Por lo tanto, podrías añadir datos adicionales al esquema generado automáticamente.

Por ejemplo, podrías decidir leer y validar el request con tu propio código, sin usar las funcionalidades automáticas de FastAPI con Pydantic, pero aún podrías querer definir el request en el esquema de OpenAPI.

Podrías hacer eso con `openapi_extra`:

import java.util.regex.Pattern;
import java.util.stream.Stream;

/**
 * Generates Javadocs in a particular way.
 *
 * TODO: We should remove the workarounds here and migrate some of the changes here into the Javadoc task proper.
 */
public abstract class GradleJavadocsPlugin implements Plugin<Project> {

    @Inject
    protected abstract FileSystemOperations getFs();

    @Override
    public void apply(Project project) {

                throw new IOException("Failed to create new transport");
            }

            // Reconnect
            channel.setState(ChannelState.CONNECTING);
            // Connection would be ensured through proper transport interface

            // Re-establish channel binding
            manager.performChannelBinding(channel);

            // Update channel with new transport
            channel.setTransport(newTransport);

   * co.uk} or {@code pvt.k12.wy.us}. Examples of domain names that are <i>not</i> public suffixes
   * include {@code google.com}, {@code foo.co.uk}, and {@code myblog.blogspot.com}.
   *
   * <p>Public suffixes are a proper superset of {@linkplain #isRegistrySuffix() registry suffixes}.
   * The list of public suffixes additionally contains privately owned domain names under which

            String result = response.toString();

            assertNotNull(result);
            assertTrue(result.contains("buffer="));
        }

        @Test
        @DisplayName("Should have proper toString format")
        void testToStringFormat() {
            String result = response.toString();

            assertNotNull(result);
            assertTrue(result.startsWith("Trans2GetDfsReferralResponse["));

                // Exponential backoff with overflow protection
                if (retryDelayMs > 0) {
                    try {
                        // Safe calculation: ensure retryCount is at least 1 for proper exponential backoff
                        int safeShift = Math.min(MAX_BACKOFF_SHIFT, Math.max(1, retryCount));
                        long multiplier = 1L << safeShift;

Ainsi, vous pouvez ajouter des données supplémentaires au schéma généré automatiquement.

Par exemple, vous pouvez décider de lire et de valider la requête avec votre propre code, sans utiliser les fonctionnalités automatiques de validation proposée par Pydantic, mais vous pouvez toujours définir la requête dans le schéma OpenAPI.

Vous pouvez le faire avec `openapi_extra` :

    private ArrayWithDuplicate(E[] elements, E duplicate) {
      this.elements = elements;
      this.duplicate = duplicate;
    }
  }

  /**
   * @return an array of the proper size with a duplicate element. The size must be at least three.
   */
  protected ArrayWithDuplicate<E> createArrayWithDuplicateElement() {
    E[] elements = createSamplesArray();

     * The logger returned by this method must not be cached in an instance field during the construction of the mojo.
     * This would cause the mojo to use a wrongly configured default logger when being run by Maven. The proper logger
     * gets injected by the Plexus container after the mojo has been constructed. Therefore, simply call this
     * method directly whenever you need the logger, it is fast enough and needs no caching.
     *

        assertFalse(DefaultMirrorSelector.isExternalRepo(getRepo("foo", "file:///somepath")));
        assertFalse(DefaultMirrorSelector.isExternalRepo(getRepo("foo", "file://D:/somepath")));

        // not a proper url so returns false;
        assertFalse(DefaultMirrorSelector.isExternalRepo(getRepo("foo", "192.168.101.1")));
        assertFalse(DefaultMirrorSelector.isExternalRepo(getRepo("foo", "")));
    }

    @Test