## Benutzerdefinierte Header hinzufügen

Es gibt Situationen, da ist es nützlich, dem HTTP-Error benutzerdefinierte Header hinzufügen zu können, etwa in einigen Sicherheitsszenarien.

Sie müssen das wahrscheinlich nicht direkt in ihrem Code verwenden.

Aber falls es in einem fortgeschrittenen Szenario notwendig ist, können Sie benutzerdefinierte Header wie folgt hinzufügen:

```Python hl_lines="14"

## `example` e `examples` no OpenAPI

Ao usar quaisquer dos:

* `Path()`
* `Query()`
* `Header()`
* `Cookie()`
* `Body()`
* `Form()`
* `File()`

você também pode declarar um dado `example` ou um grupo de `examples` com informações adicionais que serão adicionadas ao **OpenAPI**.

### `Body` com `example`

// SelfSubjectReview contains the user information that the kube-apiserver has about the user making this request.
// When using impersonation, users will receive the user info of the user being impersonated.  If impersonation or
// request header authentication is used, any extra keys will have their case ignored and returned as lowercase.
message SelfSubjectReview {
  // Standard object's metadata.

    # convert to modular filesystems everywhere
    visibility = ["//visibility:public"],
    deps = [
        ":filesystem_interface",
        "//tensorflow/c:tf_status_helper",
        "//tensorflow/c:tf_status_internal",
        "//tensorflow/core/platform:env",
        "//tensorflow/core/platform:errors",
        "//tensorflow/core/platform:status",
        "@local_tsl//tsl/platform:errors",

      } catch (ignored: IOException) {
        // Skip this path.
      }
    }
    return result.toTypedArray<String>()
  }

  /**
   * Reads stories named "story_xx.json" from the folder provided.
   */
  fun readStories(testFolderName: String): List<Story> {
    val result = mutableListOf<Story>()
    var i = 0
    while (true) { // break after last test.
      val storyResourceName =

   *
   * Otherwise if the response has a `Content-Type` header that specifies a charset, that is used
   * to determine the charset of the response bytes.
   *
   * Otherwise the response bytes are decoded as UTF-8.
   */
  fun charStream(): Reader =
    reader ?: BomAwareReader(source(), charset()).also {
      reader = it
    }

  /**
   * Returns the response as a string.
   *

    // Body contains nothing.
    assertThat(response.body.bytes().size).isEqualTo(0)
    assertThat(response.body.contentLength()).isEqualTo(0)

    // Content-Length header doesn't exist in a 204 response.
    assertThat(response.header("content-length")).isNull()
    assertThat(response.code).isEqualTo(204)
    val request = server.takeRequest()
    assertThat(request.requestLine).isEqualTo("GET /foo HTTP/1.1")
  }

        "//tensorflow/c:c_api_macros",
        "//tensorflow/c:tf_buffer",
        "//tensorflow/c:tf_buffer_internal",
        "//tensorflow/c:tf_status",
        "//tensorflow/c:tf_status_helper",
        "//tensorflow/core:framework",
        "//tensorflow/core:protos_all_cc",
        "//tensorflow/core/grappler:grappler_item",
        "//tensorflow/core/grappler/costs:graph_properties",

如果你定义了 `response_model`，程序会自动根据`response_model`来过滤和转换你响应的对象。

**FastAPI** 会使用这个 *临时* 响应对象去装在这些cookies信息 (同样还有headers和状态码等信息), 最终会将这些信息和通过`response_model`转化过的数据合并到最终的响应里。

你也可以在depend中定义`Response`参数，并设置cookie和header。

## 直接响应 `Response`

你还可以在直接响应`Response`时直接创建cookies。

你可以参考[Return a Response Directly](response-directly.md){.internal-link target=_blank}来创建response

然后设置Cookies，并返回：

```Python hl_lines="10-12"

The main limitation is that to update or delete content of a file inside a ZIP file the entire ZIP file must be replaced.

## How to enable S3 ZIP behavior ?

Ensure to set the following header `x-minio-extract` to `true` in your S3 requests.

## How to access to files inside a ZIP archive