*/
function computeStatistics(groupedFiles, fileMetadatas) {
    let totalAdditions = 0;
    let totalDeletions = 0
    let totalChangedFiles = 0;

    let summaries = [...groupedFiles.values()];
    Object.values(summaries).forEach(summary => {
        summary.additions = 0;
        summary.deletions = 0;

        summary.files.forEach(file => {
            let fileMetadata = metadataFor(fileMetadatas, file);

EOF
}

function ips_to_cidrs() {
  IP_RANGE_START="$1"
  IP_RANGE_END="$2"
  python3 - <<EOF
from ipaddress import summarize_address_range, IPv4Address
[ print(n.compressed) for n in summarize_address_range(IPv4Address(u'$IP_RANGE_START'), IPv4Address(u'$IP_RANGE_END')) ]
EOF

  }
  *shape = &*shape_;
  return absl::OkStatus();
}

Status ParallelTensor::SummarizeValue(std::string& summary) {
  summary = "{";
  std::vector<std::string> summarized_devices = device_.SummarizeDeviceNames();
  for (int component_index = 0; component_index < tensors_.size();
       ++component_index) {
    // TODO(allenl): Add a C API for summarizing tensors. Currently custom

    return methods_.summarize != nullptr;
  }

  Status SummarizeValue(std::string& summary) const override {
    if (methods_.summarize == nullptr) {
      return tensorflow::CustomDeviceTensorHandle::SummarizeValue(summary);
    }
    TF_Status c_status;
    std::unique_ptr<TF_Buffer, decltype(&TF_DeleteBuffer)> summary_buffer(
        methods_.summarize(data_, &c_status), TF_DeleteBuffer);

  int (*num_dims)(void* data, TF_Status* status);

  // Computes the axis length at `dim_index`.
  int64_t (*dim)(void* data, int dim_index, TF_Status* status);

  void (*deallocator)(void* data);

  // Summarizes the value of this tensor. The caller takes ownership of the
  // returned buffer. If `status` is not TF_OK, instead returns a null pointer.
  //
  // Does not include the shape and dtype of the tensor (which is generally

 * forms of variance -- that were computed from a single set of zero or more floating-point values.
 *
 * <p>There are two ways to obtain a {@code Stats} instance:
 *
 * <ul>
 *   <li>If all the values you want to summarize are already known, use the appropriate {@code
 *       Stats.of} factory method below. Primitive arrays, iterables and iterators of any kind of
 *       {@code Number}, and primitive varargs are supported.

But all this functionality of using asynchronous code with `async` and `await` is many times summarized as using "coroutines". It is comparable to the main key feature of Go, the "Goroutines".

## Conclusion

Let's see the same phrase from above:

        *   `setNumThreads` - Use `Interpreter.Options.setNumThreads`
    *   Add Conv3DTranspose as a builtin op.
*   `tf.summary`:
    *   Fix `tf.summary.should_record_summaries()` so it correctly reflects when
        summaries will be written, even when `tf.summary.record_if()` is not n
        effect, by returning True tensor if default writer is present.
*   Grappler: