* Validate **complex structures**:
    * Use of hierarchical Pydantic models, Python `typing`’s `List` and `Dict`, etc.
    * And validators allow complex data schemas to be clearly and easily defined, checked and documented as JSON Schema.
    * You can have deeply **nested JSON** objects and have them all validated and annotated.
* **Extensible**:

## Deprecating parameters { #deprecating-parameters }

Now let's say you don't like this parameter anymore.

You have to leave it there a while because there are clients using it, but you want the docs to clearly show it as <abbr title="obsolete, recommended not to use it">deprecated</abbr>.

Then pass the parameter `deprecated=True` to `Query`:

{* ../../docs_src/query_params_str_validations/tutorial010_an_py310.py hl[19] *}

   * is defined as the function h such that {@code h(a) == g(f(a))} for each {@code a}.
   *
   * <p><b>JRE users and Android users who opt in to library desugaring:</b> use {@code
   * g.compose(f)} or (probably clearer) {@code f.andThen(g)} instead. Note that it is not
   * serializable.
   *
   * @param g the second function to apply
   * @param f the first function to apply
   * @return the composition of {@code f} and {@code g}

     and conditions of the license they choose before applying it.
     Licensors should also secure all rights necessary before
     applying our licenses so that the public can reuse the
     material as expected. Licensors should clearly mark any
     material not subject to the license. This includes other CC-
     licensed material, or material used under an exception or
     limitation to copyright. More considerations for licensors:

      executorService.execute(
          () -> {
            lock.lock();
            try {
              startUp();
              /*
               * requireNonNull is safe because executorService is never cleared after the
               * assignment above.
               */
              requireNonNull(executorService);
              runningTask = scheduler().schedule(delegate, executorService, task);

        }
        /*
         * requireNonNull is safe: The first call to this method always enters the !hasNext() case
         * and populates currentKey, after which it's never cleared.
         */
        return immutableEntry(requireNonNull(currentKey), valueItr.next());
      }
    };
  }

  @Override
  Spliterator<Entry<K, V>> entrySpliterator() {

/// note | Technical Details

It used <a href="https://github.com/MagicStack/uvloop" class="external-link" target="_blank">`uvloop`</a> instead of the default Python `asyncio` loop. That's what made it so fast.

It clearly inspired Uvicorn and Starlette, that are currently faster than Sanic in open benchmarks.

///

/// check | Inspired **FastAPI** to

Find a way to have a crazy performance.

      executorService.execute(
          () -> {
            lock.lock();
            try {
              startUp();
              /*
               * requireNonNull is safe because executorService is never cleared after the
               * assignment above.
               */
              requireNonNull(executorService);
              runningTask = scheduler().schedule(delegate, executorService, task);

      }
      /*
       * uncheckedCastNullableTToT is safe: The first call to this method always enters the !hasNext() case and
       * populates key, after which it's never cleared.
       */
      return output(uncheckedCastNullableTToT(key), valueIterator.next());
    }

    @Override
    public void remove() {
      valueIterator.remove();
      /*

  }

  /**
   * This test depends on the fact that {@code AbstractSequentialList.iterator} transforms the
   * {@code iterator()} call into a call on {@code listIterator(int)}. This is fine because the
   * behavior is clearly documented so it's not expected to change.
   */
  public void testTransformedSequentialIterationUsesBackingListIterationOnly() {
    List<Integer> randomAccessList = new ArrayList<>(SOME_SEQUENTIAL_LIST);