}

  TFE_Execute(op_.get(), unwrapped_results.data(), &real_num_outputs, status);
  if (TF_GetCode(status) != TF_OK) {
    cancellation_manager_->StartCancel();
    return;
  }
  unwrapped_results.resize(real_num_outputs);
  outputs->reserve(real_num_outputs);
  for (TFE_TensorHandle* unwrapped_result : unwrapped_results) {
    outputs->emplace_back(unwrapped_result);
  }
}

/// note | "Technical Details"

This only happens when you install using this included `requirements.txt` instead of running `pip install fastapi` directly.

That is because inside the `requirements.txt` file, the local version of FastAPI is marked to be installed in "editable" mode, with the `-e` option.

///

### Format the code

---

But for this example, we'll use a very simple HTML document with some JavaScript, all inside a long string.

This, of course, is not optimal and you wouldn't use it for production.

In production you would have one of the options above.

But it's the simplest way to focus on the server-side of WebSockets and have a working example:

```Python hl_lines="2  6-38  41-43"

        Platform.format("key %s not found in entries %s", newEntry.getKey(), expected));
  }

  /**
   * Wrapper for {@link Map#get(Object)} that forces the caller to pass in a key of the same type as
   * the map. Besides being slightly shorter than code that uses {@link #getMap()}, it also ensures
   * that callers don't pass an {@link Entry} by mistake.
   */
  protected V get(K key) {
    return getMap().get(key);
  }

from tensorflow.python.tools import module_util as _module_util
from tensorflow.python.util.lazy_loader import KerasLazyLoader as _KerasLazyLoader

# Make sure code inside the TensorFlow codebase can use tf2.enabled() at import.
_os.environ["TF2_BEHAVIOR"] = "1"
from tensorflow.python import tf2 as _tf2
_tf2.enable()

# API IMPORTS PLACEHOLDER

# WRAPPER_PLACEHOLDER

Create metadata for your tags and pass it to the `openapi_tags` parameter:

```Python hl_lines="3-16  18"
{!../../docs_src/metadata/tutorial004.py!}
```

Notice that you can use Markdown inside of the descriptions, for example "login" will be shown in bold (**login**) and "fancy" will be shown in italics (_fancy_).

/// tip

You don't have to add metadata for all the tags that you use.

///

* Serveur web Python **prêt au déploiement** utilisant Uvicorn et Gunicorn.
* Backend Python <a href="https://github.com/fastapi/fastapi" class="external-link" target="_blank">**FastAPI**</a> :
    * **Rapide** : Très hautes performances, comparables à **NodeJS** ou **Go** (grâce à Starlette et Pydantic).

  // CSI specification and implemented in Kubernetes via the usual PV/PVC mechanism.
  //
  // The other mode is "Ephemeral". In this mode, volumes are defined inline inside the pod spec
  // with CSIVolumeSource and their lifecycle is tied to the lifecycle of that pod.
  // A driver has to be aware of this because it is only going to get a NodePublishVolume call for such a volume.
  //

* Des performances vraiment impressionnantes. C'est l'<a href="https://github.com/encode/starlette#performance" class="external-link" target="_blank">un des framework Python les plus rapide, à égalité avec **NodeJS** et **GO**</a>.
* Le support des **WebSockets**.
* Le support de **GraphQL**.
* Les <abbr title="En anglais: In-process background tasks">tâches d'arrière-plan.</abbr>

  // CSI specification and implemented in Kubernetes via the usual PV/PVC mechanism.
  //
  // The other mode is "Ephemeral". In this mode, volumes are defined inline inside the pod spec
  // with CSIVolumeSource and their lifecycle is tied to the lifecycle of that pod.
  // A driver has to be aware of this because it is only going to get a NodePublishVolume call for such a volume.
  //