```Python hl_lines="1  3" title="app/routers/users.py"
{!../../../docs_src/bigger_applications/app/routers/users.py!}
```

### *Pfadoperationen* mit `APIRouter`

Und dann verwenden Sie ihn, um Ihre *Pfadoperationen* zu deklarieren.

Verwenden Sie ihn auf die gleiche Weise wie die Klasse `FastAPI`:

```Python hl_lines="6  11  16" title="app/routers/users.py"
{!../../../docs_src/bigger_applications/app/routers/users.py!}
```

* **Machine Learning**: it normally requires lots of "matrix" and "vector" multiplications. Think of a huge spreadsheet with numbers and multiplying all of them together at the same time.
* **Deep Learning**: this is a sub-field of Machine Learning, so, the same applies. It's just that there is not a single spreadsheet of numbers to multiply, but a huge set of them, and in many cases, you use a special processor to build and / or use those models.

 *
 * Most applications should share a process-wide [TaskRunner] and use queues for per-client work.
 */
class TaskRunner(
  val backend: Backend,
  internal val logger: Logger = TaskRunner.logger,
) {
  val lock: ReentrantLock = ReentrantLock()
  val condition: Condition = lock.newCondition()

  private var nextQueueName = 10000
  private var coordinatorWaiting = false

  fun connectionPreface() {
    this.withLock {
      if (closed) throw IOException("closed")
      if (!client) return // Nothing to write; servers don't send connection headers!
      if (logger.isLoggable(FINE)) {
        logger.fine(format(">> CONNECTION ${CONNECTION_PREFACE.hex()}"))
      }
      sink.write(CONNECTION_PREFACE)
      sink.flush()
    }
  }

  /** Applies `peerSettings` and then sends a settings ACK. */

      if (increment == 0L) throw IOException("windowSizeIncrement was 0")
    } catch (e: Exception) {
      logger.fine(frameLog(true, streamId, length, TYPE_WINDOW_UPDATE, flags))
      throw e
    }
    if (logger.isLoggable(FINE)) {
      logger.fine(
        frameLogWindowUpdate(
          inbound = true,
          streamId = streamId,
          length = length,

```Python hl_lines="1  3" title="app/routers/users.py"
{!../../../docs_src/bigger_applications/app/routers/users.py!}
```

### *➡ 🛠️* ⏮️ `APIRouter`

& ⤴️ 👆 ⚙️ ⚫️ 📣 👆 *➡ 🛠️*.

⚙️ ⚫️ 🎏 🌌 👆 🔜 ⚙️ `FastAPI` 🎓:

```Python hl_lines="6  11  16" title="app/routers/users.py"
{!../../../docs_src/bigger_applications/app/routers/users.py!}
```

👆 💪 💭 `APIRouter` "🐩 `FastAPI`" 🎓.

    if (assertionsEnabled && taskRunner.lock.isHeldByCurrentThread) {
      throw AssertionError("Thread ${Thread.currentThread().name} MUST NOT hold lock on $this")
    }
  }

  val logger = Logger.getLogger("TaskFaker." + instance++)

  /** Though this executor service may hold many threads, they are not executed concurrently. */
  private val tasksExecutor = Executors.newCachedThreadPool(threadFactory("TaskFaker"))

```Python hl_lines="1  3" title="app/routers/users.py"
{!../../../docs_src/bigger_applications/app/routers/users.py!}
```

### *Path operations* with `APIRouter`

And then you use it to declare your *path operations*.

Use it the same way you would use the `FastAPI` class:

```Python hl_lines="6  11  16" title="app/routers/users.py"
{!../../../docs_src/bigger_applications/app/routers/users.py!}
```

你可以导入它并通过与 `FastAPI` 类相同的方式创建一个「实例」：

```Python hl_lines="1  3" title="app/routers/users.py"
{!../../../docs_src/bigger_applications/app/routers/users.py!}
```

### 使用 `APIRouter` 的*路径操作*

然后你可以使用它来声明*路径操作*。

使用方式与 `FastAPI` 类相同：

```Python hl_lines="6  11  16" title="app/routers/users.py"
{!../../../docs_src/bigger_applications/app/routers/users.py!}
```

你可以将 `APIRouter` 视为一个「迷你 `FastAPI`」类。

所有相同的选项都得到支持。

Кроме того, Ваше приложение **FastAPI** может состоять из нескольких файлов, модулей и т.п.

### Файл приложения **FastAPI**

Допустим, структура файлов Вашего приложения похожа на ту, что описана на странице [Более крупные приложения](bigger-applications.md){.internal-link target=_blank}:

```
.
├── app
│   ├── __init__.py
│   └── main.py
```