cb7c20f6ad714f0b016a874b2b06a9fbed4e0a59,tensorflow_transform/test_case.py,TransformTestCase,AssertVocabularyContents,#TransformTestCase#Any#Any#,290

Before Change

      raise

  def AssertVocabularyContents(self, vocab_file_path, file_contents):
    with tf.io.gfile.GFile(vocab_file_path, "rb") as f:
      file_lines = f.readlines()

      // Store frequency case.
      if isinstance(file_contents[0], tuple):
        word_and_frequency_list = []
        for content in file_lines:
          frequency, word = content.split(b" ", 1)
          // Split by comma for when the vocabulary file stores the result of
          // per-key analyzers.
          values = list(map(float, frequency.strip(b"\n").split(b",")))
          word_and_frequency_list.append(
              (word.strip(b"\n"), values[0] if len(values) == 1 else values))

        expected_words, expected_frequency = zip(*word_and_frequency_list)
        actual_words, actual_frequency = zip(*file_contents)
        self.assertAllEqual(expected_words, actual_words)
        np.testing.assert_almost_equal(expected_frequency, actual_frequency)
      else:
        file_lines = [content.strip(b"\n") for content in file_lines]
        self.assertAllEqual(file_lines, file_contents)

  def WriteRenderedDotFile(self, dot_string, output_file=None):
    tf.compat.v1.logging.info(
        "Writing a rendered dot file is not yet supported.")

After Change

      raise

  def AssertVocabularyContents(self, vocab_file_path, file_contents):
    if vocab_file_path.endswith(".tfrecord.gz"):
      file_lines = list(
          tf_utils.read_tfrecord_vocabulary_dataset(
              vocab_file_path).as_numpy_iterator())
    else:
      with tf.io.gfile.GFile(vocab_file_path, "rb") as f:
        file_lines = f.read().splitlines()

    // Store frequency case.
    if isinstance(file_contents[0], tuple):
      word_and_frequency_list = []
      for content in file_lines:
        frequency, word = content.split(b" ", 1)

Instances