Fix task completion result key (#44460)

* Fix task completion result key

* address comments

---------

Co-authored-by: salma-elshafey <[email protected]>
Co-authored-by: ashaabansoliman <[email protected]>

Author: 3 people

sdk/evaluation/azure-ai-evaluation/azure/ai/evaluation/_evaluators/_task_completion/_task_completion.py

@@ -153,10 +153,11 @@ async def _do_eval(self, eval_input: Dict) -> Dict[str, Union[float, str]]:  # t
         llm_output = prompty_output_dict.get("llm_output", {})
 
         if isinstance(llm_output, dict):
-            success = llm_output.get("success", 0)
-            if isinstance(success, str):
-                success = 1 if success.upper() == "TRUE" else 0
-
+            success_value = llm_output.get("success", False)
+            if isinstance(success_value, str):
+                success = 1 if success_value.lower() == "true" else 0
+            else:
+                success = 1 if success_value else 0
             success_result = "pass" if success == 1 else "fail"
             reason = llm_output.get("explanation", "")
             return {