Python: Properly configure structured outputs based on new options dict

python/packages/core/agent_framework/_workflows/_group_chat.py

@@ -397,7 +397,7 @@ async def _invoke_agent_helper(conversation: list[ChatMessage]) -> AgentOrchestr
             agent_response = await self._agent.run(
                 messages=conversation,
                 thread=self._thread,
-                response_format=AgentOrchestrationOutput,
+                options={"response_format": AgentOrchestrationOutput},
             )
             # Parse and validate the structured output
             agent_orchestration_output = AgentOrchestrationOutput.model_validate_json(agent_response.text)

python/packages/core/agent_framework/_workflows/_handoff.py

@@ -286,32 +286,36 @@ def _clone_chat_agent(self, agent: ChatAgent) -> ChatAgent:
         logit_bias = options.get("logit_bias")
         metadata = options.get("metadata")
 
+        # Disable parallel tool calls to prevent the agent from invoking multiple handoff tools at once.
+        cloned_options: dict[str, Any] = {
+            "allow_multiple_tool_calls": False,
+            "frequency_penalty": options.get("frequency_penalty"),
+            "instructions": options.get("instructions"),
+            "logit_bias": dict(logit_bias) if logit_bias else None,
+            "max_tokens": options.get("max_tokens"),
+            "metadata": dict(metadata) if metadata else None,
+            "model_id": options.get("model_id"),
+            "presence_penalty": options.get("presence_penalty"),
+            "response_format": options.get("response_format"),
+            "seed": options.get("seed"),
+            "stop": options.get("stop"),
+            "store": options.get("store"),
+            "temperature": options.get("temperature"),
+            "tool_choice": options.get("tool_choice"),
+            "tools": all_tools if all_tools else None,
+            "top_p": options.get("top_p"),
+            "user": options.get("user"),
+        }
+
         return ChatAgent(
             chat_client=agent.chat_client,
-            instructions=options.get("instructions"),
             id=agent.id,
             name=agent.name,
             description=agent.description,
             chat_message_store_factory=agent.chat_message_store_factory,
             context_providers=agent.context_provider,
             middleware=middleware,
-            # Disable parallel tool calls to prevent the agent from invoking multiple handoff tools at once.
-            allow_multiple_tool_calls=False,
-            frequency_penalty=options.get("frequency_penalty"),
-            logit_bias=dict(logit_bias) if logit_bias else None,
-            max_tokens=options.get("max_tokens"),
-            metadata=dict(metadata) if metadata else None,
-            model_id=options.get("model_id"),
-            presence_penalty=options.get("presence_penalty"),
-            response_format=options.get("response_format"),
-            seed=options.get("seed"),
-            stop=options.get("stop"),
-            store=options.get("store"),
-            temperature=options.get("temperature"),
-            tool_choice=options.get("tool_choice"),  # type: ignore[arg-type]
-            tools=all_tools if all_tools else None,
-            top_p=options.get("top_p"),
-            user=options.get("user"),
+            default_options=cloned_options,  # type: ignore[arg-type]
         )
 
     def _apply_auto_tools(self, agent: ChatAgent, targets: Sequence[HandoffConfiguration]) -> None:

python/samples/getting_started/agents/azure_ai/azure_ai_with_response_format.py

@@ -34,7 +34,7 @@ async def main() -> None:
             name="ProductMarketerAgent",
             instructions="Return launch briefs as structured JSON.",
             # Specify type to use as response
-            response_format=ReleaseBrief,
+            options={"response_format": ReleaseBrief},
         )
 
         query = "Draft a launch brief for the Contoso Note app."

python/samples/getting_started/agents/openai/openai_responses_client_with_structured_output.py

@@ -35,7 +35,7 @@ async def non_streaming_example() -> None:
     print(f"User: {query}")
 
     # Get structured response from the agent using response_format parameter
-    result = await agent.run(query, response_format=OutputStruct)
+    result = await agent.run(query, options={"response_format": OutputStruct})
 
