dgmc.utils

class dgmc.utils.PairDataset(dataset_s, dataset_t, sample=False)[source]

Combines two datasets, a source dataset and a target dataset, by building pairs between separate dataset examples.

Parameters
  • dataset_s (torch.utils.data.Dataset) – The source dataset.

  • dataset_t (torch.utils.data.Dataset) – The target dataset.

  • sample (bool, optional) – If set to True, will sample exactly one target example for every source example instead of holding the product of all source and target examples. (default: False)

class dgmc.utils.ValidPairDataset(dataset_s, dataset_t, sample=False)[source]

Combines two datasets, a source dataset and a target dataset, by building valid pairs between separate dataset examples. A pair is valid if each node class in the source graph also exists in the target graph.

Parameters
  • dataset_s (torch.utils.data.Dataset) – The source dataset.

  • dataset_t (torch.utils.data.Dataset) – The target dataset.

  • sample (bool, optional) – If set to True, will sample exactly one target example for every source example instead of holding the product of all source and target examples. (default: False)