Fix benchmark scoring call signature

autotemp.py

@@ -383,7 +383,12 @@ def benchmark(self, dataset: List[Dict[str, str]], temperature_string: str, top_
                     run_text = self.run(prompt, temperature_string, top_p, advanced=advanced, rounds=rounds)
                     best_output = self._extract_best_output_from_run(run_text)
                     # We do not have direct overall score; compute via judges again for consistency
-                    score_detail = self.evaluate_output(best_output, temperature=float(self.default_temp or 0.7), top_p=float(top_p))
+                    score_detail = self.evaluate_output(
+                        prompt,
+                        best_output,
+                        temperature=float(self.default_temp or 0.7),
+                        top_p=float(top_p),
+                    )
                     per_item_scores.append(float(score_detail.get("overall", 0.0)))
                     if reference:
                         met = self._compute_external_metrics(best_output, reference)

tests/__init__.py

@@ -0,0 +1 @@
+

tests/test_benchmark.py

@@ -0,0 +1,98 @@
+import importlib.util
+import pathlib
+import sys
+import types
+import unittest
+
+
+def load_autotemp_module():
+    openai_stub = types.SimpleNamespace(
+        chat=types.SimpleNamespace(
+            completions=types.SimpleNamespace(create=lambda **kwargs: None)
+        ),
+        api_key=None,
+    )
+    dotenv_stub = types.SimpleNamespace(load_dotenv=lambda: None)
+
+    class DummyInterface:
+        def __init__(self, *args, **kwargs):
+            pass
+
+        def launch(self):
+            return None
+
+    class DummySlider:
+        def __init__(self, *args, **kwargs):
+            pass
+
+    class DummyCheckbox:
+        def __init__(self, *args, **kwargs):
+            pass
+
+        def style(self, **kwargs):
+            return self
+
+    gradio_stub = types.SimpleNamespace(
+        Interface=DummyInterface,
+        Slider=DummySlider,
+        Checkbox=DummyCheckbox,
+    )
+
+    sys.modules["openai"] = openai_stub
+    sys.modules["dotenv"] = dotenv_stub
+    sys.modules["gradio"] = gradio_stub
+
+    module_path = pathlib.Path(__file__).resolve().parents[1] / "autotemp.py"
+    spec = importlib.util.spec_from_file_location("autotemp_module", module_path)
+    module = importlib.util.module_from_spec(spec)
+    assert spec.loader is not None
+    spec.loader.exec_module(module)
+    return module
+
+
+AUTOTEMP_MODULE = load_autotemp_module()
+AutoTemp = AUTOTEMP_MODULE.AutoTemp
+
+
+class FakeAutoTemp(AutoTemp):
+    def __init__(self):
+        self.default_temp = 0.0
+        self.model_version = "dummy-model"
+        self.judges = 1
+        self.usage_totals = {"prompt_tokens": 0, "completion_tokens": 0, "total_tokens": 0}
+        self.last_evaluate_call = None
+
+    def run(self, prompt, temperature_string, top_p, advanced=False, rounds=1, exploration_c=1.0):
+        return "Best AutoTemp Output\nsynthetic benchmark answer\n\nJudges: {}"
+
+    def evaluate_output(self, prompt_text, output, temperature, top_p):
+        self.last_evaluate_call = (prompt_text, output, temperature, top_p)
+        return {"overall": 77.0}
+
+    def estimate_cost_usd(self):
+        return 0.0
+
+
+class BenchmarkRegressionTests(unittest.TestCase):
+    def test_benchmark_scores_best_output_with_prompt_context(self):
+        agent = FakeAutoTemp()
+
+        summary = agent.benchmark(
+            dataset=[{"prompt": "Explain entropy simply", "reference": ""}],
+            temperature_string="0.4,0.8",
+            top_p=0.9,
+            models=["dummy-model"],
+            advanced=False,
+            rounds=1,
+            judges=1,
+        )
+
+        self.assertEqual(summary["dummy-model"]["mean_overall"], 77.0)
+        self.assertEqual(
+            agent.last_evaluate_call,
+            ("Explain entropy simply", "synthetic benchmark answer", 0.7, 0.9),
+        )
+
+
+if __name__ == "__main__":
+    unittest.main()