fix tests

Author: AkhileshNegi

backend/app/tests/crud/evaluations/test_fast_judge.py

@@ -952,7 +952,7 @@ def test_summary_brief_carries_the_scored_traces(
         assert BOT_INSTRUCTIONS in brief
         payload = json.loads(brief.split("## Per-question judge traces (JSON)\n", 1)[1])
 
-        assert [trace["trace_id"] for trace in payload] == ["item-1"]
+        assert [trace["question_id"] for trace in payload] == ["item-1"]
         assert payload[0]["question"] == "Q1"
         assert payload[0]["ground_truth_answer"] == "golden-1"
         assert payload[0]["llm_answer"] == "generated for Q1"

backend/app/tests/crud/evaluations/test_judge.py

@@ -293,15 +293,16 @@ def test_empty_question_drops_ground_truth(self) -> None:
 class TestBuildJudgeParams:
     """FR-9: system-config judging uses the fallback model + built-in ground-truth prompt.
 
-    The judge is a reasoning model (gpt-5-mini) that rejects a custom temperature, so
-    the request never carries one.
+    `build_judge_params` never sets a temperature, so the body never carries one.
+    The judge model has no `model_config` row, so the mapper reads it as
+    non-reasoning and drops `EVAL_JUDGE_REASONING_EFFORT` from the request.
     """
 
     def test_defaults_to_fallback_model_and_builtin_prompt(self, db: Session) -> None:
         base_params = build_judge_params(session=db)
 
         assert base_params["model"] == settings.EVAL_JUDGE_MODEL
-        assert base_params["model"] == "gpt-5-mini"
+        assert base_params["model"] == "gpt-5.6-luna"
         assert "temperature" not in base_params
         # Instructions are the per-row applicable-metric subset, so they are composed
         # in judge_row, never baked into the shared base params.