centrality.py

import numpy as np
import networkx as nx
import torch

def eigen_centrality(data):
    """Calculates the eigenvector centrality for the given graph

    Parameters:
    data (np.array): Adjacency matrix representation of the graph

    Returns:
    np.array: Array of eigenvector centralities
    """
    g = nx.from_numpy_matrix(data)

    # # compute egeinvector centrality and transform the output to vector
    ec = nx.eigenvector_centrality_numpy(g, weight="weight")
    eigenvector_centralities = np.array([ec[g] for g in ec])

    return eigenvector_centralities


def pagerank_centrality(data):
    """Calculates the pagerank centrality for the given graph

    Parameters:
    data (np.array): Adjacency matrix representation of the graph

    Returns:
    np.array: Array of pagerank centralities
    """
    g = nx.from_numpy_matrix(data)

    # # compute egeinvector centrality and transform the output to vector
    cc = nx.pagerank(g, weight="weight")
    pagerank_centralities = np.array([cc[g] for g in cc])

    return pagerank_centralities


def betweenness_centrality(data):
    """Calculates the betweenness centrality for the given graph

    Parameters:
    data (np.array): Adjacency matrix representation of the graph

    Returns:
    np.array: Array of betweenness centralities
    """
    g = nx.from_numpy_matrix(data)

    # # compute egeinvector centrality and transform the output to vector
    bc = nx.betweenness_centrality(g, weight="weight")
    betweenness_centralities = np.array([bc[g] for g in bc])

    return betweenness_centralities


def topological_measures(data):
    """Returns the topological measures for the given graph

    Parameters:
    data (np.array): Adjacency matrix representation of the graph

    Returns:
    np.array: Array with first element pagerank centrality,
                         second element betweenness centrality
                         third element eigenvector centrality
    """
    topology = []
    topology.append(torch.tensor(pagerank_centrality(data)))  # 0
    topology.append(torch.tensor(betweenness_centrality(data)))  # 1
    topology.append(torch.tensor(eigen_centrality(data)))  # 2

    return topology