diff --git a/src/openjarvis/evals/cli.py b/src/openjarvis/evals/cli.py index 319188a9..25ed1b5f 100644 --- a/src/openjarvis/evals/cli.py +++ b/src/openjarvis/evals/cli.py @@ -121,7 +121,7 @@ BENCHMARKS = { }, "pinchbench": { "category": "agentic", - "description": "PinchBench real-world agent tasks (23 tasks, multi-turn with tool use)", + "description": "PinchBench real-world agent tasks", }, } @@ -532,7 +532,10 @@ def _run_agentic( # Wire judge for PinchBench LLM-judge and hybrid grading if config.benchmark == "pinchbench" and hasattr(dataset, "set_judge"): judge_engine = getattr(config, "judge_engine", "cloud") or "cloud" - judge_model = getattr(config, "judge_model", None) or "anthropic/claude-opus-4-5" + judge_model = ( + getattr(config, "judge_model", None) + or "anthropic/claude-opus-4-5" + ) judge_backend = _build_judge_backend(judge_model, engine_key=judge_engine) dataset.set_judge(judge_backend, judge_model) diff --git a/src/openjarvis/evals/core/agentic_runner.py b/src/openjarvis/evals/core/agentic_runner.py index 57717ba6..ec98da97 100644 --- a/src/openjarvis/evals/core/agentic_runner.py +++ b/src/openjarvis/evals/core/agentic_runner.py @@ -329,7 +329,8 @@ class AgenticRunner: if agent_bus is not None: for etype in (EventType.TOOL_CALL_START, EventType.TOOL_CALL_END): def _relay(event, _etype=etype): - event_recorder.record(_etype, **(event.data if hasattr(event, "data") else {})) + data = event.data if hasattr(event, "data") else {} + event_recorder.record(_etype, **data) agent_bus.subscribe(etype, _relay) _bus_unsubs.append((etype, _relay))