push

Author: hassiebp

langfuse/_client/client.py

@@ -2670,15 +2670,13 @@ async def _process_experiment_item(
                     try:
                         from langfuse.model import CreateDatasetRunItemRequest
 
-                        dataset_run_item = (
-                            await self.async_api.dataset_run_items.create(
-                                request=CreateDatasetRunItemRequest(
-                                    runName=experiment_name,
-                                    runDescription=experiment_description,
-                                    metadata=experiment_metadata,
-                                    datasetItemId=item.id,  # type: ignore
-                                    traceId=trace_id,
-                                )
+                        dataset_run_item = self.api.dataset_run_items.create(
+                            request=CreateDatasetRunItemRequest(
+                                runName=experiment_name,
+                                runDescription=experiment_description,
+                                metadata=experiment_metadata,
+                                datasetItemId=item.id,  # type: ignore
+                                traceId=trace_id,
                             )
                         )
 

langfuse/_client/experiments.py

@@ -37,8 +37,8 @@ class LocalExperimentItem(TypedDict, total=False):
     metadata: Optional[Dict[str, Any]]
 
 
-ExperimentItem = Union[LocalExperimentItem, DatasetItemClient]
-ExperimentData = Union[List[LocalExperimentItem], List[DatasetItemClient]]
+ExperimentItem = Union[LocalExperimentItem, "DatasetItemClient"]
+ExperimentData = Union[List[LocalExperimentItem], List["DatasetItemClient"]]
 
 
 class Evaluation(TypedDict, total=False):

tests/test_experiments.py

@@ -281,7 +281,7 @@ def test_large_dataset_with_concurrency():
     result = langfuse_client.run_experiment(
         name="Large dataset test",
         data=large_dataset,
-        task=lambda **kwargs: f"Processed {kwargs['input']}",
+        task=lambda **kwargs: f"Processed {kwargs['item']}",
         evaluators=[lambda **kwargs: {"name": "simple_eval", "value": 1.0}],
         max_concurrency=5,
     )
@@ -525,7 +525,7 @@ def test_format_experiment_results_basic():
         name="Formatting test",
         description="Test result formatting",
         data=[{"input": "Hello", "expected_output": "Hi"}],
-        task=lambda **kwargs: f"Processed: {kwargs['input']}",
+        task=lambda **kwargs: f"Processed: {kwargs['item']}",
         evaluators=[simple_evaluator],
         run_evaluators=[run_evaluator_average_length],
     )