Puedes retornar un `dict`, `list`, valores singulares como `str`, `int`, etc.

También puedes retornar modelos de Pydantic (verás más sobre eso más adelante).

Hay muchos otros objetos y modelos que serán automáticamente convertidos a JSON (incluyendo ORMs, etc). Intenta usar tus favoritos, es altamente probable que ya sean compatibles.

### Paso 6: Despliégalo { #step-6-deploy-it }

```console
$ pip install "fastapi[all]"
---> 100%
```

</div>

### Create the `Settings` object { #create-the-settings-object }

Import `BaseSettings` from Pydantic and create a sub-class, very much like with a Pydantic model.

The same way as with Pydantic models, you declare class attributes with type annotations, and possibly default values.

    assertThat((long) arraysDim1 * arraysDim2).isNotEqualTo((long) (arraysDim1 * arraysDim2));

    byte[][] arrays = new byte[arraysDim1][];
    // it's shared to avoid using too much memory in tests
    byte[] sharedArray = new byte[arraysDim2];
    Arrays.fill(arrays, sharedArray);

    try {
      Bytes.concat(arrays);
      fail();
    } catch (IllegalArgumentException expected) {

but that should be extremely rare.

GODEBUG settings added for compatibility will be maintained
for a minimum of two years (four Go releases).
Some, such as `http2client` and `http2server`,
will be maintained much longer, even indefinitely.

When possible, each GODEBUG setting has an associated
[runtime/metrics](/pkg/runtime/metrics/) counter
named `/godebug/non-default-behavior/<name>:events`

   *
   * <p><i>Performance notes:</i> while the cartesian product of lists of size {@code m, n, p} is a
   * list of size {@code m x n x p}, its actual memory consumption is much smaller. When the
   * cartesian product is constructed, the input lists are merely copied. Only as the resulting list
   * is iterated are the individual lists created, and these are not retained after iteration.
   *

    assertThrows(IndexOutOfBoundsException.class, () -> iia3.subArray(1, 4));
  }

  /*
   * Whenever an implementation uses `instanceof` on a parameter instance, the test has to know that
   * (so much for "black box") and try instances that both do and don't pass the check. The "don't"
   * half of that is more awkward to arrange...
   */
  private static <T> Iterable<T> iterable(Collection<T> collection) {

		}
		// For client requests, Request.ContentLength of 0
		// means either actually 0, or unknown. The only way
		// to explicitly say that the ContentLength is zero is
		// to set the Body to nil. But turns out too much code
		// depends on NewRequest returning a non-nil Body,
		// so we use a well-known ReadCloser variable instead
		// and have the http package also treat that sentinel
		// variable to mean explicitly zero.

 *
 * <p>Unlike {@code java.util.HashMap}, iteration is only proportional to the actual {@code size()},
 * which is optimal, and <i>not</i> the size of the internal hashtable, which could be much larger
 * than {@code size()}. Furthermore, this structure places significantly reduced load on the garbage
 * collector by only using a constant number of internal objects.
 *

h3. Plugins need a specific metadata model

Each application will have its own model for plugins. Much of this model will be common across applications but there will always be issues specific to a particular applications’s plugins. Some of the issues that need to be taken into consideration:

* Plugin dependencies

	// Make sure we can fit "size" on to the disk without getting above the diskFillFraction
	if available < uint64(size) {
		return false, nil
	}

	// How much will be left after adding the file.
	available -= uint64(size)

	// wantLeft is how much space there at least must be left.
	wantLeft := uint64(float64(total) * (1.0 - diskFillFraction))
	return available > wantLeft, nil