grapharray.classes module

Graph Variables

BaseGraph([incoming_graph_data])	Directed graph object on which arrays are defined.
BaseGraphArray(base_graph)	Base object for creating vectors and matrices on networks.
GraphArray(base_graph[, init_val, is_array_2d])	Extracted codes shared between NodeArray and EdgeArray.
NodeArray(base_graph[, init_val, is_array_2d])	Object of variables defined on the nodes.
EdgeArray(base_graph[, init_val, is_array_2d])	Object of variables defined on the edges.
AdjacencyMatrix(weight[, sparse_format])	N x N matrix
IncidenceMatrix(base_graph)	Node-edge incidence matrix

class grapharray.classes.AdjacencyMatrix(weight: grapharray.classes.EdgeArray, sparse_format: str = 'csr')

Bases: grapharray.classes.BaseGraphArray

N x N matrix

class grapharray.classes.BaseGraph(incoming_graph_data=None, **attr)

Bases: networkx.classes.digraph.DiGraph

Directed graph object on which arrays are defined.

This object works (1) as a key to check if variables are defined on the same network and (2) as a holder of the orders of nodes and edges. Note that this object itself does not contain any variables.

property edge_to_index

Correspondence between edges and array indices

freeze()

Freeze the graph and map between nodes / edges and array indices

This method must be called before the instance is passed to array initialization methods.

property node_to_index

Correspondence between nodes and array indices

class grapharray.classes.BaseGraphArray(base_graph: grapharray.classes.BaseGraph)

Bases: object

Base object for creating vectors and matrices on networks.

This set up attributes used in both node variables and edge variables. All attributes are aliases of that of BaseGraph and thus are read-only.

property T

Transpose the array

property array

Core array

property base_graph

BaseGraph object on that this array itself is defined

property edge_to_index

Correspondence between edges and array indices

property edges

Tuple of all edges

property is_transposed

Whether the array is transposed or not.

property node_to_index

Correspondence between nodes and array indices

property nodes

Tuple of all nodes

property number_of_edges

The number of edges in the base graph

property number_of_nodes

The number of nodes in the base graph

class grapharray.classes.EdgeArray(base_graph: grapharray.classes.BaseGraph, init_val=0, is_array_2d: bool = False)

Bases: grapharray.classes.GraphArray

Object of variables defined on the edges.

as_nx_graph()

Return a nx.DiGraph with the array elements as its edge attributes.

property index

Correspondence between the array indices and the edges.

class grapharray.classes.GraphArray(base_graph: grapharray.classes.BaseGraph, init_val=0, is_array_2d: bool = False)

Bases: grapharray.classes.BaseGraphArray

Extracted codes shared between NodeArray and EdgeArray.

Parameters

• base_graph (BaseGraph) – The graph on that the variable is defined.

• init_val – The initial value of the array.

• is_array_2d (bool) – Whether the array is 2-dimensional column vectors. Default is False, which means that the array is 1-dimensional.

Notes

init_val must be either scalar, NodeVar object or {node: value} dictionary. if a scalar is given, all elements of array are set to the init_val. if a NodeVar object is given, a copy of its array is used as initial values. if a dictionary is given, the value on each node/edge is used as initial value of corresponding node/node. if np.ndarray is given, it is directly used as the initial values. This is used to make arithmetic operations faster by avoiding unnecessary array creation.

array

An array that has values linked to nodes/edges.

Type

np.ndarray

as_dict() → dict

Return values of variables as a dictionary keyed by node/edge.

as_nx_graph(assign_to=None)

Return a nx.DiGraph with the array elements as node/edge attributes.

get_copy()

Make a copy of self.

The array of the copy created by this method is a copy of the original, while the base_graph of the copy is the same instance of the original. This is different from the copy created by copy.deepcopy() in that both the array and the base_graph is a copy of the original.

property index

Correspondence between the array indices and the nodes/edges.

This is only a dummy implementation here and overridden in subclasses.

property is_2d

Whether the array is 2-dimensional or not

class grapharray.classes.IncidenceMatrix(base_graph: grapharray.classes.BaseGraph)

Bases: grapharray.classes.BaseGraphArray

Node-edge incidence matrix

class grapharray.classes.NodeArray(base_graph: grapharray.classes.BaseGraph, init_val=0, is_array_2d: bool = False)

Bases: grapharray.classes.GraphArray

Object of variables defined on the nodes.

as_nx_graph()

Return a nx.DiGraph with the array elements as its node attributes.

property index

Correspondence between the array indices and the nodes.