'dataRow'" in exception_str
+
+
 def test_create_from_url(model_run_with_model_run_data_rows,
                          annotation_import_test_helpers):
     name = str(uuid.uuid4())