     # Access the structured output directly from the response value
     if result.value:
@@ -63,7 +63,7 @@ async def streaming_example() -> None:
     # Get structured response from streaming agent using AgentResponse.from_agent_response_generator
     # This method collects all streaming updates and combines them into a single AgentResponse
     result = await AgentResponse.from_agent_response_generator(
-        agent.run_stream(query, response_format=OutputStruct),
+        agent.run_stream(query, options={"response_format": OutputStruct}),
         output_format_type=OutputStruct,
     )
 

python/samples/getting_started/azure_functions/06_multi_agent_orchestration_conditionals/function_app.py

@@ -99,7 +99,7 @@ def spam_detection_orchestration(context: DurableOrchestrationContext):
     spam_result_raw = yield spam_agent.run(
         messages=spam_prompt,
         thread=spam_thread,
-        response_format=SpamDetectionResult,
+        options={"response_format": SpamDetectionResult},
     )
 
     spam_result = cast(SpamDetectionResult, spam_result_raw.value)
@@ -120,7 +120,7 @@ def spam_detection_orchestration(context: DurableOrchestrationContext):
     email_result_raw = yield email_agent.run(
         messages=email_prompt,
         thread=email_thread,
-        response_format=EmailResponse,
+        options={"response_format": EmailResponse},
     )
 
     email_result = cast(EmailResponse, email_result_raw.value)

python/samples/getting_started/azure_functions/07_single_agent_orchestration_hitl/function_app.py

@@ -98,7 +98,7 @@ def content_generation_hitl_orchestration(context: DurableOrchestrationContext):
     initial_raw = yield writer.run(
         messages=f"Write a short article about '{payload.topic}'.",
         thread=writer_thread,
-        response_format=GeneratedContent,
+        options={"response_format": GeneratedContent},
     )
 
     content = initial_raw.value
@@ -135,17 +135,15 @@ def content_generation_hitl_orchestration(context: DurableOrchestrationContext):
                 )
                 return {"content": content.content}
 
-            context.set_custom_status(
-                "Content rejected by human reviewer. Incorporating feedback and regenerating..."
-            )
+            context.set_custom_status("Content rejected by human reviewer. Incorporating feedback and regenerating...")
             rewrite_prompt = (
                 "The content was rejected by a human reviewer. Please rewrite the article incorporating their feedback.\n\n"
                 f"Human Feedback: {approval_payload.feedback or 'No feedback provided.'}"
             )
             rewritten_raw = yield writer.run(
                 messages=rewrite_prompt,
                 thread=writer_thread,
-                response_format=GeneratedContent,
+                options={"response_format": GeneratedContent},
             )
 
             rewritten_value = rewritten_raw.value
@@ -157,9 +155,7 @@ def content_generation_hitl_orchestration(context: DurableOrchestrationContext):
             context.set_custom_status(
                 f"Human approval timed out after {payload.approval_timeout_hours} hour(s). Treating as rejection."
             )
-            raise TimeoutError(
-                f"Human approval timed out after {payload.approval_timeout_hours} hour(s)."
-            )
+            raise TimeoutError(f"Human approval timed out after {payload.approval_timeout_hours} hour(s).")
 
     raise RuntimeError(f"Content could not be approved after {payload.max_review_attempts} iteration(s).")
 

python/samples/getting_started/chat_client/azure_responses_client.py

@@ -41,13 +41,13 @@ async def main() -> None:
     print(f"User: {message}")
     if stream:
         response = await ChatResponse.from_chat_response_generator(
-            client.get_streaming_response(message, tools=get_weather, response_format=OutputStruct),
+            client.get_streaming_response(message, tools=get_weather, options={"response_format": OutputStruct}),
             output_format_type=OutputStruct,
         )
         print(f"Assistant: {response.value}")
 
     else:
-        response = await client.get_response(message, tools=get_weather, response_format=OutputStruct)
+        response = await client.get_response(message, tools=get_weather, options={"response_format": OutputStruct})
         print(f"Assistant: {response.value}")
 
