Cleanup.

    None or "log". Only for regression outputs.

  """

  models = {}

  print "fit_singletask_models()"

  print "train_data.keys()"

  print train_data.keys()

  import numpy as np

  X_train = train_data["features"]

  print "np.shape(X_train)"

  print np.shape(X_train)

  sorted_tasks = train_data["sorted_tasks"]

  for task in sorted_tasks:

    print "Building model for task %s" % task

  feature_dir = os.path.join(dataset_dir, featuretype)

  features_file = os.path.join(feature_dir, "%s-%s.pkl.gz" % (name, featuretype))

  print("About to instantiate featurizer.")

  if featuretype == "fingerprints":

    featurizer = CircularFingerprint(size=1024)

  elif featuretype == "descriptors":

      print("Featurizing molecule %d" % row_ind)

    mol = Chem.MolFromSmiles(row_data)

    features.append(featurizer.featurize([mol]))

  print("Done generating features. About to transfer them to dataframe.")

  feature_df = pd.DataFrame([])

  feature_df["features"] = pd.DataFrame(

      [{"features": feature} for feature in features])

  print("Done transfering to dataframe. About to populate remaining df fields.")

  feature_df["smiles"] = dataframe[[smiles_field]]

  feature_df["scaffolds"] = dataframe[[smiles_field]].apply(

      functools.partial(generate_scaffold, smiles_field=smiles_field),

      axis=1)

  feature_df["mol_id"] = dataframe[[id_field]]

  print("Populated 'smiles', 'scaffolds', 'mol_id' fields")

  print("About to write pkl.gz file")

  with gzip.open(features_file, "wb") as gzip_file:

                     splittype="random", target_names=None):

  """Extracts datasets and split into train/test.

  Returns a dict with the following keys

  Returns a dict with the following key/value pairs

  "features" -> X

  "mol_ids"  -> ids

  features -> X

  mol_ids  -> ids

  target -> (y, W)

  "sorted_targets" -> sorted_targets

  sorted_targets -> sorted_targets

  Parameters

  ----------

      scaffolds[scaffold].append(mol_id)

  # Sort from largest to smallest scaffold sets

  return [elt for (scaffold, elt) in sorted(scaffolds.items(), key=lambda x: -len(x[1]))]

#def labels_to_weights(ytrue):

#  """Uses the true labels to compute and output sample weights.

#

#  Parameters

#  ----------

#  ytrue: list or np.ndarray

#    True labels.

#  """

#  n_total = np.shape(ytrue)[0]

#  n_positives = np.sum(ytrue)

#  n_negatives = n_total - n_positives

#  pos_weight = np.floor(n_negatives/n_positives)

#

#  sample_weights = np.zeros(np.shape(ytrue)[0])

#  for ind, entry in enumerate(ytrue):

#    if entry == 0:  # negative

#      sample_weights[ind] = 1

#    elif entry == 1:  # positive

#      sample_weights[ind] = pos_weight

#    else:

#      print("labels_to_weights()")

#      print("ytrue")

#      print(ytrue)

#      raise ValueError("ytrue can only contain 0s or 1s.")

#  return sample_weights