added nucleic acid and amino acid charsets

  learning tasks.

  """

  def __init__(self, featurizer: Featurizer = OneHotFeaturizer()):

  def __init__(self, featurizer: Featurizer = OneHotFeaturizer, protein = True):

    """Initialize FASTALoader.

    Parameters

    ----------

    featurizer: Featurizer (default: OneHotFeaturizer())

    featurizer: Featurizer (default: OneHotFeaturizer)

      The Featurizer to be used for the loaded FASTA data.

    protein: bool (default: True)

      Whether or not the sequence passed in is a protein sequence. If False,

      it is treated as a nucleic acid sequence.

    """

    protein_charset = ('A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J', 'K',

                       'L', 'M', 'N', 'O', 'P', 'Q', 'R', 'S', 'T', 'U', 'V',

                       'W', 'X', 'Y', 'Z', '*', '-')

    nucleic_charset = ('A', 'C', 'G', 'T', 'U', '(i)', 'R', 'Y', 'K', 'M', 'S',

                       'W', 'B', 'D', 'H', 'V', 'N', '-')

    self.user_specified_features = None

    if isinstance(featurizer, UserDefinedFeaturizer):

      self.user_specified_features = featurizer.feature_fields

    if (featurizer == OneHotFeaturizer):

      if (protein):

        featurizer = OneHotFeaturizer(protein_charset)

      else:

        featurizer = OneHotFeaturizer(nucleic_charset)

    self.featurizer = featurizer

  def create_dataset(self,

      sequences = np.array([])

      for input_file in input_files:

        sequences = np.append(sequences, _read_file(input_file))

      logger.warning(f"**TESTING** FEATURIZING ARRAY: {sequences}")

      X = self.featurizer(sequences)

      logger.warning(f"**TESTING** FINAL FEATURIZED ARRAY: {X}")

      ids = np.ones(len(X))

      # (X, y, w, ids)

      yield X, None, None, ids # TODO discuss shape

      # (X, y, w, ids) TODO discuss shape

      yield X, None, None, ids 

    def _read_file(input_file: str, auto_add_annotations: bool=False):

      """