I have reviewed the following resources to understand how to use function calling in LiveKit Agents:

• Voice Agents function calling
• Github example

However, I still have some questions and need further clarification. Coming from the OpenAI API specification background for defining tools, I am looking for guidance on the following aspects of function calling in LiveKit Agents:

• Use of Enums: How can I properly use enums to limit the possible outputs for function arguments?
• Required Arguments: How do I define required arguments for a function call?
• Tool Call Control: Where should I define the behavior for strict, auto, and forced tool calls?
• Tool Launching: Where should I define when to launch a tool in the process?
• Multiple Tools: How can I define and work with multiple tools within a function call?

Any insights or examples from others who have experimented with these aspects would be greatly appreciated. The framework is really useful for focusing on business cases, and I want to make sure I'm using the function calling features effectively.

Thanks for your help!