feat(openai-responses): Support multiple message roles in API inputs

Also update the nesting to add multiple messages(where appropriate)
rather then a single message with multiple content parts.

Signed-off-by: Derek Higgins <derekh@redhat.com>

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

@@ -34,8 +34,10 @@ from llama_stack.apis.inference.inference import (
     OpenAIChatCompletionContentPartTextParam,
     OpenAIChatCompletionToolCallFunction,
     OpenAIChoice,
+    OpenAIDeveloperMessageParam,
     OpenAIImageURL,
     OpenAIMessageParam,
+    OpenAISystemMessageParam,
     OpenAIToolMessageParam,
     OpenAIUserMessageParam,
 )
@@ -77,6 +79,16 @@ async def _openai_choices_to_output_messages(choices: list[OpenAIChoice]) -> lis
     return output_messages
 
 
+async def _get_message_type_by_role(role: str):
+    role_to_type = {
+        "user": OpenAIUserMessageParam,
+        "system": OpenAISystemMessageParam,
+        "assistant": OpenAIAssistantMessageParam,
+        "developer": OpenAIDeveloperMessageParam,
+    }
+    return role_to_type.get(role)
+
+
 class OpenAIResponsesImpl:
     def __init__(
         self,
@@ -116,26 +128,32 @@ class OpenAIResponsesImpl:
         if previous_response_id:
             previous_response = await self.get_openai_response(previous_response_id)
             messages.extend(await _previous_response_to_messages(previous_response))
+
         # TODO: refactor this user_content parsing out into a separate method
-        user_content: str | list[OpenAIChatCompletionContentPartParam] = ""
+        content: str | list[OpenAIChatCompletionContentPartParam] = ""
         if isinstance(input, list):
-            user_content = []
+            for input_message in input:
-            for user_input in input:
+                if isinstance(input_message.content, list):
-                if isinstance(user_input.content, list):
+                    content = []
-                    for user_input_content in user_input.content:
+                    for input_message_content in input_message.content:
-                        if isinstance(user_input_content, OpenAIResponseInputMessageContentText):
+                        if isinstance(input_message_content, OpenAIResponseInputMessageContentText):
-                            user_content.append(OpenAIChatCompletionContentPartTextParam(text=user_input_content.text))
+                            content.append(OpenAIChatCompletionContentPartTextParam(text=input_message_content.text))
-                        elif isinstance(user_input_content, OpenAIResponseInputMessageContentImage):
+                        elif isinstance(input_message_content, OpenAIResponseInputMessageContentImage):
-                            if user_input_content.image_url:
+                            if input_message_content.image_url:
                                 image_url = OpenAIImageURL(
-                                    url=user_input_content.image_url, detail=user_input_content.detail
+                                    url=input_message_content.image_url, detail=input_message_content.detail
                                 )
-                                user_content.append(OpenAIChatCompletionContentPartImageParam(image_url=image_url))
+                                content.append(OpenAIChatCompletionContentPartImageParam(image_url=image_url))
                 else:
-                    user_content.append(OpenAIChatCompletionContentPartTextParam(text=user_input.content))
+                    content = input_message.content
+                message_type = await _get_message_type_by_role(input_message.role)
+                if message_type is None:
+                    raise ValueError(
+                        f"Llama Stack OpenAI Responses does not yet support message role '{input_message.role}' in this context"
+                    )
+                messages.append(message_type(content=content))
         else:
-            user_content = input
+            messages.append(OpenAIUserMessageParam(content=input))
-        messages.append(OpenAIUserMessageParam(content=user_content))
 
         chat_tools = await self._convert_response_tools_to_chat_tools(tools) if tools else None
         chat_response = await self.inference_api.openai_chat_completion(

tests/unit/providers/agents/meta_reference/test_openai_responses.py

@@ -9,10 +9,15 @@ from unittest.mock import AsyncMock
 import pytest
 
 from llama_stack.apis.agents.openai_responses import (
+    OpenAIResponseInputMessage,
+    OpenAIResponseInputMessageContentText,
     OpenAIResponseInputToolWebSearch,
     OpenAIResponseOutputMessage,
 )
 from llama_stack.apis.inference.inference import (
+    OpenAIAssistantMessageParam,
+    OpenAIChatCompletionContentPartTextParam,
+    OpenAIDeveloperMessageParam,
     OpenAIUserMessageParam,
 )
 from llama_stack.apis.tools.tools import Tool, ToolGroups, ToolInvocationResult, ToolParameter, ToolRuntime
@@ -156,3 +161,49 @@ async def test_create_openai_response_with_string_input_with_tools(openai_respon
     assert len(result.output) >= 1
     assert isinstance(result.output[1], OpenAIResponseOutputMessage)
     assert result.output[1].content[0].text == "Dublin"
+
+
+@pytest.mark.asyncio
+async def test_create_openai_response_with_multiple_messages(openai_responses_impl, mock_inference_api):
+    """Test creating an OpenAI response with multiple messages."""
+    # Setup
+    input_messages = [
+        OpenAIResponseInputMessage(role="developer", content="You are a helpful assistant", name=None),
+        OpenAIResponseInputMessage(role="user", content="Name some towns in Ireland", name=None),
+        OpenAIResponseInputMessage(
+            role="assistant",
+            content=[
+                OpenAIResponseInputMessageContentText(text="Galway, Longford, Sligo"),
+                OpenAIResponseInputMessageContentText(text="Dublin"),
+            ],
+            name=None,
+        ),
+        OpenAIResponseInputMessage(role="user", content="Which is the largest town in Ireland?", name=None),
+    ]
+    model = "meta-llama/Llama-3.1-8B-Instruct"
+
+    mock_inference_api.openai_chat_completion.return_value = load_chat_completion_fixture("simple_chat_completion.yaml")
+
+    # Execute
+    await openai_responses_impl.create_openai_response(
+        input=input_messages,
+        model=model,
+        temperature=0.1,
+    )
+
+    # Verify the the correct messages were sent to the inference API i.e.
+    # All of the responses message were convered to the chat completion message objects
+    inference_messages = mock_inference_api.openai_chat_completion.call_args_list[0].kwargs["messages"]
+    for i, m in enumerate(input_messages):
+        if isinstance(m.content, str):
+            assert inference_messages[i].content == m.content
+        else:
+            assert inference_messages[i].content[0].text == m.content[0].text
+            assert isinstance(inference_messages[i].content[0], OpenAIChatCompletionContentPartTextParam)
+        assert inference_messages[i].role == m.role
+        if m.role == "user":
+            assert isinstance(inference_messages[i], OpenAIUserMessageParam)
+        elif m.role == "assistant":
+            assert isinstance(inference_messages[i], OpenAIAssistantMessageParam)
+        else:
+            assert isinstance(inference_messages[i], OpenAIDeveloperMessageParam)