Fix label serialization

Author: Val Brodsky

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

@@ -14,14 +14,14 @@ class FeatureSchema(BaseModel):
     name: Optional[str] = None
     feature_schema_id: Optional[Cuid] = None
 
-    @model_validator(mode='before')
-    @classmethod
-    def must_set_one(cls, values):
-        if values['feature_schema_id'] is None and values['name'] is None:
+    @model_validator(mode='after')
+    def must_set_one(self):
+        if self.feature_schema_id is None and self.name is None:
             raise ValueError(
                 "Must set either feature_schema_id or name for all feature schemas"
             )
-        return values
+
+        return self
 
     def dict(self, *args, **kwargs):
         res = super().dict(*args, **kwargs)

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

@@ -7,13 +7,14 @@
 
 
 class DataRow(_CamelCaseMixin):
-    id: str = None
-    global_key: str = None
+    id: Optional[str] = None
+    global_key: Optional[str] = None
 
     @model_validator(mode='after')
     def must_set_one(self):
-        if not is_exactly_one_set(self.id, self.global_key):
-            raise ValueError("Must set either id or global_key")
+        if self.id is None and self.global_key is None:
+            raise ValueError(
+                "Must set either id or global_key for all data rows")
 
         return self
 

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

@@ -116,8 +116,11 @@ def from_common(cls, checklist: Checklist, name: str,
 
     def dict(self, *args, **kwargs):
         res = super().dict(*args, **kwargs)
-        if 'answers' in res:
-            res['answer'] = res.pop('answers')
+        if kwargs.get('by_alias', False):
+            key = 'answers'
+        else:
+            key = 'answer'
+        res[key] = [a.dict(*args, **kwargs) for a in self.answer]
         return res
 
 
@@ -149,6 +152,11 @@ def from_common(cls, radio: Radio, name: str,
                    name=name,
                    schema_id=feature_schema_id)
 
+    def dict(self, *args, **kwargs):
+        res = super().dict(*args, **kwargs)
+        res['answer'] = self.answer.dict(*args, **kwargs)
+        return res
+
 
 # ====== End of subclasses
 

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

@@ -108,10 +108,10 @@ def serialize(
             label.annotations = uuid_safe_annotations
             for example in NDLabel.from_common([label]):
                 annotation_uuid = getattr(example, "uuid", None)
-
-                res = example.dict(
+                res = example.model_dump(
                     by_alias=True,
                     exclude={"uuid"} if annotation_uuid == "None" else None,
+                    exclude_none=True,
                 )
                 for k, v in list(res.items()):
                     if k in IGNORE_IF_NONE and v is None:

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

@@ -163,7 +163,6 @@ def _infer_media_type(
             data = VideoData
         elif DICOMObjectAnnotation in types:
             data = DicomData
-
         if data_row.id:
             return data(uid=data_row.id)
         else:

libs/labelbox/tests/data/serialization/ndjson/test_video.py

@@ -14,13 +14,26 @@
 from labelbox.data.serialization.ndjson.converter import NDJsonConverter
 
 
+def sorted_assert_list(a, b):
+    assert sorted(a, key=lambda d: sorted(d.items())) == sorted(
+        b, key=lambda d: sorted(d.items()))
+
+
 def test_video():
     with open('tests/data/assets/ndjson/video_import.json', 'r') as file:
         data = json.load(file)
 
     res = list(NDJsonConverter.deserialize(data))
     res = list(NDJsonConverter.serialize(res))
-    assert res == [data[2], data[0], data[1], data[3], data[4], data[5]]
+    # assert res == [data[2], data[0], data[1], data[3], data[4], data[5]]
+    assert (res[0]) == data[2]
+    assert (res[1]) == data[0]
+    answers = data[1].pop('answer')
+    data[1]['answers'] = answers
+    assert (res[2]) == data[1]
+    assert (res[3]) == data[3]
+    assert (res[4]) == data[4]
+    assert (res[5]) == data[5]
 
 
 def test_video_name_only():