# Async Tests { #async-tests }

You have already seen how to test your **FastAPI** applications using the provided `TestClient`. Up to now, you have only seen how to write synchronous tests, without using `async` functions.

/// tip | 팁

테스트 함수가 이제 `TestClient`를 사용할 때처럼 단순히 `def`가 아니라 `async def`로 작성된 점에 주목해주세요.

///

그 다음에  `AsyncClient` 로 앱을 만들고 비동기 요청을 `await` 키워드로 보낼 수 있습니다:

{* ../../docs_src/async_tests/test_main.py hl[9:12] *}

위의 코드는:

```Python
response = client.get('/')
```

`TestClient` 에 요청을 보내던 것과 동일합니다.

/// tip | 팁

새로운 `AsyncClient`를 사용할 때 async/await를 사용하고 있다는 점에 주목하세요. 이 요청은 비동기적으로 처리됩니다.

///

# Asynchrone Tests { #async-tests }

Sie haben bereits gesehen, wie Sie Ihre **FastAPI**-Anwendungen mit dem bereitgestellten `TestClient` testen. Bisher haben Sie nur gesehen, wie man synchrone Tests schreibt, ohne `async`-Funktionen zu verwenden.

# Tests Asíncronos { #async-tests }

Ya has visto cómo probar tus aplicaciones de **FastAPI** usando el `TestClient` proporcionado. Hasta ahora, solo has visto cómo escribir tests sincrónicos, sin usar funciones `async`.

# Асинхронное тестирование { #async-tests }

Вы уже видели как тестировать **FastAPI** приложение, используя имеющийся класс `TestClient`. К этому моменту вы видели только как писать тесты в синхронном стиле без использования `async` функций.

您已经了解了如何使用 `TestClient` 测试 **FastAPI** 应用程序。但是到目前为止，您只了解了如何编写同步测试，而没有使用 `async` 异步函数。

在测试中能够使用异步函数可能会很有用，比如当您需要异步查询数据库的时候。想象一下，您想要测试向 FastAPI 应用程序发送请求，然后验证您的后端是否成功在数据库中写入了正确的数据，与此同时您使用了异步的数据库的库。

让我们看看如何才能实现这一点。

## pytest.mark.anyio

如果我们想在测试中调用异步函数，那么我们的测试函数必须是异步的。 AnyIO 为此提供了一个简洁的插件，它允许我们指定一些测试函数要异步调用。

## HTTPX

即使您的 **FastAPI** 应用程序使用普通的 `def` 函数而不是 `async def` ，它本质上仍是一个 `async` 异步应用程序。

@app.get("/async/dict-no-response-model")
async def async_dict_no_response_model():
    return {"name": "foo", "value": 123}


@app.get("/async/dict-with-response-model", response_model=ItemOut)
async def async_dict_with_response_model(
    dep: Annotated[int, Depends(dep_b)],
):
    return {"name": "foo", "value": 123, "dep": dep}


@app.get("/async/model-no-response-model")
async def async_model_no_response_model(

class OtherDependencyError(Exception):
    pass


async def asyncgen_state(state: dict[str, str] = Depends(get_state)):
    state["/async"] = "asyncgen started"
    yield state["/async"]
    state["/async"] = "asyncgen completed"


def generator_state(state: dict[str, str] = Depends(get_state)):
    state["/sync"] = "generator started"
    yield state["/sync"]
    state["/sync"] = "generator completed"

# "/path/bar". This script uses "gsutil rsync" instead, which acts on directory
# contents. About arguments to gsutil:
# "gsutil -m rsync" runs in parallel.
# "gsutil rsync -r" is recursive and makes directories work.
# "gsutil rsync -d" is "sync and delete files from destination if not present in source"

DOWNLOADS="$(mktemp -d)"
mkdir -p "$DOWNLOADS"
gsutil -m rsync -r "$TFCI_ARTIFACT_STAGING_GCS_URI" "$DOWNLOADS"

	eventsSkipped     atomic.Int64
	eventsErrorsTotal atomic.Int64

	sync.RWMutex
	targets map[TargetID]Target
	queue   chan asyncEvent
	ctx     context.Context

	statLock    sync.RWMutex
	targetStats map[TargetID]targetStat
}

type targetStat struct {
	// The number of concurrent async Send calls per targets
	currentSendCalls int64
	// The number of total events per target
	totalEvents int64