Fix completion params getting overwritten

Author: xzrderek

eval_protocol/adapters/langfuse.py

@@ -379,22 +379,32 @@ def _create_input_metadata(self, trace: Any, observations: List[Any]) -> InputMe
         Returns:
             InputMetadata object
         """
-        # Extract completion parameters from observations
+        # Extract completion parameters from trace input first, then observations
         completion_params = {}
 
-        # Look for model parameters in observations
-        for obs in observations:
-            if hasattr(obs, "model") and obs.model:
-                completion_params["model"] = obs.model
-            if hasattr(obs, "model_parameters") and obs.model_parameters:
-                params = obs.model_parameters
-                if "temperature" in params:
-                    completion_params["temperature"] = params["temperature"]
-                if "max_tokens" in params:
-                    completion_params["max_tokens"] = params["max_tokens"]
-                if "top_p" in params:
-                    completion_params["top_p"] = params["top_p"]
-                break
+        # First check trace input for evaluation test completion_params
+        if hasattr(trace, "input") and trace.input:
+            if isinstance(trace.input, dict):
+                kwargs = trace.input.get("kwargs", {})
+                if "completion_params" in kwargs:
+                    trace_completion_params = kwargs["completion_params"]
+                    if trace_completion_params and isinstance(trace_completion_params, dict):
+                        completion_params.update(trace_completion_params)
+
+        # Fallback: Look for model parameters in observations if not found in trace input
+        if not completion_params:
+            for obs in observations:
+                if hasattr(obs, "model") and obs.model:
+                    completion_params["model"] = obs.model
+                if hasattr(obs, "model_parameters") and obs.model_parameters:
+                    params = obs.model_parameters
+                    if "temperature" in params:
+                        completion_params["temperature"] = params["temperature"]
+                    if "max_tokens" in params:
+                        completion_params["max_tokens"] = params["max_tokens"]
+                    if "top_p" in params:
+                        completion_params["top_p"] = params["top_p"]
+                    break
 
         # Create dataset info from trace metadata
         dataset_info = {

eval_protocol/pytest/evaluation_test.py

@@ -271,9 +271,11 @@ def _log_eval_error(status: Status, rows: list[EvaluationRow] | None, passed: bo
                         passed=None,
                     )
                     for row in data:
-                        row.input_metadata.completion_params = (
-                            completion_params if completion_params is not None else {}
-                        )
+                        # Only set completion_params if they don't already exist
+                        if not row.input_metadata.completion_params:
+                            row.input_metadata.completion_params = (
+                                completion_params if completion_params is not None else {}
+                            )
                         # Add mode to session_data
                         if row.input_metadata.session_data is None:
                             row.input_metadata.session_data = {}