1. pahelix.featurizers

1.1. het_gnn_featurizer

Featurizers for DDI Heterogenous graph.
class pahelix.featurizers.het_gnn_featurizer.DDiFeaturizer[source]

Featurizer for drugs

collate_fn(ddi_data, dti_data, ppi_data, features)[source]

Aggregate all needed nodes into a Hetrogenous graph

pahelix.featurizers.het_gnn_featurizer.num_nodes_stat(data)[source]

count the number of nodes from data

Examples

data: {‘pair’: (a, b)}

pahelix.featurizers.het_gnn_featurizer.nx_graph_build(hg, nodes_dict, label)[source]

Build Heterogenous graph with node name not idx.

1.2. pretrain_gnn_featurizer

Featurizers for pretrain-gnn.
Adapted from https://github.com/snap-stanford/pretrain-gnns/tree/master/chem/utils.py
class pahelix.featurizers.pretrain_gnn_featurizer.AttrmaskTransformFn[source]

Gen features for attribute mask model of pretrain gnns

class pahelix.featurizers.pretrain_gnn_featurizer.AttrmaskCollateFn(atom_names, bond_names, mask_ratio=0.15)[source]

CollateFn for attribute mask model of pretrain gnns

class pahelix.featurizers.pretrain_gnn_featurizer.SupervisedTransformFn[source]

Gen features for supervised model of pretrain gnns

class pahelix.featurizers.pretrain_gnn_featurizer.SupervisedCollateFn(atom_names, bond_names)[source]

CollateFn for supervised model of pretrain gnns