from typing import Final

import pytest

from litellm.proxy.engine.inference import Deployment, DeploymentParams, completion_charge, quote
from litellm.types.utils import ModelResponse


def test_custom_priced_model_charges_reported_tokens() -> None:
    deployment: Final = Deployment(
        litellm_params=DeploymentParams(
            model="openai/engine-test", input_cost_per_token=0.001, output_cost_per_token=0.002
        )
    )
    response: Final = ModelResponse(
        model="engine-test", usage={"prompt_tokens": 20, "completion_tokens": 10, "total_tokens": 30}
    )
    assert completion_charge((deployment,), response, 10) == pytest.approx(0.04)
    assert quote((deployment,), "hello") > 0.04