Random Generators

Random Number Generator

class trecs.random.generators.Generator(seed=None, bit_generator=None)

Wrapper around numpy.random.Generator. Please see the Numpy documentation for more details.

Parameters

• seed (int, optional) –

• bit_generator (numpy.random.BitGenerator, optional) – numpy.random.BitGenerator. Please see Numpy’s BitGenerator documentation for more details.

Social Graph Generator

class trecs.random.generators.SocialGraphGenerator

Thin wrapper around the Networkx random graph generators. We use this static class to generate random network adjacency matrices.

By default, it generates a binomial graph, but it can generate any other random graph included in the Networkx API. Please refer to the Networkx documentation.

static generate_random_graph(num, *args, **kwargs)

Note: to change type of graph, please include the graph_type parameter.

Parameters

num (int) – Number of nodes in the graph. This is equivalent to the number of users in the system.

Returns

Adjacency matrix – Size |U|x|U|.

Return type

numpy.ndarray

Raises

ValueError – If num is not an integer.

Examples

A minimal use case:

>>> from trecs.random import SocialGraphGenerator
>>> num = 1000 # <-- number of nodes (users)
>>> graph = SocialGraphGenerator.generate_random_graph(num=num)
>>> graph # <-- 1000x1000 binomial adjacency matrix

Changing random graph generator (e.g., with the random_regular_graph):

>>> from networkx.generators.random_graphs import random_regular_graph
>>> rrg = random_regular_graph
>>> d = 30 # <-- degree, also required by random_regular_graph
>>> graph = SocialGraphGenerator.generate_random_graph(num=num, d=d, graph_type=rrg)