async def read_item(item_id: int, q: Union[str, None] = None):
    return {"item_id": item_id, "q": q}
```

**Qeyd**:

Əgər bu mövzu haqqında məlumatınız yoxdursa <a href="https://fastapi.tiangolo.com/az/async/#in-a-hurry" target="_blank">`async` və `await` sənədindəki</a> _"Tələsirsən?"_ bölməsinə baxa bilərsiniz.

</details>

### Kodu işə salaq

Serveri aşağıdakı əmr ilə işə salaq:

<div class="termy">

        ":unified_api_testutil",
        "//tensorflow/c:c_api",
        "//tensorflow/c:c_test_util",
        "//tensorflow/c:tf_status_helper",
        "//tensorflow/c/experimental/gradients:array_grad",
        "//tensorflow/c/experimental/gradients:math_grad",
        "//tensorflow/c/experimental/gradients:not_differentiable",
        "//tensorflow/c/experimental/gradients/tape:tape_context",

        dataset, in one function call.
    *   **[`timeseries_dataset_from_array`](https://www.tensorflow.org/api_docs/python/tf/keras/preprocessing/timeseries_dataset_from_array)**
        is a `tf.data.Dataset`-based replacement of the legacy
        `TimeseriesGenerator`. It takes you from an array of timeseries data to
        a dataset of shifting windows with their targets.
*   Added

# Library of gradient functions.
package(
    # copybara:uncomment default_applicable_licenses = ["//tensorflow:license"],
    licenses = ["notice"],
)

cc_library(
    name = "array_grad",
    srcs = ["array_grad.cc"],
    hdrs = [
        "array_grad.h",
    ],
    visibility = [
        "//tensorflow:internal",
    ],
    deps = [
        "//tensorflow/c/eager:abstract_context",

    return List(elements.size / 2) { Header(elements[it * 2]!!, elements[it * 2 + 1]!!) }
  }

  @JvmStatic
  fun repeat(
    c: Char,
    count: Int,
  ): String {
    val array = CharArray(count)
    Arrays.fill(array, c)
    return String(array)
  }

  /**
   * Okio buffers are internally implemented as a linked list of arrays. Usually this implementation

    assertThat(recordedRequest.path)
      .isEqualTo("/lookup?ct&dns=AAABAAABAAAAAAAABmdvb2dsZQNjb20AAAEAAQ")
  }

  @Test
  fun failOnExcessiveResponse() {
    val array = CharArray(128 * 1024 + 2) { '0' }
    server.enqueue(dnsResponse(String(array)))
    try {
      dns.lookup("google.com")
      fail<Any>()
    } catch (ioe: IOException) {
      assertThat(ioe.message).isEqualTo("google.com")

        override fun checkClientTrusted(
          chain: Array<X509Certificate>,
          authType: String,
        ) = throw CertificateException()

        override fun checkServerTrusted(
          chain: Array<X509Certificate>,
          authType: String,
        ) = throw AssertionError()

        override fun getAcceptedIssuers(): Array<X509Certificate> = throw AssertionError()
      }

    As always, in FastAPI you can combine `def` and `async def` as needed.

    If you need a refresher about when to use which, check out the section _"In a hurry?"_ in the docs about [`async` and `await`](../async.md#in-a-hurry){.internal-link target=_blank}.

9. This *path operation function* is not returning dataclasses (although it could), but a list of dictionaries with internal data.

!!! info "informação"
    Note como `Offer` tem uma lista de `Item`s, que por sua vez possui opcionalmente uma lista `Image`s

## Corpos de listas puras

Se o valor de primeiro nível do corpo JSON que você espera for um `array` do JSON (uma` lista` do Python), você pode declarar o tipo no parâmetro da função, da mesma forma que nos modelos do Pydantic:


```Python
images: List[Image]
```

como em:

```Python hl_lines="15"

        get() {
            val sourcePsi = when {
                // array type for vararg parameters is not present in the code, so no need to handle it
                delegatedTypeRef?.source?.kind == KtFakeSourceElementKind.ArrayTypeFromVarargParameter -> null

                // but the array's underlying type is present with a fake source, and needs to be handled