Map<K, Collection<V>> asMap(Multimap<K, V> multimap) {
    return multimap.asMap();
  }

  /**
   * Returns a multimap view of the specified map. The multimap is backed by the map, so changes to
   * the map are reflected in the multimap, and vice versa. If the map is modified while an
   * iteration over one of the multimap's collection views is in progress (except through the

      Map<K, Collection<V>> asMap(Multimap<K, V> multimap) {
    return multimap.asMap();
  }

  /**
   * Returns a multimap view of the specified map. The multimap is backed by the map, so changes to
   * the map are reflected in the multimap, and vice versa. If the map is modified while an
   * iteration over one of the multimap's collection views is in progress (except through the

    List<List<Integer>> partitions = partition(list, 3);

    // Changes before the partition is retrieved are reflected
    list.set(0, 3);

    Iterator<List<Integer>> iterator = partitions.iterator();

    // Changes before the partition is retrieved are reflected
    list.set(1, 4);

    List<Integer> first = iterator.next();

    // Changes after are too (unlike Iterables.partition)
    list.set(2, 5);

}

@test "API compatibility test passes, ensuring no unexpected changes to the TF API" {
    bazel test $TFCI_BAZEL_COMMON_ARGS //tensorflow/tools/api/tests:api_compatibility_test
    echo "You have to re-generate the TF API goldens and have the API changes reviewed."
    echo "Look at the instructions for ':api_compatibility_test -- --update_goldens=True'"
}

import okhttp3.internal.platform.Platform

/**
 * Example of using a hardware key to perform client auth.
 * Prefer recent JDK builds, and results are temperamental to slight environment changes.
 * Different instructions and configuration may be required for other hardware devices.
 *
 * Using a yubikey device as a SSL key store.
 * https://lauri.võsandi.com/2017/03/yubikey-for-ssh-auth.html
 *

   * flavors can read old data or data from the other flavor. However, we strongly discourage
   * relying on this, as we have made incompatible changes to serialized forms in the past and
   * expect to do so again, as discussed in https://github.com/google/guava#important-warnings.
   */
  @SuppressWarnings("unused")

     * The entrySet() of ConcurrentHashMap, unlike that of other Map
     * implementations, supports add() under JDK8. This seems problematic, but I
     * didn't see that discussed in the review, which included many other
     * changes: https://mail.openjdk.org/pipermail/core-libs-dev/2013-May/thread.html#17367
     *
     * TODO(cpovirk): decide what the best long-term action here is: force users

   * <p>Iterators from the returned map are at least <i>weakly consistent</i>: they are safe for
   * concurrent use, but if the cache is modified (including by eviction) after the iterator is
   * created, it is undefined which of the changes (if any) will be reflected in that iterator.
   */
  ConcurrentMap<K, V> asMap();

  /**
   * Performs any pending maintenance operations needed by the cache. Exactly which activities are

Doing this, our `GzipRequest` will take care of decompressing the data (if necessary) before passing it to our *path operations*.

After that, all of the processing logic is the same.

But because of our changes in `GzipRequest.body`, the request body will be automatically decompressed when it is loaded by **FastAPI** when needed.

## Accessing the request body in an exception handler { #accessing-the-request-body-in-an-exception-handler }

   * calling {@code get(columnKey, rowKey)} on the generated table always returns the same value as
   * calling {@code get(rowKey, columnKey)} on the original table. Updating the original table
   * changes the contents of the transposed table and vice versa.
   *
   * <p>The returned table supports update operations as long as the input table supports the