olorenchemengine.external.SPGNN package#

Submodules#

olorenchemengine.external.SPGNN.main module#

Wraps the model presented in Strategies for pre-training graph neural networks

GitHub repository

class olorenchemengine.external.SPGNN.main.SPGNN(model_type='contextpred', batch_size=32, epochs=100, lr=0.001, lr_scale=1, decay=0, num_layer=5, emb_dim=300, dropout_ratio=0.5, graph_pooling='mean', JK='last', gnn_type='gin', **kwargs)#

Bases: BaseModel

SPGNN is the model presented in Strategies for pre-training graph neural networks GitHub repository

available_pretrained_models#

List of available pretrained models; passed in the model_type parameter

Type:

List[str]

Parameters:

model_type (str) – Type of model to use; default: “contextpred”

classmethod AllInstances()#

AllTypes returns a list of all standard instances of all subclasses of BaseClass.

Standard instances means that all required parameters for instantiation of the subclasses are set with canonical values.

available_pretrained_models = ['contextpred', 'edgepred', 'infomax', 'masking', 'supervised_contextpred', 'supervised_edgepred', 'supervised_infomax', 'supervised_masking', 'supervised', 'gat_supervised_contextpred', 'gat_supervised', 'gat_contextpred']#
preprocess(X, y, **kwargs)#
Parameters:

X (list of smiles) –

Returns:

Processed list converted into whatever input for the model

class olorenchemengine.external.SPGNN.main.SPGNNVecRep(model_type='contextpred', batch_size=32, epochs=100, lr=0.001, lr_scale=1, decay=0, num_layer=5, emb_dim=300, dropout_ratio=0.5, graph_pooling='mean', JK='last', gnn_type='gin', **kwargs)#

Bases: BaseVecRepresentation

SPGNN_REP gives the output of the model presented in Strategies for pre-training graph neural networks GitHub repository <https://github.com/snap-stanford/pretrain-gnns> as a molecular representation.

available_pretrained_models#

List of available pretrained models; passed in the model_type parameter

Type:

List[str]

Parameters:

model_type (str) – Type of model to use; default: “contextpred”

available_pretrained_models = ['contextpred', 'edgepred', 'infomax', 'masking', 'supervised_contextpred', 'supervised_edgepred', 'supervised_infomax', 'supervised_masking', 'supervised', 'gat_supervised_contextpred', 'gat_supervised', 'gat_contextpred']#
convert(smiles, **kwargs)#

BaseVecRepresentation’s convert returns a list of numpy arrays.

Parameters:

  • Xs (Union[list, pd.DataFrame, dict, str]) – input data

  • ys (Union[list, pd.Series, np.ndarray], optional) – included for compatibility, unused argument. Defaults to None.

Returns:

list of molecular vector representations

Return type:

List[np.ndarray]

class olorenchemengine.external.SPGNN.main.SPGNN_AF(log=True)#

Bases: AtomFeaturizer

convert(atom: _MockObject.Chem.Atom)#
property length#
class olorenchemengine.external.SPGNN.main.SPGNN_BF(log=True)#

Bases: BondFeaturizer

convert(bond: _MockObject.Chem.Bond)#
property length#
class olorenchemengine.external.SPGNN.main.SPGNN_PYG#

Bases: TorchGeometricGraph

olorenchemengine.external.SPGNN.main.predict(model, device, loader, setting='classification')#
olorenchemengine.external.SPGNN.main.train(model, device, loader, optimizer, setting)#

olorenchemengine.external.SPGNN.model module#

Defines the model architecture of SPGNN.

Module contents#