Address PR comments and fix build failure

Author: Kevin Kim

labelbox/data/serialization/ndjson/base.py

@@ -2,7 +2,7 @@
 from uuid import uuid4
 from pydantic import BaseModel, root_validator, validator, Field
 
-from labelbox.utils import camel_case
+from labelbox.utils import camel_case, is_exactly_one_set
 from ...annotation_types.types import Cuid
 
 
@@ -12,7 +12,7 @@ class DataRow(BaseModel):
 
     @root_validator()
     def must_set_one(cls, values):
-        if bool(values.get('id')) == bool(values.get('global_key')):
+        if is_exactly_one_set(values.get('id'), values.get('global_key')):
             raise ValueError("Must set either id or global_key")
         return values
 

labelbox/schema/annotation_import.py

@@ -14,6 +14,7 @@
 from labelbox.orm import query
 from labelbox.orm.db_object import DbObject
 from labelbox.orm.model import Field, Relationship
+from labelbox.utils import is_exactly_one_set
 from labelbox.schema.confidence_presence_checker import LabelsConfidencePresenceChecker
 from labelbox.schema.enums import AnnotationImportState
 from labelbox.schema.serialization import serialize_labels
@@ -155,7 +156,7 @@ def _get_ndjson_from_objects(cls, objects: Union[List[Dict[str, Any]],
             )
 
         objects = serialize_labels(objects)
-        cls.validate_data_rows(objects)
+        cls._validate_data_rows(objects)
 
         data_str = ndjson.dumps(objects)
         if not data_str:
@@ -174,7 +175,7 @@ def refresh(self) -> None:
         self._set_field_values(res)
 
     @classmethod
-    def validate_data_rows(cls, objects: List[Dict[str, Any]]):
+    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.
@@ -187,7 +188,8 @@ def validate_data_rows(cls, objects: List[Dict[str, Any]]):
         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']:
+            elif not is_exactly_one_set(object['dataRow'].get('id'),
+                                        object['dataRow'].get('globalKey')):
                 errors.append(
                     f"Must provide only one of 'id' or 'globalKey' for 'dataRow' in {object}"
                 )

labelbox/utils.py

@@ -26,6 +26,10 @@ def snake_case(s):
     return _convert(s, "_", lambda i: False)
 
 
+def is_exactly_one_set(x, y):
+    return not (bool(x) == bool(y))
+
+
 class _CamelCaseMixin(BaseModel):
 
     class Config:

tests/data/annotation_types/test_metrics.py

@@ -18,6 +18,7 @@ def test_legacy_scalar_metric():
         'data': {
             'external_id': None,
             'uid': 'ckrmd9q8g000009mg6vej7hzg',
+            'global_key': None,
             'im_bytes': None,
             'file_path': None,
             'url': None,
@@ -68,6 +69,7 @@ def test_custom_scalar_metric(feature_name, subclass_name, aggregation, value):
         'data': {
             'external_id': None,
             'uid': 'ckrmd9q8g000009mg6vej7hzg',
+            'global_key': None,
             'im_bytes': None,
             'file_path': None,
             'url': None,
@@ -124,6 +126,7 @@ def test_custom_confusison_matrix_metric(feature_name, subclass_name,
         'data': {
             'external_id': None,
             'uid': 'ckrmd9q8g000009mg6vej7hzg',
+            'global_key': None,
             'im_bytes': None,
             'file_path': None,
             'url': None,

tests/data/assets/ndjson/classification_import_global_key.json

@@ -0,0 +1,34 @@
+[
+    {
+        "answer": {
+            "schemaId": "ckrb1sfl8099g0y91cxbd5ftb",
+            "confidence": 0.8
+        },
+        "schemaId": "c123",
+        "dataRow": {
+            "globalKey": "05e8ee85-072e-4eb2-b30a-501dee9b0d9d"
+        },
+        "uuid": "f6879f59-d2b5-49c2-aceb-d9e8dc478673"
+    },
+    {
+        "answer": [
+            {
+                "schemaId": "ckrb1sfl8099e0y919v260awv",
+                "confidence": 0.82
+            }
+        ],
+        "schemaId": "ckrb1sfkn099c0y910wbo0p1a",
+        "dataRow": {
+            "globalKey": "05e8ee85-072e-4eb2-b30a-501dee9b0d9d"
+        },
+        "uuid": "d009925d-91a3-4f67-abd9-753453f5a584"
+    },
+    {
+        "answer": "a value",
+        "schemaId": "ckrb1sfkn099c0y910wbo0p1a",
+        "dataRow": {
+            "globalKey": "05e8ee85-072e-4eb2-b30a-501dee9b0d9d"
+        },
+        "uuid": "d009925d-91a3-4f67-abd9-753453f5a584"
+    }
+]