I am trying to use llama indexes async

# blocking LLM get message handler
def getmsg(message) -> AGENT_CHAT_RESPONSE_TYPE:
    return chat_sys.chat(message)


# async (non-blocking) LLM message handler
async def getmsga(message) -> AGENT_CHAT_RESPONSE_TYPE:
    return await client.loop.run_in_executor(None,getmsg, message)

from llama_index.llms import OpenAILike

llm = OpenAILike(model="model", api_key="fake", api_base="http://127.0.0.1:8000/v1", ...)

def completion_to_prompt(completion: str) -> str:
  system_prompt = "..."
  return f"<|system|>\n{system_prompt}</s>\n<|user|>\n{completion}</s>\n<|assistant|>\n"

def messages_to_prompt(messages):
  prompt_str = ""
  for msg in messages:
    if msg.role == "system":
      prompt += f"<|system|>\n{msg.content}</s>\n"
    if msg.role == "user":
      prompt += f"<|user|>\n{msg.content}</s>\n"
    if msg.role == "assistant":
      prompt += f"<|assistant|>\n{msg.content}</s>\n"

  prompt_str += "<|assistant|>\n"
  return prompt_str


llm = OpenAILike(....., completion_to_prompt=completion_to_prompt, messages_to_prompt=messages_to_prompt)