///

/// tip

If you are wondering, "AlexNet", "ResNet", and "LeNet" are just names of Machine Learning <abbr title="Technically, Deep Learning model architectures">models</abbr>.

///

### Declare a *path parameter*

Then create a *path parameter* with a type annotation using the enum class you created (`ModelName`):

```Python hl_lines="16"

```

Vous pouvez retourner un dictionnaire (`dict`), une liste (`list`), des valeurs seules comme des chaines de caractères (`str`) et des entiers (`int`), etc.

Vous pouvez aussi retourner des models **Pydantic** (qui seront détaillés plus tard).

Il y a de nombreux autres objets et modèles qui seront automatiquement convertis en JSON. Essayez d'utiliser vos favoris, il est fort probable qu'ils soient déjà supportés.

    * Chuyển đổi kiểu dữ liệu Python (`str`, `int`, `float`, `bool`, `list`,...).
    * `datetime` objects.
    * `UUID` objects.
    * Database models.
    * ...và nhiều hơn thế.
* Tự động tạo tài liệu tương tác API, bao gồm 2 giao diện người dùng:
    * Swagger UI.
    * ReDoc.

---

Quay trở lại ví dụ trước, **FastAPI** sẽ thực hiện:

* **Güvenli şifreleme** .
* **JWT token** kimlik doğrulama.
* **SQLAlchemy** models (Flask dan bağımsızdır. Celery worker'ları ile kullanılabilir).
* Kullanıcılar için temel başlangıç ​​modeli (gerektiği gibi değiştirin ve kaldırın).
* **Alembic** migration.
* **CORS** (Cross Origin Resource Sharing).

    *,
    cloned_types: Optional[MutableMapping[Type[BaseModel], Type[BaseModel]]] = None,
) -> ModelField:
    if PYDANTIC_V2:
        return field
    # cloned_types caches already cloned types to support recursive models and improve
    # performance by avoiding unnecessary cloning
    if cloned_types is None:
        cloned_types = _CLONED_TYPES_CACHE

    original_type = field.type_

from tests.utils import needs_py39, needs_py310


def clear_sqlmodel():
    # Clear the tables in the metadata for the default base model
    SQLModel.metadata.clear()
    # Clear the Models associated with the registry, to avoid warnings
    default_registry.dispose()


@pytest.fixture(
    name="client",
    params=[
        "tutorial001",
        pytest.param("tutorial001_py39", marks=needs_py39),

/// info

Note that the `TestClient` receives data that can be converted to JSON, not Pydantic models.

If you have a Pydantic model in your test and you want to send its data to the application during testing, you can use the `jsonable_encoder` described in [JSON Compatible Encoder](encoder.md){.internal-link target=_blank}.

from typing import Optional

from fastapi.openapi.models import APIKey, APIKeyIn
from fastapi.security.base import SecurityBase
from starlette.exceptions import HTTPException
from starlette.requests import Request
from starlette.status import HTTP_403_FORBIDDEN
from typing_extensions import Annotated, Doc


class APIKeyBase(SecurityBase):
    pass


class APIKeyQuery(APIKeyBase):
    """

from tests.utils import needs_py39, needs_py310


def clear_sqlmodel():
    # Clear the tables in the metadata for the default base model
    SQLModel.metadata.clear()
    # Clear the Models associated with the registry, to avoid warnings
    default_registry.dispose()


@pytest.fixture(
    name="client",
    params=[
        "tutorial002",
        pytest.param("tutorial002_py39", marks=needs_py39),

### Concurrency + Parallelism: Web + Machine Learning

With **FastAPI** you can take advantage of concurrency that is very common for web development (the same main attraction of NodeJS).