Getting prompts using LiteLLM syntax

[doasfrancisco]

Describe the feature or potential improvement

It would be ideal to get prompts using LiteLLM syntax, example:

async def get_response(message):
    prompt= litellm.get_prompt("test").compile(user_message="Hi! How are you?")
    response = await acompletion(
        model="gpt-4o",
        messages=prompt,
        metadata= {
            "generation_name": "test_generation",
            "session_id": "test_session_id",
            "trace_user_id": "test_user",
            "prompt": "test"
            }
        )
    return response

Additional information

No response