We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
1 parent f989cd1 commit fae0debCopy full SHA for fae0deb
1 file changed
eval_protocol/mcp_servers/tau2/tests/test_tau2_e2e.py
@@ -752,10 +752,22 @@ def tau2_airline_eval(
752
753
reward = 1.0
754
755
+ # Convert incoming action dicts to typed Action objects for the evaluator
756
+ action_objs: Optional[List[Action]] = None
757
+ if actions is not None:
758
+ action_objs = []
759
+ for a in actions:
760
+ if isinstance(a, Action):
761
+ action_objs.append(a)
762
+ elif isinstance(a, dict):
763
+ action_objs.append(Action(**a))
764
+ else:
765
+ raise TypeError("actions must be a list of Action or dict items")
766
+
767
evaluation_criteria = EvaluationCriteria(
768
nl_assertions=nl_assertions,
769
communicate_info=communicate_info,
- actions=actions,
770
+ actions=action_objs,
771
env_assertions=None,
772
reward_basis=[
773
RewardType.NL_ASSERTION,
0 commit comments