gfn.gym.helpers.bayesian_structure.factories

Functions

get_data(name, num_nodes, num_edges, num_samples[, ...])

Generate Bayesian linear Gaussian data.

get_prior(name)

get_scorer(graph_name, prior_name, num_nodes, ...[, ...])

Module Contents

gfn.gym.helpers.bayesian_structure.factories.get_data(name, num_nodes, num_edges, num_samples, node_names=None, rng=None)

Generate Bayesian linear Gaussian data.

Parameters:

  • name (str) – Data generation method type.

  • num_nodes (int) – Number of variables in the graph.

  • num_edges (int) – Number of edges to sample in the graph.

  • num_samples (int) – Number of samples to generate.

  • node_names (Optional[List[str]]) – Optional list of node names.

  • rng (Optional[np.random.Generator]) – Optional random generator instance.

Returns:

(graph, data, score) where ‘score’ indicates the scoring method used.

Return type:

tuple

gfn.gym.helpers.bayesian_structure.factories.get_prior(name)
Parameters:

name (str)

Return type:

gfn.gym.helpers.bayesian_structure.priors.BasePrior

gfn.gym.helpers.bayesian_structure.factories.get_scorer(graph_name, prior_name, num_nodes, num_edges, num_samples, node_names=None, rng=None)
Parameters:

  • graph_name (str)

  • prior_name (str)

  • num_nodes (int)

  • num_edges (int)

  • num_samples (int)

  • node_names (Optional[list[str]])

  • rng (Optional[numpy.random.Generator])

Return type:

tuple[gfn.gym.helpers.bayesian_structure.scores.BGeScore, pandas.DataFrame, torch_geometric.data.Data]