Method _write

tensorflow_datasets/core/writer_test.py:257–294

(
      self,
      to_write,
      salt: str = '',
      dataset_name: str = 'foo',
      split: str = 'train',
      disable_shuffling: bool = False,
      ignore_duplicates: bool = False,
      example_writer: Optional[writer_lib.ExampleWriter] = None,
      shard_config: Optional[shard_utils.ShardConfig] = None,
  )

  ]

  def _write(
      self,
      to_write,
      salt: str = '',
      dataset_name: str = 'foo',
      split: str = 'train',
      disable_shuffling: bool = False,
      ignore_duplicates: bool = False,
      example_writer: Optional[writer_lib.ExampleWriter] = None,
      shard_config: Optional[shard_utils.ShardConfig] = None,
  ):
    example_writer = example_writer or writer_lib.ExampleWriter(
        file_format=file_adapters.DEFAULT_FILE_FORMAT
    )
    filetype_suffix = file_adapters.ADAPTER_FOR_FORMAT[
        example_writer.file_format
    ].FILE_SUFFIX
    filename_template = naming.ShardedFileTemplate(
        dataset_name=dataset_name,
        split=split,
        filetype_suffix=filetype_suffix,
        data_dir=self.tmp_dir,
    )
    shard_config = shard_config or shard_utils.ShardConfig(
        num_shards=self.NUM_SHARDS
    )
    writer = writer_lib.Writer(
        serializer=testing.DummySerializer('dummy specs'),
        filename_template=filename_template,
        hash_salt=salt,
        disable_shuffling=disable_shuffling,
        example_writer=example_writer,
        shard_config=shard_config,
        ignore_duplicates=ignore_duplicates,
    )
    for key, record in to_write:
      writer.write(key, record)
    return writer.finalize()

  def test_write_tfrecord(self):
    """Stores records as tfrecord in a fixed number of shards with shuffling."""