json_schema_extra=json_schema_extra,
            **extra,
        )


@dataclass(frozen=True)
class Depends:
    dependency: Optional[Callable[..., Any]] = None
    use_cache: bool = True
    scope: Union[Literal["function", "request"], None] = None


@dataclass(frozen=True)
class Security(Depends):

    "field_title_generator": None,
    "description": None,
    "examples": None,
    "exclude": None,
    "exclude_if": None,
    "discriminator": None,
    "deprecated": None,
    "json_schema_extra": None,
    "frozen": None,
    "validate_default": None,
    "repr": True,
    "init": None,
    "init_var": None,
    "kw_only": None,
}


# TODO: remove when dropping support for Pydantic < v2.12.3

// Supports following actions:
// - restart (restarts all the MinIO instances in a setup)
// - stop (stops all the MinIO instances in a setup)
// - freeze (freezes all incoming S3 API calls)
// - unfreeze (unfreezes previously frozen S3 API calls)
func (a adminAPIHandlers) ServiceHandler(w http.ResponseWriter, r *http.Request) {
	ctx := r.Context()

	vars := mux.Vars(r)
	action := vars["action"]

	var serviceSig serviceSignal

	}
	return sys.collectPeerMetrics(ctx, peerChannels, g)
}

// ServiceFreeze freezes all S3 API calls when 'freeze' is true,
// 'freeze' is 'false' would resume all S3 API calls again.
// NOTE: once a tenant is frozen either two things needs to
// happen before resuming normal operations.
//   - Server needs to be restarted 'mc admin service restart'
//   - 'freeze' should be set to 'false' for this call

def broken_dep() -> Any:
    yield "s"
    raise ValueError("Broken after yield")


app = FastAPI()


@app.get("/catching")
def catching(d: Annotated[str, Depends(catching_dep)]) -> Any:
    raise CustomError("Simulated error during streaming")


@app.get("/broken")
def broken(d: Annotated[str, Depends(broken_dep)]) -> Any:
    return {"message": "all good?"}

In vielen Fällen kennen Sie möglicherweise bereits die Antwort auf diese Fragen. 🤓

Wenn Sie vielen Menschen bei ihren Fragen helfen, werden Sie offizieller [FastAPI-Experte](fastapi-people.md#fastapi-experts){.internal-link target=_blank}. 🎉

    def iter_data():
        yield from session

    return StreamingResponse(iter_data())


@app.get("/broken-session-data")
def get_broken_session_data(session: BrokenSessionDep) -> Any:
    return list(session)


@app.get("/broken-session-stream")
def get_broken_session_stream(session: BrokenSessionDep) -> Any:
    def iter_data():
        yield from session

@app.websocket("/ws")
async def websocket_endpoint(websocket: WebSocket, session: SessionDep):
    await websocket.accept()
    for item in session:
        await websocket.send_text(f"{item}")


@app.websocket("/ws-broken")
async def websocket_endpoint_broken(websocket: WebSocket, session: BrokenSessionDep):
    await websocket.accept()
    for item in session:
        await websocket.send_text(f"{item}")  # pragma no cover

    inverse.delegate.clear();
    Entry<K, V> broken = null;
    Iterator<Entry<K, V>> itr = this.delegate.entrySet().iterator();
    while (itr.hasNext()) {
      Entry<K, V> entry = itr.next();
      K k = entry.getKey();
      V v = entry.getValue();
      K conflict = inverse.delegate.putIfAbsent(v, k);
      if (conflict != null) {
        broken = entry;

    to the layer came from a keras input.
*   Clean up `BatchNormalization` layer's `trainable` property to act like
    standard python state when it's used inside `tf.functions` (frozen at
    tracing time), instead of acting like a pseudo-variable whose updates *kind
    of sometimes* get reflected in already-traced `tf.function` traces.
*   Add the `Conv1DTranspose` layer.