GraphData

class caldera.data.GraphData(node_attr, edge_attr, global_attr, edges, requires_grad=None)[source]

Data representing a single graph.

Parameters

  • node_attr (FloatTensor) –

  • edge_attr (FloatTensor) –

  • global_attr (FloatTensor) –

  • edges (LongTensor) –

  • requires_grad (Optional[bool]) –

Blank.

Parameters

  • node_attr (FloatTensor) –

  • edge_attr (FloatTensor) –

  • global_attr (FloatTensor) –

  • edges (LongTensor) –

  • requires_grad (Optional[bool]) –

__init__(node_attr, edge_attr, global_attr, edges, requires_grad=None)[source]

Blank.

Parameters

  • node_attr (FloatTensor) –

  • edge_attr (FloatTensor) –

  • global_attr (FloatTensor) –

  • edges (LongTensor) –

  • requires_grad (Optional[bool]) –

Methods

__init__(node_attr, edge_attr, global_attr, …)

Blank.

allclose(other, **kwargs)

param other

append_edges(edge_attr, edges)

param edge_attr

append_nodes(node_attr)

param node_attr

apply(func, *args[, keys])

Applies the function to the data, creating a new instance of GraphData.

apply_(func, *args[, keys])

Applies the function in place to the data, wihout creating a new instance of GraphData.

apply_edge_mask(mask)

Apply edge mask to the graph, returning a new GraphData instance.

apply_edge_mask_(mask)

In place version of caldera.data.GraphData.apply_edge_mask()

apply_node_mask(node_mask)

Apply node mask to the graph, returning a new GraphData instance, removing any edges if necessary.

apply_node_mask_(node_mask)

In place version of caldera.data.GraphData.apply_node_mask().

clone()

Clones the data.

contiguous()

copy([non_blocking])

non_blocking (bool) – if True and this copy is between CPU and GPU, the copy may occur asynchronously with respect to the host.

debug()

density()

Return density of the graph.

detach()

from_networkx(g[, n_node_feat, n_edge_feat, …])

Create a new GraphData from a networkx graph.

index_edges(idx)

Apply index to nodes.

index_edges_(idx)

In place version of caldera.data.GraphData.index_edges()

index_nodes(idx)

Apply index to nodes.

index_nodes_(idx)

In place version of caldera.data.GraphData.index_nodes()

info()

memsize()

Return total number of bytes in the.

nelement()

Return total number of elements in the.

random(n_feat, e_feat, g_feat[, …])

param n_feat

reverse()

rtype

GraphData

reverse_()

share_storage(other[, return_dict])

Check if this data shares storage with another data.

shuffle()

rtype

GraphData

shuffle_()

rtype

None

shuffle_edges()

rtype

GraphData

shuffle_edges_()

rtype

None

shuffle_nodes()

rtype

GraphData

shuffle_nodes_()

rtype

None

to(device, *args, **kwargs)

param device

to_networkx([feature_key, global_attr_key, …])

param feature_key

view([x_slice, e_slice, g_slice])

param x_slice

Attributes

e

edge_shape

edges

g

global_shape

node_shape

num_edges

num_graphs

num_nodes

requires_grad

shape

size

x

__init__(node_attr, edge_attr, global_attr, edges, requires_grad=None)[source]

Blank.

Parameters

  • node_attr (FloatTensor) –

  • edge_attr (FloatTensor) –

  • global_attr (FloatTensor) –

  • edges (LongTensor) –

  • requires_grad (Optional[bool]) –