feat!: Wire through parallel_tool_calls to Responses API (#4124)

# What does this PR do?
Initial PR against #4123
Adds `parallel_tool_calls` spec to Responses API and basic initial
implementation where no more than one function call is generated when
set to `False`.

## Test Plan
* Unit tests have been added to verify no more than one function call is
generated.
* A followup PR will verify passing through `parallel_tool_calls` to
providers.
* A followup PR will address verification and/or implementation of
incremental function calling across multiple conversational turns.

---------

Signed-off-by: Anastas Stoyanovsky <astoyano@redhat.com>

src/llama_stack/providers/inline/agents/meta_reference/agents.py

@@ -92,6 +92,7 @@ class MetaReferenceAgentsImpl(Agents):
         model: str,
         prompt: OpenAIResponsePrompt | None = None,
         instructions: str | None = None,
+        parallel_tool_calls: bool | None = True,
         previous_response_id: str | None = None,
         conversation: str | None = None,
         store: bool | None = True,
@@ -120,6 +121,7 @@ class MetaReferenceAgentsImpl(Agents):
             include,
             max_infer_iters,
             guardrails,
+            parallel_tool_calls,
             max_tool_calls,
         )
         return result  # type: ignore[no-any-return]

src/llama_stack/providers/inline/agents/meta_reference/responses/openai_responses.py

@@ -252,6 +252,7 @@ class OpenAIResponsesImpl:
         include: list[str] | None = None,
         max_infer_iters: int | None = 10,
         guardrails: list[str | ResponseGuardrailSpec] | None = None,
+        parallel_tool_calls: bool | None = None,
         max_tool_calls: int | None = None,
     ):
         stream = bool(stream)
@@ -296,6 +297,7 @@ class OpenAIResponsesImpl:
             tools=tools,
             max_infer_iters=max_infer_iters,
             guardrail_ids=guardrail_ids,
+            parallel_tool_calls=parallel_tool_calls,
             max_tool_calls=max_tool_calls,
         )
 
@@ -346,6 +348,7 @@ class OpenAIResponsesImpl:
         tools: list[OpenAIResponseInputTool] | None = None,
         max_infer_iters: int | None = 10,
         guardrail_ids: list[str] | None = None,
+        parallel_tool_calls: bool | None = True,
         max_tool_calls: int | None = None,
     ) -> AsyncIterator[OpenAIResponseObjectStream]:
         # These should never be None when called from create_openai_response (which sets defaults)
@@ -385,6 +388,7 @@ class OpenAIResponsesImpl:
             created_at=created_at,
             text=text,
             max_infer_iters=max_infer_iters,
+            parallel_tool_calls=parallel_tool_calls,
             tool_executor=self.tool_executor,
             safety_api=self.safety_api,
             guardrail_ids=guardrail_ids,

src/llama_stack/providers/inline/agents/meta_reference/responses/streaming.py

@@ -114,6 +114,7 @@ class StreamingResponseOrchestrator:
         safety_api,
         guardrail_ids: list[str] | None = None,
         prompt: OpenAIResponsePrompt | None = None,
+        parallel_tool_calls: bool | None = None,
         max_tool_calls: int | None = None,
     ):
         self.inference_api = inference_api
@@ -128,6 +129,8 @@ class StreamingResponseOrchestrator:
         self.prompt = prompt
         # System message that is inserted into the model's context
         self.instructions = instructions
+        # Whether to allow more than one function tool call generated per turn.
+        self.parallel_tool_calls = parallel_tool_calls
         # Max number of total calls to built-in tools that can be processed in a response
         self.max_tool_calls = max_tool_calls
         self.sequence_number = 0
@@ -190,6 +193,7 @@ class StreamingResponseOrchestrator:
             usage=self.accumulated_usage,
             instructions=self.instructions,
             prompt=self.prompt,
+            parallel_tool_calls=self.parallel_tool_calls,
             max_tool_calls=self.max_tool_calls,
         )