And also after the `response` is generated, before returning it.

For example, you could add a custom header `X-Process-Time` containing the time in seconds that it took to process the request and generate a response:

```Python hl_lines="10  12-13"
{!../../../docs_src/middleware/tutorial001.py!}
```

## Other middlewares

    github_event = PartialGitHubEvent.parse_raw(contents)

    # Avoid race conditions with multiple labels
    sleep_time = random.random() * 10  # random number between 0 and 10 seconds
    logging.info(
        f"Sleeping for {sleep_time} seconds to avoid "
        "race conditions and multiple comments"
    )
    time.sleep(sleep_time)

    # Get PR
    logging.debug(f"Processing PR: #{github_event.pull_request.number}")

* `datetime.timedelta`:
    * A Python `datetime.timedelta`.
    * In requests and responses will be represented as a `float` of total seconds.
    * Pydantic also allows representing it as a "ISO 8601 time diff encoding", <a href="https://docs.pydantic.dev/latest/concepts/serialization/#json_encoders" class="external-link" target="_blank">see the docs for more info</a>.

Then, for the request in the next tab, your app will wait for one second less, and so on.

This means that it will end up finishing some of the last tabs' requests earlier than some of the previous ones.

```Python
old_dict = {
    "old key": "old value",
    "second old key": "second old value",
}
new_dict = {**old_dict, "new key": "new value"}
```

Ici, `new_dict` contiendra toutes les paires clé-valeur de `old_dict` plus la nouvelle paire clé-valeur :

```Python
{
    "old key": "old value",
    "second old key": "second old value",
    "new key": "new value",
}
```

```Python
my_user: User = User(id=3, name="John Doe", joined="2018-07-19")

second_user_data = {
    "id": 4,
    "name": "Mary",
    "joined": "2018-11-30",
}

my_second_user: User = User(**second_user_data)
```

!!! info
    `**second_user_data` nghĩa là:

def test_get_invalid_second_header_items():
    response = client.get(
        "/items/", headers={"X-Token": "fake-super-secret-token", "X-Key": "invalid"}
    )
    assert response.status_code == 400, response.text
    assert response.json() == {"detail": "X-Key header invalid"}


def test_get_invalid_second_header_users():
    response = client.get(

    The HTTP authorization credentials in the result of using `HTTPBearer` or
    `HTTPDigest` in a dependency.

    The HTTP authorization header value is split by the first space.

    The first part is the `scheme`, the second part is the `credentials`.

    For example, in an HTTP Bearer token scheme, the client will send a header
    like:

    ```
    Authorization: Bearer deadbeef12346
    ```

    In this case:

Cela va valider les données, montrer une erreur claire si ces dernières ne sont pas valides, et documenter le paramètre dans le schéma `OpenAPI` de cette *path operation*.

## Rajouter plus de validation

Vous pouvez aussi rajouter un second paramètre `min_length` :

```Python hl_lines="9"
{!../../../docs_src/query_params_str_validations/tutorial003.py!}
```

## Ajouter des validations par expressions régulières

Quand vous créez des *fonctions de chemins*, vous pouvez vous retrouver dans une situation où vous avez un chemin fixe.

Tel que `/users/me`, disons pour récupérer les données sur l'utilisateur actuel.

Et vous avez un second chemin : `/users/{user_id}` pour récupérer de la donnée sur un utilisateur spécifique grâce à son identifiant d'utilisateur