Export GraphML and GEXF

Infomap can write module assignments back to graph objects before exporting them to common visualization formats. This keeps the graph topology and the Infomap result in one file, so tools such as Gephi and Cytoscape can color or filter nodes by community without a separate join step.

NetworkX GraphML and GEXF

For reusable infomap.Infomap workflows, keep the mapping returned by infomap.Infomap.add_networkx_graph(). The export helpers use it to align Infomap’s internal integer ids with the original NetworkX node labels:

import networkx as nx
from infomap import Infomap
from infomap.export import write_gexf, write_graphml

graph = nx.karate_club_graph()

im = Infomap(silent=True, num_trials=20, seed=123)
mapping = im.add_networkx_graph(graph)
im.run()

write_graphml(graph, im, "karate_infomap.graphml", node_mapping=mapping)
write_gexf(graph, im, "karate_infomap.gexf", node_mapping=mapping)

The helpers copy the graph by default, annotate the copy, and pass it to NetworkX’s native writers. Use copy=False when you explicitly want the attributes written to the graph object you pass in.

If you only need top-level modules, the NetworkX-style entry point can write the node attribute directly. Then call NetworkX’s writer yourself:

import networkx as nx
import infomap

graph = nx.karate_club_graph()
infomap.find_communities(
    graph,
    module_attribute="infomap_module",
    seed=123,
    num_trials=20,
)

nx.write_graphml(graph, "karate_infomap.graphml")
nx.write_gexf(graph, "karate_infomap.gexf")

igraph GraphML

For python-igraph, use infomap.Infomap.add_igraph_graph() and infomap.export.write_graphml(). The helper annotates vertex attributes and uses python-igraph’s native GraphML writer:

import igraph as ig
from infomap import Infomap
from infomap.export import write_graphml

graph = ig.Graph.Famous("Zachary")

im = Infomap(silent=True, num_trials=20, seed=123)
im.add_igraph_graph(graph)
im.run()

write_graphml(graph, im, "karate_infomap.graphml")

GEXF export is available for NetworkX graphs only. python-igraph does not provide a native GEXF writer, and Infomap does not convert igraph graphs through NetworkX implicitly because that adds overhead and can change ids or attributes in surprising ways.

Written attributes

By default, exported nodes receive string-valued attributes for broad compatibility with visualization tools:

Attribute	Meaning
infomap_module	Top-level module assignment.
infomap_path	Colon-separated hierarchical module path, such as "3:12".
infomap_level_1, infomap_level_2, …	One attribute per hierarchy level.

Set module_attribute or path_attribute to change the main attribute names. Set include_hierarchy=False when you only want the top-level module attribute.

Visualization workflow

Open the exported GraphML or GEXF file in Gephi or Cytoscape and style nodes by the infomap_module attribute. For hierarchical runs, use infomap_path to identify full paths or infomap_level_N attributes to color one hierarchy level at a time.

The native infomap CLI does not read or write GraphML/GEXF in this release. Use the Python workflow above when you need these interchange formats.