pull requests are approved and pass continuous integration checks, a TensorFlow
team member will apply `ready to pull` label to your change. This means we are
working on getting your pull request submitted to our internal repository. After
the change has been submitted internally, your pull request will be merged
automatically on GitHub.

If you want to contribute, start working through the TensorFlow codebase,
navigate to the

If not specified and the dependency has `yield`, it will have a `scope` of `"request"` by default.

### `scope` for sub-dependencies { #scope-for-sub-dependencies }

When you declare a dependency with a `scope="request"` (the default), any sub-dependency needs to also have a `scope` of `"request"`.

     *
     * @param model The model whose URLs should be normalized, may be {@code null}.
     * @param request The model building request that holds further settings, must not be {@code null}.
     */
    void normalize(Model model, ModelBuildingRequest request);

Si no se especifica y la dependencia tiene `yield`, tendrá un `scope` de `"request"` por defecto.

### `scope` para sub-dependencias { #scope-for-sub-dependencies }

Cuando declaras una dependencia con `scope="request"` (el valor por defecto), cualquier sub-dependencia también necesita tener un `scope` de `"request"`.

        override fun writeTo(sink: BufferedSink) {
          sink.writeUtf8("Hello request!")
          sink.close()
        }

        override fun isDuplex(): Boolean = true
      }
    val request =
      request()
        .post(asyncRequestBody)
        .build()
    val response = client.newCall(request).execute()
    Assumptions.assumeTrue(response.protocol == Protocol.HTTP_2)

from fastapi import FastAPI
from fastapi.middleware.wsgi import WSGIMiddleware
from flask import Flask, request
from markupsafe import escape

flask_app = Flask(__name__)


@flask_app.route("/")
def flask_main():
    name = request.args.get("name", "World")
    return f"Hello, {escape(name)} from Flask!"


app = FastAPI()


@app.get("/v2")
def read_main():
    return {"message": "Hello World"}

## 使用 `Jinja2Templates`

* 导入 `Jinja2Templates`
* 创建可复用的 `templates` 对象
* 在返回模板的*路径操作*中声明 `Request` 参数
* 使用 `templates` 渲染并返回 `TemplateResponse`， 传递模板的名称、request对象以及一个包含多个键值对（用于Jinja2模板）的"context"字典，

{* ../../docs_src/templates/tutorial001.py hl[4,11,15:16] *}

/// note | 笔记

在FastAPI 0.108.0，Starlette 0.29.0之前，`name`是第一个参数。
并且，在此之前，`request`对象是作为context的一部分以键值对的形式传递的。

///

/// tip | 提示

	for _, ep := range testServer.Disks {
		for _, disk := range ep.Endpoints {
			os.RemoveAll(disk.Path)
		}
	}
}

// Truncate request to simulate unexpected EOF for a request signed using streaming signature v4.
func truncateChunkByHalfSigv4(req *http.Request) (*http.Request, error) {
	bufReader := bufio.NewReader(req.Body)
	hexChunkSize, chunkSignature, err := readChunkLine(bufReader)
	if err != nil {
		return nil, err

Die Möglichkeit, in Ihren Tests asynchrone Funktionen zu verwenden, könnte beispielsweise nützlich sein, wenn Sie Ihre Datenbank asynchron abfragen. Stellen Sie sich vor, Sie möchten das Senden von <abbr title="Request – Anfrage: Daten, die der Client zum Server sendet">Requests</abbr> an Ihre FastAPI-Anwendung testen und dann überprüfen, ob Ihr Backend die richtigen Daten erfolgreich in die Datenbank geschrieben hat, während Sie eine asynchrone Datenbankbibliothek verwenden.

		query := url.Values{}
		for key, value := range testCase.queryParams {
			query.Set(key, value)
		}
		// Create a request to use.
		req, err := http.NewRequest(http.MethodGet, "http://host/a/b?"+query.Encode(), nil)
		if err != nil {
			t.Errorf("(%d) failed to create http.Request, got %v", i, err)
		}
		if testCase.expected != ErrNone {
			// Should be set since we are simulating a http server.