Export networkx graph

IDTxl allows to export inferred networks to the networkx Python package. The networkx package provides plotting functionalities, graph algorithms, and further export functions. The inferred network is exported as a weighted, directed graph object DiGraph().

Import IO functions and load example data

import pickle
from idtxl import idtxl_io as io

results = pickle.load(open('test/data/mute_results_0.p', 'rb'))

Export inferred network graph

Example: export inferred network, using the lag of the past variable with maximum information transfer into the target to weight edges

weights = 'max_te_lag'
adj_matrix = results.get_adjacency_matrix(weights=weights, fdr=False)
io.export_networkx_graph(adjacency_matrix=adj_matrix, weights=weights)

Export graph of selected source variables

Example: export graph of inferred past variables for a single target

io.export_networkx_source_graph(
    results=results, target=1, sign_sources=True, fdr=False)

By setting sign_sources to True only significant sources are included, otherwise all sources in the cadidate set are plotted.