tests/test_asr.py

import asyncio
from pathlib import Path

import pytest

from tutor.config import Settings
from tutor.services.asr.base import ASRClient, ASRError, FakeASRClient, Transcription
from tutor.services.asr.factory import create_asr_client


def test_fake_satisfies_protocol_and_records_calls() -> None:
    fake = FakeASRClient(text="hello world")
    assert isinstance(fake, ASRClient)
    result = asyncio.run(fake.transcribe(Path("/tmp/x.wav"), language="en"))
    assert isinstance(result, Transcription)
    assert result.text == "hello world"
    assert result.language == "en"
    assert fake.calls == [Path("/tmp/x.wav")]


def test_factory_returns_fake_by_default() -> None:
    client = create_asr_client(Settings(asr_provider="fake"))
    assert isinstance(client, FakeASRClient)


def test_factory_builds_faster_whisper_without_loading_model() -> None:
    # Construction must be cheap: the model loads lazily on first transcribe,
    # so building the client must not download or load anything.
    client = create_asr_client(
        Settings(asr_provider="faster_whisper", asr_model="tiny", asr_cpu_threads=1)
    )
    from tutor.services.asr.faster_whisper_client import FasterWhisperASRClient

    assert isinstance(client, FasterWhisperASRClient)
    assert client._model is None  # not loaded yet
    assert client.model_size == "tiny"
    assert client.cpu_threads == 1


def test_factory_rejects_unknown_provider() -> None:
    settings = Settings(asr_provider="fake")
    object.__setattr__(settings, "asr_provider", "bogus")  # bypass Literal validation
    with pytest.raises(ValueError, match="unknown ASR provider"):
        create_asr_client(settings)


def test_faster_whisper_raises_on_missing_file() -> None:
    from tutor.services.asr.faster_whisper_client import FasterWhisperASRClient

    client = FasterWhisperASRClient(model_size="tiny")
    with pytest.raises(ASRError, match="audio file not found"):
        asyncio.run(client.transcribe(Path("/nonexistent/audio.wav")))