Source code for torch_geometric.utils.subgraph

import torch

from .num_nodes import maybe_num_nodes


[docs]def subgraph(subset, edge_index, edge_attr=None, relabel_nodes=False, num_nodes=None): r"""Returns the induced subgraph of :obj:`(edge_index, edge_attr)` containing the nodes in :obj:`subset`. Args: subset (LongTensor, BoolTensor or [int]): The nodes to keep. edge_index (LongTensor): The edge indices. edge_attr (Tensor, optional): Edge weights or multi-dimensional edge features. (default: :obj:`None`) relabel_nodes (bool, optional): If set to :obj:`True`, the resulting :obj:`edge_index` will be relabeled to hold consecutive indices starting from zero. (default: :obj:`False`) num_nodes (int, optional): The number of nodes, *i.e.* :obj:`max_val + 1` of :attr:`edge_index`. (default: :obj:`None`) :rtype: (:class:`LongTensor`, :class:`Tensor`) """ device = edge_index.device if isinstance(subset, list) or isinstance(subset, tuple): subset = torch.tensor(subset, dtype=torch.long) if subset.dtype == torch.bool or subset.dtype == torch.uint8: n_mask = subset if relabel_nodes: n_idx = torch.zeros(n_mask.size(0), dtype=torch.long, device=device) n_idx[subset] = torch.arange(subset.sum().item(), device=device) else: num_nodes = maybe_num_nodes(edge_index, num_nodes) n_mask = torch.zeros(num_nodes, dtype=torch.bool) n_mask[subset] = 1 if relabel_nodes: n_idx = torch.zeros(num_nodes, dtype=torch.long, device=device) n_idx[subset] = torch.arange(subset.size(0), device=device) mask = n_mask[edge_index[0]] & n_mask[edge_index[1]] edge_index = edge_index[:, mask] edge_attr = edge_attr[mask] if edge_attr is not None else None if relabel_nodes: edge_index = n_idx[edge_index] return edge_index, edge_attr