Skip to content

Tutorial

Flow

The general flow of processing with Netseer is:

  • Load a time-series list of graphs.
  • Predict the next graph in the series.

Getting Started

For this tutorial, I'm defaulting to using uv. Otherwise set up a python project using pip/venv as standard and create a main.py

Setup

Create a python project:

uv init netseer-tutorial
cd netseer-tutorial

Add the Netseer package:

uv add netseer # or pip install netseer

The tutorial takes place in the main.py file.

Loading

There are two ways to create graphs:

  • Creating your own list of graphs. Typically this would be a time-series list of graphs.
  • Randomly generating a list of graphs using Netseer.

Using own graphs

If creating your own graphs, Netseer uses the python package igraph to handle graphs.
Therefore, the created graphs need to be saved in a format that can be used by igraph.Graph.Read. For our tutorial we will used GML format graphs.

Save the list of graphs into a directory/folder relative to your project. For example in a /data/ directory. E.g.:

data/
├── graph_1.gml
├── graph_2.gml
└── graph_3.gml

Then create a list of absolute paths. This is easily done using pathlib and glob.

# Get the directory/folder with all the graphs.
directory_path = Path.cwd() / Path("path/to/data")
# Use glob to get a list of absolute paths for .gml files.
graph_files = directory_path.glob("*.gml"))

The graph_files will now contain the path to each graph, as a list.
Similar to: ["home/code/netseer/data/graph_1.gml", "home/code/netseer/data/graph_2.gml","home/code/netseer/data/graph_3.gml"]

Then use the imported utils function read_graph_list() on the graph_files.
This reads each .gml file in the graph_files and loads them into a list of graphs, loaded in memory.
The graphs are loaded in the index order in graph_files.
Use a numbering system for the .gml file names so that sorted can arrange the graphs paths in descending order.

# If you need to sort the graph names.
graph_files = sorted(graph_files)

After the optional sorting:

# Load the graphs into igraph using the list of gml files.
graph_list = utils.read_graph_list(graph_files)

The graphs loaded this way are ordered by their filename, in descending order.
So graph_1 would be at index 0, graph_2 at index 1 etc.
The order of graphs are important for the prediction.

Using random graphs

Alternatively, Netseer has a function to create a random list of graphs.
Instead of using pathlib, glob and read_graph_list(), import generate_graph and use the function generate_graph_list to create a random graph list that is already stored in memory.

from netseer import generate_graph

graph_list = generate_graph.generate_graph_list()

The generate_graph_list() function contains parameters to tailor the generated random graphs, such as growth rate.
Refer to the references documentation for an explanation.

Prediction

After loading the graphs into memory, use the imported prediction function predict_graph() to predict the next graph in the series.
Change the h parameter in predict_graph to predict that many steps into the future. E.g. 1 step, 2 steps, 3 steps etc.

# Then use the list of graphs for predictions.
predict = prediction.predict_graph(graph_list, h=1)