 

python/samples/getting_started/context_providers/simple_context_provider.py

@@ -48,7 +48,7 @@ async def invoked(
                     messages=request_messages,  # type: ignore
                     instructions="Extract the user's name and age from the message if present. "
                     "If not present return nulls.",
-                    response_format=UserInfo,
+                    options={"response_format": UserInfo},
                 )
 
                 # Update user info with extracted data

python/samples/getting_started/devui/workflow_agents/workflow.py

@@ -86,7 +86,7 @@ def is_approved(message: Any) -> bool:
         "- feedback: concise, actionable feedback\n"
         "- clarity, completeness, accuracy, structure: individual scores (0-100)"
     ),
-    response_format=ReviewResult,
+    default_options={"response_format": ReviewResult},
 )
 
 # Create Editor agent - improves content based on feedback

python/samples/getting_started/workflows/agents/workflow_as_agent_reflection_pattern.py

@@ -100,7 +100,7 @@ class _Response(BaseModel):
         messages.append(ChatMessage(role=Role.USER, text="Please review the agent's responses."))
 
         print("Reviewer: Sending review request to LLM...")
-        response = await self._chat_client.get_response(messages=messages, response_format=_Response)
+        response = await self._chat_client.get_response(messages=messages, options={"response_format": _Response})
 
         parsed = _Response.model_validate_json(response.messages[-1].text)
 

python/samples/getting_started/workflows/control-flow/edge_condition.py

@@ -138,7 +138,7 @@ def create_spam_detector_agent() -> ChatAgent:
             "Include the original email content in email_content."
         ),
         name="spam_detection_agent",
-        response_format=DetectionResult,
+        default_options={"response_format": DetectionResult},
     )
 
 
@@ -152,7 +152,7 @@ def create_email_assistant_agent() -> ChatAgent:
             "Return JSON with a single field 'response' containing the drafted reply."
         ),
         name="email_assistant_agent",
-        response_format=EmailResponse,
+        default_options={"response_format": EmailResponse},
     )
 
 

python/samples/getting_started/workflows/control-flow/multi_selection_edge_group.py

@@ -190,7 +190,7 @@ def create_email_analysis_agent() -> ChatAgent:
             "and 'reason' (string)."
         ),
         name="email_analysis_agent",
-        response_format=AnalysisResultAgent,
+        default_options={"response_format": AnalysisResultAgent},
     )
 
 
@@ -199,7 +199,7 @@ def create_email_assistant_agent() -> ChatAgent:
     return AzureOpenAIChatClient(credential=AzureCliCredential()).create_agent(
         instructions=("You are an email assistant that helps users draft responses to emails with professionalism."),
         name="email_assistant_agent",
-        response_format=EmailResponse,
+        default_options={"response_format": EmailResponse},
     )
 
 
@@ -208,7 +208,7 @@ def create_email_summary_agent() -> ChatAgent:
     return AzureOpenAIChatClient(credential=AzureCliCredential()).create_agent(
         instructions=("You are an assistant that helps users summarize emails."),
         name="email_summary_agent",
-        response_format=EmailSummaryModel,
+        default_options={"response_format": EmailSummaryModel},
     )
 
 
@@ -243,7 +243,8 @@ def select_targets(analysis: AnalysisResult, target_ids: list[str]) -> list[str]
     )
 
     workflow = (
-        workflow_builder.set_start_executor("store_email")
+        workflow_builder
+        .set_start_executor("store_email")
         .add_edge("store_email", "email_analysis_agent")
         .add_edge("email_analysis_agent", "to_analysis_result")
         .add_multi_selection_edge_group(

python/samples/getting_started/workflows/control-flow/switch_case_edge_group.py

@@ -162,7 +162,7 @@ def create_spam_detection_agent() -> ChatAgent:
             "and 'reason' (string)."
         ),
         name="spam_detection_agent",
-        response_format=DetectionResultAgent,
+        default_options={"response_format": DetectionResultAgent},
     )
 
