Python: Fix openai assistant function calling bug (#8817)

### Motivation and Context

We weren't properly handling the available FunctionResultContent in an
OpenAI assistant if multiple tool calls were included in the chat
history.

<!-- Thank you for your contribution to the semantic-kernel repo!
Please help reviewers and future users, providing the following
information:
  1. Why is this change required?
  2. What problem does it solve?
  3. What scenario does it contribute to?
  4. If it fixes an open issue, please link to the issue here.
-->

### Description

Fix this issue by making sure we use the `function call contents` list
as the source of truth (the model is expecting that we return the tool
call ids that were sent). We then use these function call content ids to
grab the corresponding function result content from the chat history and
this is sent back to the model.
- Fixes #8701 

<!-- Describe your changes, the overall approach, the underlying design.
These notes will help understanding how your code works. Thanks! -->

### Contribution Checklist

<!-- Before submitting this PR, please make sure: -->

- [X] The code builds clean without any errors or warnings
- [X] The PR follows the [SK Contribution
Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md)
and the [pre-submission formatting
script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts)
raises no violations
- [X] All unit tests pass, and I have added new tests where possible
- [X] I didn't break anyone 😄

python/samples/getting_started_with_agents/step2_plugins.py

@@ -70,15 +70,14 @@ async def main():
     # Create the instance of the Kernel
     kernel = Kernel()
 
-    # Add the OpenAIChatCompletion AI Service to the Kernel
     service_id = "agent"
     kernel.add_service(AzureChatCompletion(service_id=service_id))
 
     settings = kernel.get_prompt_execution_settings_from_service_id(service_id=service_id)
     # Configure the function choice behavior to auto invoke kernel functions
     settings.function_choice_behavior = FunctionChoiceBehavior.Auto()
 
-    kernel.add_plugin(plugin=MenuPlugin(), plugin_name="menu")
+    kernel.add_plugin(MenuPlugin(), plugin_name="menu")
 
     # Create the agent
     agent = ChatCompletionAgent(

python/samples/getting_started_with_agents/step7_assistant.py

@@ -20,7 +20,7 @@
 HOST_INSTRUCTIONS = "Answer questions about the menu."
 
 # Note: you may toggle this to switch between AzureOpenAI and OpenAI
-use_azure_openai = True
+use_azure_openai = False
 
 
 # Define a sample plugin for the sample

python/semantic_kernel/agents/open_ai/open_ai_assistant_base.py

@@ -735,7 +735,7 @@ async def _invoke_internal(
                     chat_history = ChatHistory()
                     _ = await self._invoke_function_calls(fccs=fccs, chat_history=chat_history)
 
-                    tool_outputs = self._format_tool_outputs(chat_history)
+                    tool_outputs = self._format_tool_outputs(fccs, chat_history)
                     await self.client.beta.threads.runs.submit_tool_outputs(
                         run_id=run.id,
                         thread_id=thread_id,
@@ -982,22 +982,27 @@ async def _invoke_function_calls(self, fccs: list[FunctionCallContent], chat_his
         ]
         return await asyncio.gather(*tasks)
 
-    def _format_tool_outputs(self, chat_history: ChatHistory) -> list[dict[str, str]]:
+    def _format_tool_outputs(self, fccs: list[FunctionCallContent], chat_history: ChatHistory) -> list[dict[str, str]]:
         """Format tool outputs from chat history for submission.
 
         Args:
+            fccs: The function call contents.
             chat_history: The chat history.
 
         Returns:
             The formatted tool outputs as a list of dictionaries.
         """
-        tool_outputs = []
-        for tool_call in chat_history.messages[0].items:
-            if isinstance(tool_call, FunctionResultContent):
-                tool_outputs.append({
-                    "tool_call_id": tool_call.id,
-                    "output": tool_call.result,
-                })
-        return tool_outputs
+        tool_call_lookup = {
+            tool_call.id: tool_call
+            for message in chat_history.messages
+            for tool_call in message.items
+            if isinstance(tool_call, FunctionResultContent)
+        }
+
+        return [
+            {"tool_call_id": fcc.id, "output": str(tool_call_lookup[fcc.id].result)}
+            for fcc in fccs
+            if fcc.id in tool_call_lookup
+        ]
 
     # endregion

python/tests/unit/agents/test_open_ai_assistant_base.py

@@ -965,8 +965,8 @@ def test_format_tool_outputs(azure_openai_assistant_agent, openai_unit_test_env)
     frc = FunctionResultContent.from_function_call_content_and_result(fcc, 123, {"test2": "test2"})
     chat_history.add_message(message=frc.to_chat_message_content())
 
-    tool_outputs = azure_openai_assistant_agent._format_tool_outputs(chat_history)
-    assert tool_outputs[0] == {"tool_call_id": "test", "output": 123}
+    tool_outputs = azure_openai_assistant_agent._format_tool_outputs([fcc], chat_history)
+    assert tool_outputs[0] == {"tool_call_id": "test", "output": "123"}
 
 
 @pytest.mark.asyncio