Allow users to update the final answer prompt of MagenticOne orc. (#4476

)

* Allow users to update the final answer prompt of MagenticOne orchestrator.

python/packages/autogen-agentchat/src/autogen_agentchat/teams/_group_chat/_magentic_one/_magentic_one_group_chat.py

@@ -7,6 +7,7 @@
 from ....base import ChatAgent, TerminationCondition
 from .._base_group_chat import BaseGroupChat
 from ._magentic_one_orchestrator import MagenticOneOrchestrator
+from ._prompts import ORCHESTRATOR_FINAL_ANSWER_PROMPT
 
 trace_logger = logging.getLogger(TRACE_LOGGER_NAME)
 event_logger = logging.getLogger(EVENT_LOGGER_NAME)
@@ -25,6 +26,7 @@ class MagenticOneGroupChat(BaseGroupChat):
             Without a termination condition, the group chat will run based on the orchestrator logic or until the maximum number of turns is reached.
         max_turns (int, optional): The maximum number of turns in the group chat before stopping. Defaults to 20.
         max_stalls (int, optional): The maximum number of stalls allowed before re-planning. Defaults to 3.
+        final_answer_prompt (str, optional): The LLM prompt used to generate the final answer or response from the team's transcript. A default (sensible for GPT-4o class models) is provided.
 
     Raises:
         ValueError: In orchestration logic if progress ledger does not have required keys or if next speaker is not valid.
@@ -64,6 +66,7 @@ def __init__(
         termination_condition: TerminationCondition | None = None,
         max_turns: int | None = 20,
         max_stalls: int = 3,
+        final_answer_prompt: str = ORCHESTRATOR_FINAL_ANSWER_PROMPT,
     ):
         super().__init__(
             participants,
@@ -77,6 +80,7 @@ def __init__(
             raise ValueError("At least one participant is required for MagenticOneGroupChat.")
         self._model_client = model_client
         self._max_stalls = max_stalls
+        self._final_answer_prompt = final_answer_prompt
 
     def _create_group_chat_manager_factory(
         self,
@@ -95,5 +99,6 @@ def _create_group_chat_manager_factory(
             max_turns,
             self._model_client,
             self._max_stalls,
+            self._final_answer_prompt,
             termination_condition,
         )

python/packages/autogen-agentchat/src/autogen_agentchat/teams/_group_chat/_magentic_one/_magentic_one_orchestrator.py

@@ -48,6 +48,7 @@ def __init__(
         max_turns: int | None,
         model_client: ChatCompletionClient,
         max_stalls: int,
+        final_answer_prompt: str,
         termination_condition: TerminationCondition | None,
     ):
         super().__init__(
@@ -60,6 +61,7 @@ def __init__(
         )
         self._model_client = model_client
         self._max_stalls = max_stalls
+        self._final_answer_prompt = final_answer_prompt
         self._name = "MagenticOneOrchestrator"
         self._max_json_retries = 10
         self._task = ""
@@ -95,7 +97,10 @@ def _get_task_ledger_plan_update_prompt(self, team: str) -> str:
         return ORCHESTRATOR_TASK_LEDGER_PLAN_UPDATE_PROMPT.format(team=team)
 
     def _get_final_answer_prompt(self, task: str) -> str:
-        return ORCHESTRATOR_FINAL_ANSWER_PROMPT.format(task=task)
+        if self._final_answer_prompt == ORCHESTRATOR_FINAL_ANSWER_PROMPT:
+            return ORCHESTRATOR_FINAL_ANSWER_PROMPT.format(task=task)
+        else:
+            return self._final_answer_prompt
 
     async def _log_message(self, log_message: str) -> None:
         trace_logger.debug(log_message)