Expansion Utils

Connect a Star object to the Multiverse system by extracting its networkx graph for use with the Realtor class.

This function extracts the networkx graph from a Star object’s starGraph attribute and prepares it to be used with the Realtor class for spatial analysis.

Parameters:

  • star_obj (Star) – A Star object with an initialized starGraph attribute.

  • node_features (array-like, optional) – Features for each node in the graph. If None, must be provided when creating the Realtor instance.

  • group_features (array-like, optional) – Features for each group/component in the graph. If None, must be provided when creating the Realtor instance.

Returns:

A dictionary containing: - ‘graph’: The networkx graph extracted from the Star object - ‘node_features’: The node features if provided - ‘group_features’: The group features if provided

Return type:

dict

Example

>>> from thema.multiverse.universe.star import Star
>>> from thema.multiverse.universe.stars.jmapStar import JMapStar
>>> from thema.expansion.realtor import Realtor
>>> from thema.expansion.utils import star_link
>>>
>>> # Initialize and fit a JMapStar (a Star implementation)
>>> jmap_star = JMapStar(data_path="path/to/data.csv",
...                       clean_path="path/to/clean.pkl",
...                       projection_path="path/to/projection.pkl")
>>> jmap_star.fit()
>>>
>>> # Extract the graph and feature data for the Realtor
>>> star_data = star_link(jmap_star,
...                        node_features=my_node_features,
...                        group_features=my_group_features)
>>>
>>> # Create a Realtor instance to find the best location for a target vector
>>> target_vector = [0.1, 0.2, 0.3]  # Example target vector
>>> realtor = Realtor(target_vector=target_vector,
...                    graph=star_data['graph'],
...                    node_features=star_data['node_features'],
...                    group_features=star_data['group_features'])
>>>
>>> # Find the best node for docking the target
>>> best_node = realtor.node_docking()
>>> print(f"The best node for the target is: {best_node}")
>>>
>>> # Alternatively, perform a random walk
>>> samples = realtor.random_walk(n_samples=500, m_steps=800)
>>> print(f"Random walk samples: {samples[:5]}...")