object : WebSocketListener() {
            override fun onOpen(
              webSocket: WebSocket,
              response: Response,
            ) {
              try {
                latch.await() // The server can't respond to pings!
              } catch (e: InterruptedException) {
                throw AssertionError(e)
              }
            }
          },
        ).build(),
    )

    return {"item_id": item_id, "q": q}
```

<details markdown="1">
<summary>Або використайте <code>async def</code>...</summary>

Якщо ваш код використовує `async` / `await`, скористайтеся `async def`:

```Python hl_lines="7  12"
from fastapi import FastAPI

app = FastAPI()


@app.get("/")
async def read_root():
    return {"Hello": "World"}

                logger.error("Failed to update suggest index.", t);
                exitCode.set(1);
                latch.countDown();
            });

            try {
                latch.await();
            } catch (final InterruptedException ignore) {
                if (logger.isDebugEnabled()) {
                    logger.debug("Interrupted.", ignore);
                }
                exitCode.set(1);

{* ../../docs_src/response_model/tutorial001_01_py310.py hl[16,21] *}

Comme cela se passe côté Rust, la performance sera bien meilleure que si cela était fait avec le Python classique et la classe `JSONResponse`.

data: {"name": "Plumbus", "price": 32.99}

```

SSE is commonly used for AI chat streaming, live notifications, logs and observability, and other cases where the server pushes updates to the client.

/// tip

If you want to stream binary data, for example video or audio, check the advanced guide: [Stream Data](../advanced/stream-data.md).

///

## Stream SSE with FastAPI { #stream-sse-with-fastapi }

name: Contributor CI Build

on:
  push:
    branches:
      - master
  pull_request:
    branches:
      - master

concurrency:
  # On master/release, we don't want any jobs cancelled so the sha is used to name the group
  group: ${{ (github.ref == 'refs/heads/master' || github.ref == 'refs/heads/release' ) && format('contributor-pr-base-{0}', github.sha) || format('contributor-pr-{0}', github.ref) }}
  cancel-in-progress: false

env:

        parallelization: (Int) -> ParallelizationMethod,
    ): List<SmallSubprojectBucket> {
        // splitIntoBuckets() method expects us to split large element into N elements,
        // but we want to have a single bucket with N batches.
        // As a workaround, we repeat the bucket N times, and deduplicate the result at the end
        val resultIncludingDuplicates =
            splitIntoBuckets(

   *     Java 8 and removed in Java 9. We do not plan to remove this API from Guava, but if you are
   *     using it, it is probably not doing what you want.
   */
  @Deprecated
  JAVA_EXT_DIRS("java.ext.dirs"),

  /** Operating system name. */
  OS_NAME("os.name"),

  /** Operating system architecture. */
  OS_ARCH("os.arch"),

### `scope` { #scope }

La spécification indique aussi que le client peut envoyer un autre champ de formulaire « scope ».

Le nom du champ de formulaire est `scope` (au singulier), mais il s'agit en fait d'une longue chaîne contenant des « scopes » séparés par des espaces.

Chaque « scope » n'est qu'une chaîne (sans espaces).

Ils sont normalement utilisés pour déclarer des permissions de sécurité spécifiques, par exemple :

```Python
with open("file.txt") as file:
    file.read()
```

Python 的最近几个版本也有了一个**异步上下文管理器**，你可以通过 `async with` 来使用：

```Python
async with lifespan(app):
    await do_stuff()
```

你可以像上面一样创建一个上下文管理器或者异步上下文管理器，它的作用是在进入 `with` 块时，执行 `yield` 之前的代码，并且在离开 `with` 块时，执行 `yield` 后面的代码。

但在我们上面的例子里，我们并不是直接使用，而是传递给 FastAPI 来供其使用。