var first = true
        list.forEach {
            if (first) first = false else comma()
            body(it)
        }
        endArray()
    }

    private
    fun beginArray() {
        write('[')
    }

    private
    fun endArray() {
        write(']')
    }

    private
    fun property(name: String, value: String) {
        property(name) { jsonString(value) }

        reader.beginArray();
        while (reader.peek() != JsonToken.END_ARRAY) {
            consumeVariant(reader, metadata);
        }
        reader.endArray();
    }

    private void consumeVariant(JsonReader reader, MutableModuleComponentResolveMetadata metadata) throws IOException {
        String variantName = null;

            out.name("subcategories").beginArray();
            for (String sc : value.getSubcategories()) {
                out.value(sc);
            }
            out.endArray();
            out.endObject();
        }

        @Override
        public ProblemCategory read(JsonReader in) throws IOException {
            in.beginObject();
            String namespace = null;

    private List<Object> readArray(JsonReader reader) {
        List<Object> values = []
        reader.beginArray()
        while (reader.peek() != JsonToken.END_ARRAY) {
            Object value = readAny(reader)
            values.add(value)
        }
        reader.endArray()
        return values
    }

    static Map<String, String> normalizeForTests(Map<String, ?> attributes) {

from tensorflow.python.types import core
from tensorflow.python.util import tf_export

# A representative sample is a map of: input_key -> input_value.
# Ex.: {'dense_input': tf.constant([1, 2, 3])}
# Ex.: {'x1': np.ndarray([4, 5, 6]}
RepresentativeSample = Mapping[str, core.TensorLike]

# A representative dataset is an iterable of representative samples.
RepresentativeDataset = Iterable[RepresentativeSample]

                        break;
                    case "groups":
                        final List<String> list = new ArrayList<>();
                        while (jsonParser.nextToken() != JsonToken.END_ARRAY) {
                            final String group = jsonParser.getText();
                            list.add(group);
                        }
                        if (logger.isDebugEnabled()) {

          sample, signature_def
      )
      input_tensor_data = input_tensor.eval()

    self.assertLen(feed_dict, 1)
    self.assertIn('input:0', feed_dict)
    self.assertIsInstance(feed_dict['input:0'], np.ndarray)
    self.assertAllEqual(feed_dict['input:0'], input_tensor_data)

  @test_util.deprecated_graph_mode_only
  def test_create_feed_dict_from_input_data_empty(self):
    signature_def = meta_graph_pb2.SignatureDef(

    self._hist_mids = np.linspace(first_mid, last_mid, self._num_bins)

  def _get_dequantized_hist_mids_after_quantize(
      self, quant_min: float, quant_max: float
  ) -> np.ndarray:
    """Quantizes and dequantizes hist_mids using quant_min and quant_max.

    Quantization converts the range of numbers from [quant_min, quant_max] to
    [0, 2^num_bits - 1]. Values less than quant_min are converted to 0, and

    [`tf.experimental.numpy`](https://www.tensorflow.org/api_docs/python/tf/experimental/numpy),
    which is a NumPy-compatible API for writing TF programs. This module
    provides class `ndarray`, which mimics the `ndarray` class in NumPy, and
    wraps an immutable `tf.Tensor` under the hood. A subset of NumPy functions
    (e.g. `numpy.add`) are provided. Their inter-operation with TF facilities is
    seamless in most cases. See