numgraph.utils

Utility functions

numgraph.utils.coalesce(edge_list)[source]

Polishes an edge list by removing duplicates and by sorting the edges

Parameters: edge_list (NDArray) – An edge list (num_edges x 2)
Returns: NDArray – A sorted edge list with no duplicated edges (new_num_edges x 2)

numgraph.utils.dense(generator)[source]

Transforms a sparse generator into its dense version

Parameters: generator (Callable) – A callable that generates graphs
Returns: Callable – A callable that generates the squared adjacency matrix (num_nodes x num_nodes) of a graph

numgraph.utils.to_dense(edge_list, num_nodes=None)[source]

Converts a list of edges in a squared adjacency matrix

Parameters

edge_list (NDArray) – The list of edges (num_edges x 2)
num_nodes (int, optional) – The number of nodes in the graph, by default None

Returns

NDArray – The squared adjacency matrix (num_nodes x num_nodes)

numgraph.utils.to_sparse(adj_matrix)[source]

Converts an adjacency matrix to a list of edges

Parameters: adj_matrix (NDArray) – The squared adjacency matrix (num_nodes x num_nodes)
Returns: NDArray – The list of edges (num_edges x 2)

numgraph.utils.to_undirected(adj)[source]

Turns a directed edge_list into a non-directed one

Parameters: adj (NDArray) – A directed adjacency matrix (num_nodes x num_nodes), or edge list (num_edges x 2)
Returns: NDArray – An undirected adjacency matrix (num_nodes x num_nodes), or the edge list ((2*num_edges) x 2)

numgraph.utils.remove_self_loops(adj)[source]

Removes every self-loop in the graph given by adj

Parameters: adj (NDArray) – The adjancency matrix (num_nodes x num_nodes), or the edge_list (num_edges x 2)
Returns: NDAarray – The adjacency matrix (num_nodes x num_nodes), or the list of edges (new_num_edges x 2), without self-loops.

numgraph.utils.unsorted_coalesce(edge_list, weights=None)[source]

Polishes an edge list by removing duplicates and by sorting the edges

Parameters

edge_list (NDArray) – An edge list (num_edges x 2)
weights (NDArray) – The weights (num_edges x 1)

Returns

NDArray – An unsorted edge list with no duplicated edges (new_num_edges x 2)
NDArray – The unsorted weigths associated to the new edge list (new_num_edges x 1)

Spikes Generator

class numgraph.utils.spikes_generator.SpikeGenerator(t_max=10, heat_spike=(0.7, 2.0), num_spikes=1, rng=None)[source]

Bases: object

The generator of the spikes in the heat diffusion over a graph. To each spike is associated an increase of temperature of the node. This class identify at each step the node with injected temperature and the new temperature.

Parameters

t_max (int, optional) – The maximum number of timesteps in the simulation, by default 10
heat_spike (Tuple[float, float], optional) – A tuple containing the min and max temperature of a spike, by default (0.7, 2.)
num_spikes (int, optional) – The number of heat spikes during the process, by default 1
rng (Generator, optional) – Numpy random number generator, by default None

compute_spike(t, x)[source]

Computes the evolution of node temperature given a timestep t

Parameters

t (int) – The timesteps
x (NDArray) – The vector of node temperatures of shape (num_nodes x 1)

Returns

NDArray – The vector of new node temperatures of shape (num_nodes x 1)

class numgraph.utils.spikes_generator.HeatSpikeGenerator(t_max=10, heat_spike=(0.7, 2.0), num_spikes=1, rng=None)[source]

Bases: numgraph.utils.spikes_generator.SpikeGenerator

The generator of the spikes in the heat diffusion over a graph. To each spike is associated an increase of temperature of the node. This class identify at each step the node with injected temperature and the new temperature.

Parameters

t_max (int, optional) – The maximum number of timesteps in the simulation, by default 10
heat_spike (Tuple[float, float], optional) – A tuple containing the min and max temperature of a spike, by default (0.7, 2.)
num_spikes (int, optional) – The number of heat spikes during the process, by default 1
rng (Generator, optional) – Numpy random number generator, by default None

compute_spike(t, x)[source]

Computes the evolution of node temperature given a timestep t

Parameters

t (int) – The timesteps
x (NDArray) – The vector of node temperatures of shape (num_nodes x 1)

Returns

NDArray – The vector of new node temperatures of shape (num_nodes x 1)

class numgraph.utils.spikes_generator.ColdHeatSpikeGenerator(t_max=10, heat_spike=(0.7, 2.0), cold_spike=(- 1.0, 0.0), prob_cold_spike=0.2, num_spikes=2, rng=None)[source]

Bases: numgraph.utils.spikes_generator.SpikeGenerator

The generator of the spikes in the heat diffusion over a graph. To each spike is associated an increase (or decrease) of temperature of the node. This class identify at each step the node with injected temperature and the new temperature.

Parameters

t_max (int, optional) – The maximum number of timesteps in the simulation, by default 10
heat_spike (Tuple[float, float], optional) – A tuple containing the min and max temperature of a hot spike, by default (0.7, 2.)
cold_spike (Tuple[float, float], optional) – A tuple containing the min and max temperature of a cold spike, by default (-1., 0.)
prob_cold_spike (float, optional) – The probability of a cold spike to happen, by default 0.2
num_spikes (int, optional) – The number of heat spikes during the process, by default 1
rng (Generator, optional) – Numpy random number generator, by default None

compute_spike(t, x)[source]

Computes the evolution of node temperature given a timestep t

Parameters

t (int) – The timesteps
x (NDArray) – The vector of node temperatures of shape (num_nodes x 1)

Returns

NDArray – The vector of new node temperatures of shape (num_nodes x 1)