Import the `Response` class (sub-class) you want to use and declare it in the *path operation decorator*.

For large responses, returning a `Response` directly is much faster than returning a dictionary.

Having the return model ensure that a value is always available and always `int` (not `None`) is very useful for the API clients, they can write much simpler code having this certainty.

Also, **automatically generated clients** will have simpler interfaces, so that the developers communicating with your API can have a much better time working with your API. 😎

///

import java.math.BigInteger;
import java.util.Random;
import org.jspecify.annotations.NullUnmarked;

/**
 * Utilities for benchmarks.
 *
 * <p>In many cases, we wish to vary the order of magnitude of the input as much as we want to vary
 * the input itself, so most methods which generate values use an exponential distribution varying
 * the order of magnitude of the generated values uniformly at random.
 *
 * @author Louis Wasserman
 */

 * client. Applications may use this class to monitor HTTP connections as members of a
 * [connection pool][ConnectionPool].
 *
 * Do not confuse this class with the misnamed `HttpURLConnection`, which isn't so much a connection
 * as a single request/response exchange.
 *
 * ## Modern TLS
 *
 * There are trade-offs when selecting which options to include when negotiating a secure connection

    }
  }

  /**
   * A trust manager for Android applications that customize the trust manager.
   *
   * This class exploits knowledge of Android implementation details. This class is potentially
   * much faster to initialize than [BasicTrustRootIndex] because it doesn't need to load and
   * index trusted CA certificates.
   */
  internal data class CustomTrustRootIndex(
    private val trustManager: X509TrustManager,

        * **FastAPI**: (uses Starlette) an API microframework with several additional features for building APIs, with data validation, etc.

* **Uvicorn**:
    * Will have the best performance, as it doesn't have much extra code apart from the server itself.

   * invokeAny} which are pending serial execution, even the subset of the tasks that have not yet
   * started execution. It is unclear from the {@code ExecutorService} specification if these should
   * be included, and it's much easier to implement the interpretation that they not be. Finally, a
   * call to {@code shutdown} or {@code shutdownNow} may result in concurrent calls to {@code

    when {
      PlatformVersion.majorVersion > 11 ->
        assertThat(s.enabledProtocols.toList()).containsExactly(
          "TLSv1.3",
          "TLSv1.2",
        )
      // Not much we can guarantee on JDK 11.
      PlatformVersion.majorVersion == 11 ->
        assertThat(s.enabledProtocols.toList()).contains(
          "TLSv1.2",
        )
      // JDK 8 291 removed older versions

Your server(s) is (are) a **resource**, you can consume or **utilize**, with your programs, the computation time on the CPUs, and the RAM memory available.

How much of the system resources do you want to be consuming/utilizing? It might be easy to think "not much", but in reality, you will probably want to consume **as much as possible without crashing**.

   * invokeAny} which are pending serial execution, even the subset of the tasks that have not yet
   * started execution. It is unclear from the {@code ExecutorService} specification if these should
   * be included, and it's much easier to implement the interpretation that they not be. Finally, a
   * call to {@code shutdown} or {@code shutdownNow} may result in concurrent calls to {@code