E um retorno com o código de status `200` que utiliza o seu `response_model`, porém inclui um `example` customizado:

```Python hl_lines="20-31"
{!../../docs_src/additional_responses/tutorial003.py!}
```

Isso será combinado e incluído em seu OpenAPI, e disponibilizado na documentação da sua API:

<img src="/img/tutorial/additional-responses/image01.png">

## Combinar retornos predefinidos e personalizados

 *
 * This is because the ModelProcessor interface extends ModelLocator and ModelReader. If we
 * made this component available under all its interfaces then it could end up being injected
 * into itself leading to a stack overflow.
 *
 * A side effect of using @Typed is that it translates to explicit bindings in the container.
 * So instead of binding the component under a 'wildcard' key it is now bound with an explicit

# OAuth2 scopes

You can use OAuth2 scopes directly with **FastAPI**, they are integrated to work seamlessly.

This would allow you to have a more fine-grained permission system, following the OAuth2 standard, integrated into your OpenAPI application (and the API docs).

OAuth2 with scopes is the mechanism used by many big authentication providers, like Facebook, Google, GitHub, Microsoft, Twitter, etc. They use it to provide specific permissions to users and applications.

```Python hl_lines="3  6-8"
{!../../docs_src/sub_applications/tutorial001.py!}
```

### Unteranwendung

Erstellen Sie dann Ihre Unteranwendung und deren *Pfadoperationen*.

Diese Unteranwendung ist nur eine weitere Standard-FastAPI-Anwendung, aber diese wird „gemountet“:

```Python hl_lines="11  14-16"
{!../../docs_src/sub_applications/tutorial001.py!}
```

### Die Unteranwendung mounten

              @Override
              protected void finalize() {
                finalizerRan.countDown();
              }
            });

    // Don't copy this into your own test!
    // Use e.g. awaitClear or await(CountDownLatch) instead.
    GcFinalization.awaitFullGc();

    // If this test turns out to be flaky, add a second call to awaitFullGc()

# "pip tests" run a similar suite of tests the "nonpip" tests, but do something
# odd to attempt to validate the quality of the pip package. The wheel is
# installed into a virtual environment, and then that venv is used to run all
# bazel tests with a special flag "--define=no_tensorflow_py_deps=true", which
# drops all the bazel dependencies for each py_test; this makes all the tests

`tf` instead of `tensorflow`, following the common practice of importing
TensorFlow via the command `import tensorflow as tf`.

The primary function of this module is to import all of the public TensorFlow
interfaces into a single place. The interfaces themselves are located in
sub-modules, as described below.

Note that the file `__init__.py` in the TensorFlow source code tree is actually

   *       CompactHashing.createTable, whose values are either
   *       <ul>
   *         <li>UNSET, meaning "null pointer"
   *         <li>one plus an index into the entries and elements array
   *       </ul>
   *   <li>another java.util.Set delegate implementation. In most modern JDKs, normal java.util hash

  // `handSize` is a small positive number that configures the
  // shuffle sharding of requests into queues.  When enqueuing a request
  // at this priority level the request's flow identifier (a string
  // pair) is hashed and the hash value is used to shuffle the list
  // of queues and deal a hand of the size specified here.  The
  // request is put into one of the shortest queues in that hand.

          // Defeat branch predictor for: c < 0x80 ; branch taken 50% of the time.
          return 0x100;
        } else if (userFriendly.matches("(?i)(?:Greek|Cyrillic|European|ISO.?8859)")) {
          // Mostly 2-byte UTF-8 sequences - "European" text
          return 0x800;
        } else if (userFriendly.matches("(?i)(?:Chinese|Han|Asian|BMP)")) {
          // Mostly 3-byte UTF-8 sequences - "Asian" text