Datasets¶

The canonical example networks from the format documentation on mapequation.org, bundled with the package. Each loader reads its file with the same parser the Infomap CLI uses and returns a fresh Network, ready to run:

import infomap

net = infomap.datasets.two_triangles()
result = infomap.run(net)

modular_wd() and states() encode directed flow, so those loaders return a Network pre-configured with --flow-model directed; a flow_model passed at infomap.run() still wins.

Basic examples¶

infomap.datasets.two_triangles() → Network¶

Two triangles joined by a single bridge.

The minimal example of community structure: 6 nodes named A–F and 7 unweighted, undirected links forming two triangles connected by one bridge link. A random walker stays long within each triangle and rarely crosses the bridge, so the map equation finds the two triangles as modules.

Returns:: A fresh network with 6 nodes and 7 links.
Return type:: Network

Examples

>>> import infomap
>>> net = infomap.datasets.two_triangles()
>>> net.num_nodes, net.num_links
(6, 7)
>>> net.run().num_top_modules
2

infomap.datasets.nine_triangles() → Network¶

Nine triangles with two-level hierarchical structure.

27 nodes and 39 links: nine triangles grouped three-and-three into three super-groups, with unit-weight links inside each super-group and weaker 0.8 links between them. The optimal solution is hierarchical: three super-group modules at the top level, each nesting its three triangles. Use it to explore num_levels and the multilevel tree.

Returns:: A fresh network with 27 nodes and 39 links.
Return type:: Network

Examples

>>> import infomap
>>> net = infomap.datasets.nine_triangles()
>>> net.num_nodes, net.num_links
(27, 39)
>>> net.run(options={"seed": 123, "num_trials": 10}).num_top_modules
3

Weighted and directed¶

infomap.datasets.modular_w() → Network¶

A weighted network with four modules.

25 nodes (ids 0–24) and 43 weighted, undirected links forming four naturally connected groups. This is the network behind the hero figure of this documentation, sized so that module structure is visible yet non-trivial.

Returns:: A fresh network with 25 nodes and 43 links.
Return type:: Network

Examples

>>> import infomap
>>> net = infomap.datasets.modular_w()
>>> net.num_nodes, net.num_links
(25, 43)

infomap.datasets.modular_wd() → Network¶

A weighted, directed network.

The directed variant of modular_w(): 25 nodes (ids 1–25) and 48 weighted links declared as *Arcs, including asymmetric pairs where the two directions carry different weights. Directed flow changes the module structure compared with the undirected variant, which is the point of comparing the two.

The returned Network is pre-configured with --flow-model directed (the canonical way to run this example); pass flow_model at run() to override.

Returns:: A fresh network with 25 nodes and 48 links, pre-configured for directed flow.
Return type:: Network

Examples

>>> import infomap
>>> net = infomap.datasets.modular_wd()
>>> net.num_nodes, net.num_links
(25, 48)

Bipartite¶

infomap.datasets.bipartite() → Network¶

A small bipartite network with named nodes.

Figure 2 in the input-formats documentation on mapequation.org: 3 ordinary nodes and 2 feature nodes with 4 weighted links between the node types, declared with a *Bipartite section (the feature nodes start at id 4). Infomap models bipartite flow directly on the two-mode structure.

Returns:: A fresh bipartite network with 5 nodes and 4 links.
Return type:: Network

Examples

>>> import infomap
>>> net = infomap.datasets.bipartite()
>>> net.bipartite_start_id
4

Multilayer¶

infomap.datasets.multilayer() → Network¶

A multilayer network in the explicit *Multilayer format.

Figure 3 in the input-formats documentation on mapequation.org: 5 physical nodes named i–m in 2 layers, with all 12 intra-layer and 4 inter-layer links given explicitly as layer node layer node weight rows. The explicit format gives full control over movement between layers; compare multilayer_intra_inter() and multilayer_intra() for the constrained and relax-rate variants of the same network.

Returns:: A fresh multilayer network with 6 state nodes (5 physical nodes across 2 layers) and 16 links.
Return type:: Network

Examples

>>> import infomap
>>> net = infomap.datasets.multilayer()
>>> net.num_nodes
6

infomap.datasets.multilayer_intra_inter() → Network¶

A multilayer network in the *Intra/*Inter format.

Figure 4 in the input-formats documentation on mapequation.org: the same 5 physical nodes and 2 layers as multilayer(), but written as 12 intra-layer links plus 2 *Inter rows (layer node layer weight) that constrain how a random walker switches layer at a physical node. Infomap expands the inter-layer weights into explicit multilayer links.

Returns:: A fresh multilayer network with 5 physical nodes across 2 layers.
Return type:: Network

Examples

>>> import infomap
>>> net = infomap.datasets.multilayer_intra_inter()
>>> net.num_nodes
6

infomap.datasets.multilayer_intra() → Network¶

A multilayer network with intra-layer links only.

Figure 5 in the input-formats documentation on mapequation.org: the same 5 physical nodes and 2 layers as multilayer(), but with only the 12 intra-layer links. Inter-layer movement is generated by the relax rate (multilayer_relax_rate, default 0.15): the walker relaxes to a random layer at the current physical node with that probability.

Returns:: A fresh multilayer network with 5 physical nodes across 2 layers.
Return type:: Network

Examples

>>> import infomap
>>> net = infomap.datasets.multilayer_intra()
>>> net.num_nodes
6

Memory and state¶

infomap.datasets.states() → Network¶

A memory network in the *States format.

Figure 6 in the input-formats documentation on mapequation.org: 5 physical nodes named i–m carrying 6 state nodes with 16 weighted, directed links between the states. Two state nodes share the physical node i, so where the walker goes next depends on where it came from – the memory effect that state networks capture and that overlapping modules reveal.

The returned Network is pre-configured with --flow-model directed (the canonical way to run this example); pass flow_model at run() to override.

Returns:: A fresh state network with 6 state nodes on 5 physical nodes and 16 links, pre-configured for directed flow.
Return type:: Network

Examples

>>> import infomap
>>> net = infomap.datasets.states()
>>> net.num_nodes
6