### Pydantic- und FastAPI-`examples` { #pydantic-and-fastapi-examples }

Wenn Sie `examples` innerhalb eines Pydantic-Modells hinzufügen, indem Sie `schema_extra` oder `Field(examples=["something"])` verwenden, wird dieses Beispiel dem **JSON-Schema** für dieses Pydantic-Modell hinzugefügt.

Und dieses **JSON-Schema** des Pydantic-Modells ist in der **OpenAPI** Ihrer API enthalten und wird dann in der Benutzeroberfläche der Dokumentation verwendet.

{* ../../docs_src/path_params/tutorial005_py39.py hl[1,6:9] *}

/// tip | Dica
Se você está se perguntando, "AlexNet", "ResNet" e "LeNet" são apenas nomes de <abbr title="Tecnicamente, arquiteturas de modelos de Deep Learning">modelos</abbr> de Aprendizado de Máquina.
///

### Declare um parâmetro de path { #declare-a-path-parameter }

Toda a validação é controlada pelo robusto e bem estabelecido **Pydantic**.

### Segurança e autenticação { #security-and-authentication }

Segurança e autenticação integradas. Sem nenhum compromisso com bancos de dados ou modelos de dados.

Todos os esquemas de seguranças definidos no OpenAPI, incluindo:

* HTTP Basic.

///

## Separating tests { #separating-tests }

In a real application, you probably would have your tests in a different file.

And your **FastAPI** application might also be composed of several files/modules, etc.

### **FastAPI** app file { #fastapi-app-file }

Let's say you have a file structure as described in [Bigger Applications](bigger-applications.md){.internal-link target=_blank}:

```
.
├── app

}

nullaway {
    // NullAway can use NullMarked instead, but for the adoption process it is more effective to assume that all gradle code is already annotated.
    // This way we can catch discrepancies in modules easier. We should make all packages NullMarked eventually too, but this is a separate task.
    annotatedPackages.add("org.gradle")
}

dependencies {
    attributesSchema {

    <version>11-SNAPSHOT</version>
  </parent>

  <groupId>inheritance</groupId>
  <artifactId>child-artifact-id</artifactId>
  <version>11-SNAPSHOT</version>
  <name>Model inheritance test parent: module directory != artifactId</name>
  <description>
    artifactId == "child-artifact-id"
    but expect path on SCM and site == "child"

/// tip | Consejo

`JSONResponse` en sí misma es una subclase de `Response`.

///

Y cuando devuelves una `Response`, **FastAPI** la pasará directamente.

No hará ninguna conversión de datos con los modelos de Pydantic, no convertirá los contenidos a ningún tipo, etc.

Esto te da mucha flexibilidad. Puedes devolver cualquier tipo de datos, sobrescribir cualquier declaración o validación de datos, etc.

computation
[**graphs**](https://developers.google.com/machine-learning/glossary/#graph).
Since models are practically programs that TensorFlow executes, using untrusted
models or graphs is equivalent to running untrusted code.

If you need to run untrusted models, execute them inside a
[**sandbox**](https://developers.google.com/code-sandboxing). Memory corruptions

/// note | Remarque

`JSONResponse` est elle-même une sous-classe de `Response`.

///

Et quand vous retournez une `Response`, **FastAPI** la transmet directement.

Elle ne fera aucune conversion de données avec les modèles Pydantic, elle ne convertira pas le contenu en un type quelconque.

Cela vous donne beaucoup de flexibilité. Vous pouvez retourner n'importe quel type de données, surcharger n'importe quelle déclaration ou validation de données.

[tool.mypy]
plugins = ["pydantic.mypy"]
strict = true

[[tool.mypy.overrides]]
module = "fastapi.concurrency"
warn_unused_ignores = false
ignore_missing_imports = true

[[tool.mypy.overrides]]
module = "fastapi.tests.*"
ignore_missing_imports = true
check_untyped_defs = true

[[tool.mypy.overrides]]
module = "docs_src.*"
disallow_incomplete_defs = false
disallow_untyped_defs = false