New comments and cleanup

    TODO(rbharath, joegomes): It is unclear whether this should be a Dataset

    function. Might get refactored out.

    TODO(rbharath, joegomes): If y_n were an exposed part of the API, this

    function could be entirely written in userspace.

    """

    if len(self) == 0:

      return None, None, None, None

    N = mol.GetNumAtoms()

    coords = np.zeros((N,3))

    # TODO(joegomes, rbharath): Add comment about magic number

    # RDKit stores atomic coordinates in Angstrom. Atomic unit of length is the

    # bohr (1 bohr = 0.529177 Angstrom). Converting units makes gradient calculation

    # consistent with most QM software packages.

    coords_in_bohr = [mol.GetConformer(0).GetAtomPosition(i).__div__(0.52917721092)

                      for i in xrange(N)]

  """Wrapper class for computing user-defined metrics."""

  def __init__(self, metric, task_averager=None, name=None, threshold=None,

               verbosity=None, mode=None, compute_force_metrics=False):

               verbosity=None, mode=None, compute_energy_metric=False):

    """

    Args:

      metric: function that takes args y_true, y_pred (in that order) and

        raise ValueError("Must specify mode for new metric.")

    assert mode in ["classification", "regression"]

    self.mode = mode

    self.compute_force_metrics = compute_force_metrics 

    # The convention used is that the first task is the metric.

    # TODO(rbharath, joegomes): This doesn't seem like it should be hard-coded as

    # an option in the Metric class. Instead, this should be possible to move into

    # user-space as a custom task_averager function.

    self.compute_energy_metric = compute_energy_metric

  def compute_metric(self, y_true, y_pred, w=None, n_classes=2, filter_nans=True):

    """Compute a performance metric for each task.

      if filter_nans:

        computed_metrics = np.array(computed_metrics)

        computed_metrics = computed_metrics[~np.isnan(computed_metrics)]

      if self.compute_force_metrics:

      if self.compute_energy_metric:

        # TODO(rbharath, joegomes): What is this magic number?

        force_error = self.task_averager(computed_metrics[1:])*4961.47596096

        print("Force error (metric: np.mean(%s)): %f kJ/mol/A" % (self.name, force_error))

        return computed_metrics[0]

from deepchem.utils.save import log

from deepchem.models import Model

import sklearn

######################################################## DEBUG

from deepchem.transformers import undo_transforms

######################################################## DEBUG

class SingletaskToMultitask(Model):

  """

            verbosity=self.verbosity)

        task_model.reload()

      #y_pred[:, ind] = task_model.predict(dataset, transformers)

      y_pred[:, ind] = task_model.predict(dataset, [])

    y_pred = undo_transforms(y_pred, transformers)

    return y_pred