}
        return numOfDeleted;
    }

    /**
     * Refreshes the specified index to make recent changes visible for search.
     * This operation ensures that all pending writes are persisted and searchable.
     *
     * @param searchEngineClient the search engine client to use for refresh
     * @param index the index name to refresh
     * @return the refresh status code
     */

   * is used only by those methods, and it is never set in a "normal" instance.
   *
   * <p>This class needs to write deserialized data into fields that are {@code final transient}.
   * Such writes will become impossible to perform in {@link #readObject} after JEP 500. Instead, we
   * must create a new instance with the desired field values, stash it in this field, and then

`dict`를 정의하려면, 쉼표로 구분된 2개의 타입 매개변수를 전달합니다.

첫 번째 타입 매개변수는 `dict`의 키를 위한 것입니다.

두 번째 타입 매개변수는 `dict`의 값을 위한 것입니다:

{* ../../docs_src/python_types/tutorial008_py310.py hl[1] *}

이는 다음을 의미합니다:

* 변수 `prices`는 `dict`입니다:
    * 이 `dict`의 키는 `str` 타입입니다(예: 각 아이템의 이름).
    * 이 `dict`의 값은 `float` 타입입니다(예: 각 아이템의 가격).

#### Union { #union }

변수가 **여러 타입 중 어떤 것이든** 될 수 있다고 선언할 수 있습니다. 예를 들어 `int` 또는 `str`입니다.

   * the full, recursive stack trace of {@code throwable}. Note that you probably should not be
   * parsing the resulting string; if you need programmatic access to the stack frames, you can call
   * {@link Throwable#getStackTrace()}.
   */
  @GwtIncompatible // java.io.PrintWriter, java.io.StringWriter
  public static String getStackTraceAsString(Throwable throwable) {

e.namespaceURI===ot?!(t.namespaceURI===nt&&!st[o])&&!(t.namespaceURI===tt&&!ct[o])&&!Et[n]&&(ut[n]||!yt[n]):!("application/xhtml+xml"!==mt||!at[e.namespaceURI]))}(e)?(At(e),!0):"noscript"!==n&&"noembed"!==n&&"noframes"!==n||!b(/<\/no(script|embed|frames)/i,e.innerHTML)?(Ue&&e.nodeType===Q&&(t=e.textContent,u([he,ge,Te],(e=>{t=y(t,e," ")})),e.textContent!==t&&(f(o.removed,{element:e.cloneNode()}),e.textContent=t)),Rt(de.afterSanitizeElements,e,null),!1):(At(e),!0)},Ct=function(e,t,n){if(ve[t])ret...

    }

    @Override
    String atomicHelperTypeForTest() {
      return "AtomicReferenceFieldUpdaterAtomicHelper";
    }
  }

  /**
   * {@link AtomicHelper} based on {@code synchronized} and volatile writes.
   *
   * <p>This is an implementation of last resort for when certain basic VM features are broken (like
   * AtomicReferenceFieldUpdater).
   */
  private static final class SynchronizedHelper extends AtomicHelper {

   * that they were not initialized yet.
   */
  private static final int SUFFIX_NOT_INITIALIZED = -2;

  /**
   * Maximum parts (labels) in a domain name. This value arises from the 255-octet limit described
   * in <a href="http://www.ietf.org/rfc/rfc2181.txt">RFC 2181</a> part 11 with the fact that the
   * encoding of each part occupies at least two bytes (dot plus label externally, length byte plus

 * modifications are visible, synchronize access to the table. Unlike other {@code Table}
 * implementations, synchronization is unnecessary between a thread that writes to one cell and a
 * thread that reads from another.
 *
 * <p>See the Guava User Guide article on <a href=
 * "https://github.com/google/guava/wiki/NewCollectionTypesExplained#table">{@code Table}</a>.
 *

  /*
   * The basic strategy is to subdivide the table among Segments, each of which itself is a
   * concurrently readable hash table. The map supports non-blocking reads and concurrent writes
   * across different segments.
   *
   * If a maximum size is specified, a best-effort bounding is performed per segment, using a
   * page-replacement algorithm to determine which entries to evict when the capacity has been

            writer.flush();
        } catch (final JsonProcessingException e) {
            logger.warn("[RAG] Failed to serialize SSE data. event={}", event, e);
        }
    }

    /**
     * Writes a JSON response to the HTTP response.
     *
     * @param response the HTTP response
     * @param status the HTTP status code
     * @param data the data to serialize as JSON