mlx_graphs.utils.transformations.add_self_loops

Contents

mlx_graphs.utils.transformations.add_self_loops#

mlx_graphs.utils.transformations.add_self_loops(edge_index: mlx.core.array, edge_features: mlx.core.array | None = None, num_nodes: int | None = None, fill_value: float | mlx.core.array | None = 1, allow_repeated: bool | None = True) → mlx.core.array | tuple[mlx.core.array, mlx.core.array][source]#

Adds self-loops to the given graph represented by edge_index and edge_features.

Parameters:

  • edge_index (array) – a [2, num_edges] array representing the source and target nodes of each edge

  • edge_features (Optional[array]) – Optional tensor representing features associated with each edge, with shape [num_edges, num_edge_features]

  • num_nodes (Optional[int]) – Optional number of nodes in the graph. If not provided, it is inferred from edge_index.

  • fill_value (Union[float, array, None]) – Value used for filling the self-loop features. Default is 1.

  • allow_repeated (Optional[bool]) – Specify whether to add self-loops for all nodes, even if they are already in the edge_index. Defaults to True.

Return type:

Union[array, tuple[array, array]]

Returns:

A tuple containing the updated edge_index and edge_features with self-loops.

Contents