Add check to make sure all reserved keywords are attributes and not properties

nemo_curator/stages/base.py

@@ -112,6 +112,15 @@ def __init_subclass__(cls, **kwargs):
             msg = f"{cls.__name__} must not override '_batch_size'"
             raise TypeError(msg)
 
+        for attr in ("name", "resources", "batch_size"):
+            if isinstance(cls.__dict__.get(attr), property):
+                msg = (
+                    f"{cls.__name__} must not define '{attr}' as a @property. "
+                    f"Use a plain class attribute or dataclass field instead, "
+                    f"so that ProcessingStage.with_() can override it."
+                )
+                raise TypeError(msg)
+
     def num_workers(self) -> int | None:
         """Number of workers required. If None, then executor will determine the number of workers."""
         return None

nemo_curator/stages/synthetic/nemotron_cc/base.py

@@ -41,10 +41,7 @@ class BaseSyntheticStage(ProcessingStage[DocumentBatch, DocumentBatch]):
     client: AsyncLLMClient | LLMClient = None
     model_name: str = None
     generation_config: GenerationConfig | None = None
-
-    @property
-    def name(self) -> str:
-        return "NemotronCCBaseStage"
+    name: str = "NemotronCCBaseStage"
 
     def __post_init__(self) -> None:
         self.is_async_client = isinstance(self.client, AsyncLLMClient)

nemo_curator/stages/synthetic/nemotron_cc/nemotron_cc.py

@@ -64,10 +64,7 @@ class DiverseQAPostProcessingStage(ProcessingStage[DocumentBatch, DocumentBatch]
     tokenizer: AutoTokenizer | None = None
     prefix: str = "Here are the questions and answers based on the provided text:"
     max_num_pairs: int = 10
-
-    @property
-    def name(self) -> str:
-        return "DiverseQAPostProcessing"
+    name: str = "DiverseQAPostProcessing"
 
     def process(self, batch: DocumentBatch) -> DocumentBatch:
         df = batch.to_pandas()
@@ -149,10 +146,7 @@ class KnowledgeListPostProcessingStage(ProcessingStage[DocumentBatch, DocumentBa
     """
 
     input_field: str = "knowledge_list"
-
-    @property
-    def name(self) -> str:
-        return "KnowledgeListPostProcessing"
+    name: str = "KnowledgeListPostProcessing"
 
     def process(self, batch: DocumentBatch) -> DocumentBatch:
         df = batch.to_pandas()

tests/stages/common/test_base.py

@@ -340,6 +340,46 @@ def inputs(self) -> tuple[list[str], list[str]]:
                 def outputs(self) -> tuple[list[str], list[str]]:
                     return [], []
 
+    def test_name_property_decorator(self):
+        """Test that ProcessingStage raises an error if a derived class defines 'name' as a @property."""
+        with pytest.raises(TypeError, match="must not define 'name' as a @property"):
+
+            class MockStagePropertyName(ProcessingStage[MockTask, MockTask]):
+                @property
+                def name(self) -> str:
+                    return "PropertyName"
+
+                def process(self, task: MockTask) -> MockTask:
+                    return task
+
+    def test_resources_property_decorator(self):
+        """Test that ProcessingStage raises an error if a derived class defines 'resources' as a @property."""
+        with pytest.raises(TypeError, match="must not define 'resources' as a @property"):
+
+            class MockStagePropertyResources(ProcessingStage[MockTask, MockTask]):
+                name = "MockStagePropertyResources"
+
+                @property
+                def resources(self) -> Resources:
+                    return Resources(cpus=1.0)
+
+                def process(self, task: MockTask) -> MockTask:
+                    return task
+
+    def test_batch_size_property_decorator(self):
+        """Test that ProcessingStage raises an error if a derived class defines 'batch_size' as a @property."""
+        with pytest.raises(TypeError, match="must not define 'batch_size' as a @property"):
+
+            class MockStagePropertyBatchSize(ProcessingStage[MockTask, MockTask]):
+                name = "MockStagePropertyBatchSize"
+
+                @property
+                def batch_size(self) -> int:
+                    return 1
+
+                def process(self, task: MockTask) -> MockTask:
+                    return task
+
     def test_nested_class_inheritance(self):
         """Test that nested class inheritance raises an error if a derived class overrides the _name, _resources, or _batch_size property."""
         with pytest.raises(TypeError, match="MockStageNestedOverriddenName must not override '_name'"):