GOOD.utils.data

Some data process utils including construction of molecule PyG graph from smile (for compatibility).

Functions

batch_input(G, batch_size[, num_nodes, ...])	Repeat a graph batch_size times and pack into a Batch.
from_smiles(smiles[, with_hydrogen, kekulize])	Converts a SMILES string to a torch_geometric.data.data.Data instance.

GOOD.utils.data.batch_input(G: Data, batch_size: int, num_nodes: Optional[int] = None, node_attrs: list = ['color'])

Repeat a graph batch_size times and pack into a Batch.

Parameters

• G (Data) – The given graph G.

• batch_size (int) – Batch size.

• num_nodes (int) – The number of node of the graph. If None, it will use maybe_numb_nodes.

• node_attrs (list) – The preserved node attributes.

Returns

Repeated graph batch.

GOOD.utils.data.from_smiles(smiles: str, with_hydrogen: bool = False, kekulize: bool = False)

Converts a SMILES string to a torch_geometric.data.data.Data instance.

Parameters

• smiles (string, optional) – The SMILES string.

• with_hydrogen (bool, optional) – If set to True, will store hydrogens in the molecule graph. (default: False)

• kekulize (bool, optional) – If set to True, converts aromatic bonds to single/double bonds. (default: False)