 
@@ -171,7 +171,7 @@ def create_email_assistant_agent() -> ChatAgent:
     return AzureOpenAIChatClient(credential=AzureCliCredential()).create_agent(
         instructions=("You are an email assistant that helps users draft responses to emails with professionalism."),
         name="email_assistant_agent",
-        response_format=EmailResponse,
+        default_options={"response_format": EmailResponse},
     )
 
 

python/samples/getting_started/workflows/declarative/customer_support/main.py

@@ -171,28 +171,28 @@ async def main() -> None:
     self_service_agent = chat_client.create_agent(
         name="SelfServiceAgent",
         instructions=SELF_SERVICE_INSTRUCTIONS,
-        response_format=SelfServiceResponse,
+        default_options={"response_format": SelfServiceResponse},
     )
 
     ticketing_agent = chat_client.create_agent(
         name="TicketingAgent",
         instructions=TICKETING_INSTRUCTIONS,
         tools=plugin.get_functions(),
-        response_format=TicketingResponse,
+        default_options={"response_format": TicketingResponse},
     )
 
     routing_agent = chat_client.create_agent(
         name="TicketRoutingAgent",
         instructions=TICKET_ROUTING_INSTRUCTIONS,
         tools=[plugin.get_ticket],
-        response_format=RoutingResponse,
+        default_options={"response_format": RoutingResponse},
     )
 
     windows_support_agent = chat_client.create_agent(
         name="WindowsSupportAgent",
         instructions=WINDOWS_SUPPORT_INSTRUCTIONS,
         tools=[plugin.get_ticket],
-        response_format=SupportResponse,
+        default_options={"response_format": SupportResponse},
     )
 
     resolution_agent = chat_client.create_agent(
@@ -205,7 +205,7 @@ async def main() -> None:
         name="TicketEscalationAgent",
         instructions=ESCALATION_INSTRUCTIONS,
         tools=[plugin.get_ticket, plugin.send_notification],
-        response_format=EscalationResponse,
+        default_options={"response_format": EscalationResponse},
     )
 
     # Agent registry for lookup

python/samples/getting_started/workflows/declarative/deep_research/main.py

@@ -139,7 +139,7 @@ async def main() -> None:
     manager_agent = chat_client.create_agent(
         name="ManagerAgent",
         instructions=MANAGER_INSTRUCTIONS,
-        response_format=ManagerResponse,
+        default_options={"response_format": ManagerResponse},
     )
 
     summary_agent = chat_client.create_agent(

python/samples/getting_started/workflows/human-in-the-loop/guessing_game_with_human_input.py

@@ -154,7 +154,7 @@ def create_guessing_agent() -> ChatAgent:
             "No explanations or additional text."
         ),
         # response_format enforces that the model produces JSON compatible with GuessOutput.
-        response_format=GuessOutput,
+        default_options={"response_format": GuessOutput},
     )
 
 

python/samples/getting_started/workflows/state-management/shared_states_with_agents.py

@@ -162,7 +162,7 @@ def create_spam_detection_agent() -> ChatAgent:
             "You are a spam detection assistant that identifies spam emails. "
             "Always return JSON with fields is_spam (bool) and reason (string)."
         ),
-        response_format=DetectionResultAgent,
+        default_options={"response_format": DetectionResultAgent},
         # response_format enforces structured JSON from each agent.
         name="spam_detection_agent",
     )
@@ -176,7 +176,7 @@ def create_email_assistant_agent() -> ChatAgent:
             "Return JSON with a single field 'response' containing the drafted reply."
         ),
         # response_format enforces structured JSON from each agent.
-        response_format=EmailResponse,
+        default_options={"response_format": EmailResponse},
         name="email_assistant_agent",
     )
 

python/samples/semantic-kernel-migration/openai_responses/03_responses_agent_structured_output.py

@@ -49,7 +49,7 @@ async def run_agent_framework() -> None:
     # AF forwards the same response_format payload at invocation time.
     reply = await chat_agent.run(
         "Draft a launch brief for the Contoso Note app.",
-        response_format=ReleaseBrief,
+        options={"response_format": ReleaseBrief},
     )
     print("[AF]", reply.text)