[bot] Merge master/3ad5eca2 into rel/dev

packages/gooddata-eval/src/gooddata_eval/core/agentic/conversation.py

@@ -6,7 +6,7 @@
 import json
 import re
 from dataclasses import dataclass, field
-from typing import Literal
+from typing import ClassVar, Literal
 
 from gooddata_sdk import GoodDataSdk
 from pydantic import BaseModel
@@ -80,6 +80,23 @@ class TurnResult(BaseModel):
     def skill_success(self) -> bool:
         return self.skill_routing and self.output_present and self.no_error
 
+    # Reported per turn in detail["turns"]. A name listed here that no longer exists on the
+    # model raises rather than silently emitting a stale key, which a hand-written dict
+    # literal of the same fields would not -- and model_dump deep-copies activated_skills,
+    # so a caller mutating the returned dict cannot reach back into this TurnResult.
+    _DETAIL_FIELDS: ClassVar[set[str]] = {
+        "turn_id",
+        "expected_skill",
+        "skill_routing",
+        "output_present",
+        "output_correct",
+        "activated_skills",
+    }
+
+    def detail(self) -> dict:
+        """The subset of this result reported in detail["turns"] for one conversation turn."""
+        return self.model_dump(include=self._DETAIL_FIELDS)
+
 
 def _resolve_refs(
     expected_output: dict | None,
@@ -443,17 +460,7 @@ def _conversation_detail(result: ConversationResult) -> dict:
     return {
         "full_skill_coverage": result.full_skill_coverage,
         "total_clarification_turns": result.total_clarification_turns,
-        "turns": [
-            {
-                "turn_id": tr.turn_id,
-                "expected_skill": tr.expected_skill,
-                "skill_routing": tr.skill_routing,
-                "output_present": tr.output_present,
-                "output_correct": tr.output_correct,
-                "activated_skills": tr.activated_skills,
-            }
-            for tr in result.turn_results
-        ],
+        "turns": [tr.detail() for tr in result.turn_results],
         "latency_breakdown": build_latency_breakdown(result.tool_call_events, result.reasoning_step_events),
     }
 

packages/gooddata-eval/tests/test_agentic_conversation.py

@@ -98,6 +98,34 @@ def test_turn_result_skill_success():
     assert r.skill_success is True
 
 
+def _turn_result() -> TurnResult:
+    return TurnResult(
+        turn_id="t1",
+        expected_skill="visualization",
+        skill_routing=True,
+        output_present=True,
+        no_error=True,
+        activated_skills=["visualization"],
+        clarification_turns_used=0,
+        output_correct=None,
+    )
+
+
+def test_turn_result_detail_copies_activated_skills():
+    """A caller mutating the returned dict must not reach back into the TurnResult."""
+    r = _turn_result()
+    d = r.detail()
+    d["activated_skills"].append("mutated")
+    assert r.activated_skills == ["visualization"]
+
+
+def test_turn_result_detail_fields_all_exist_on_the_model():
+    """_DETAIL_FIELDS is a hand-listed subset, so a renamed field must fail here rather
+    than silently drop a key from every report."""
+    assert set(TurnResult.model_fields) >= TurnResult._DETAIL_FIELDS
+    assert set(_turn_result().detail()) == TurnResult._DETAIL_FIELDS
+
+
 def test_resolve_refs_no_refs():
     assert _resolve_refs({"key": "value"}, {}) == {"key": "value"}