* Throwable}. But we <i>can</i> write {@code sneakyThrow(t);}.
   *
   * <p>We sometimes also use {@code sneakyThrow} for testing how our code responds to
   * sneaky checked exception.
   *
   * @return never; this method declares a return type of {@link Error} only so that callers can
   *     write {@code throw sneakyThrow(t);} to convince the compiler that the statement will always
   *     throw.
   */
  @CanIgnoreReturnValue

# Utilizando o Request diretamente { #using-the-request-directly }

Até agora você declarou as partes da requisição que você precisa utilizando os seus tipos.

Obtendo dados de:

* O path como parâmetros.
* Cabeçalhos (*Headers*).
* Cookies.
* etc.

E ao fazer isso, o **FastAPI** está validando as informações, convertendo-as e gerando documentação para a sua API automaticamente.

   * Throwable}. But we <i>can</i> write {@code sneakyThrow(t);}.
   *
   * <p>We sometimes also use {@code sneakyThrow} for testing how our code responds to
   * sneaky checked exception.
   *
   * @return never; this method declares a return type of {@link Error} only so that callers can
   *     write {@code throw sneakyThrow(t);} to convince the compiler that the statement will always
   *     throw.
   */
  @CanIgnoreReturnValue

Para eso necesitas acceder al request directamente.

{* ../../docs_src/using_request_directly/tutorial001_py39.py hl[1,7:8] *}

Al declarar un parámetro de *path operation function* con el tipo siendo `Request`, **FastAPI** sabrá pasar el `Request` en ese parámetro.

/// tip | Consejo

   * Throwable}. But we <i>can</i> write {@code sneakyThrow(t);}.
   *
   * <p>We sometimes also use {@code sneakyThrow} for testing how our code responds to
   * sneaky checked exception.
   *
   * @return never; this method declares a return type of {@link Error} only so that callers can
   *     write {@code throw sneakyThrow(t);} to convince the compiler that the statement will always
   *     throw.
   */
  @CanIgnoreReturnValue

        assertViolations(result31, 0, 0, 3);

        assertContains(
                result31.getWarnings().get(0),
                "'dependencies.dependency.scope' for test:a:jar declares usage of deprecated 'system' scope");
        assertContains(
                result31.getWarnings().get(1),

    private String language;

    @Deprecated
    private boolean addedToClasspath;

    /**
     * Default ctor for Plexus compatibility, as many plugins have artifact handlers declared in legacy Plexus XML.
     * Do not use directly!
     *
     * @deprecated This ctor is present only for Plexus XML defined component compatibility, do not use it.
     */
    @Deprecated

         * annotations in the form of JetBrains annotations, which have only class retention and
         * thus are invisible at runtime. Thus, we conclude that the parameter types are
         * *non*-nullable, even when they are declared as `Foo?`.
         */
        || hasAutomaticNullChecksFromKotlin(member);
  }

  private static boolean hasAutomaticNullChecksFromKotlin(Member member) {

                }

                currentFilePath?.let { filePath ->
                    when {
                        line.contains("These transitive dependencies should be declared directly:") -> {
                            dependencyFixes.addToMap(filePath, extractDependencies(index))
                        }

Podemos hacerlo mejor...

## Qué hace a una dependencia { #what-makes-a-dependency }

Hasta ahora has visto dependencias declaradas como funciones.

Pero esa no es la única forma de declarar dependencias (aunque probablemente sea la más común).

El factor clave es que una dependencia debe ser un "callable".

Un "**callable**" en Python es cualquier cosa que Python pueda "llamar" como una función.