PolyGraphs

Appearance

Custom Graphs ​

Graphs defined using Graph Modelling Language can be imported for use with Polygraphs. Graphs from GML files can be loaded by setting network.kind parameter as gml and specifying a name in network.gml.name and path to GML file in network.gml.path. The name parameter is used to identify the network using the configuration file stored with the results during the analysis stage.

PolyGraphs assumes the file contains an undirected graph by default, you can set the parameter network.gml.directed to True to specify that you have a directed graph.

TIP

Polygraphs loads the graph from the GML file using the edge list

Remember that the graph used to run the simulation is stored inside the .bin file inside the results directory. Ensure that any analysis you do of the results are made using the graph from this file (which can be accessed using the PolyGraphs analyser) to avoid any inconsistencies with the graph.

Example Configuration ​

The following configuration file loads a GML file called custom_graph.gml which contains a directed graph from the user's home directory:

yaml
# Kind and size of network (a complete network with 10 agents)
network.kind: "gml"
network.gml.name: "CustomGMLNetwork"
network.gml.path: "~/custom_graph.gml"
network.gml.directed: True

# Initial beliefs are random uniform between 0 and 1
init.kind: "uniform"
# Chance that action B is better than action A
epsilon: 0.01
# Enable logging; print progress every 100 steps
logging.enabled: True
logging.interval: 100
# Enable snapshots; take one every 100 steps
snapshots.enabled: True
snapshots.interval: 100
# Run for 1,000 steps
simulation.steps: 1000
# Set model
op: "BalaGoyalOp"

Creating GML Files ​

The Polygraphs repository contains an example notebook which demonstrates the process for generating a GML file from the Six Degrees of Francis Bacon network.

Node ID Normalisation ​

Node ids in GML files are normalised before they are loaded in Polygraphs. Normalised ids for all nodes can be accessed from the gml_id of Polygraphs processor graphs object:

python
processor.graphs[0].pg['ndata']['gml_id']

You can also simulate the normalisation process using the Processor.normalise_gml() method on any GML file to get a dictionary of resulting node ids:

python
from polygraphs.analysis import Processor
Processor.normalise_gml("~/path_to/graph.gml.gz")