Can I run a llama_index function calling

function calling hasn't quite made it into the groq implementation yet, been meaning to add that

I managed to have it with:
llm = Groq(model="llama3-8b-8192", is_function_calling_model=True)
openai_step_engine = OpenAIAgentWorker.from_tools(query_engine_tools, llm=llm, verbose=True)
agent = AgentRunner(openai_step_engine)

but yeah - that looks like a hack