dgllife.utils.PAGTNAtomFeaturizer

class dgllife.utils.PAGTNAtomFeaturizer(atom_data_field='h')

The atom featurizer used in PAGTN

PAGTN is introduced in Path-Augmented Graph Transformer Network.

The atom features include:

• One hot encoding of the atom type.

• One hot encoding of formal charge of the atom.

• One hot encoding of the atom degree

• One hot encoding of explicit valence of an atom. The supported possibilities include 0 - 6.

• One hot encoding of implicit valence of an atom. The supported possibilities include 0 - 5.

• Whether the atom is aromatic.

Parameters

atom_data_field (str) – Name for storing atom features in DGLGraphs, default to ‘h’.

Examples

>>> from rdkit import Chem
>>> from dgllife.utils import PAGTNAtomFeaturizer

>>> mol = Chem.MolFromSmiles('C')
>>> atom_featurizer = PAGTNAtomFeaturizer(atom_data_field='feat')
>>> atom_featurizer(mol)
{'feat': tensor([[1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,
                  0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,
                  0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,
                  0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 1., 0., 0.,
                  0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0.,
                  0., 1., 0., 0.]])}
>>> # Get feature size for nodes
>>> print(atom_featurizer.feat_size())
94

See also

BaseAtomFeaturizer, CanonicalAtomFeaturizer, PretrainAtomFeaturizer, WeaveAtomFeaturizer, AttentiveFPAtomFeaturizer

__init__(atom_data_field='h')

Initialize self. See help(type(self)) for accurate signature.

Methods

__init__([atom_data_field])	Initialize self.
feat_size([feat_name])	Get the feature size for feat_name.