How-to Guides

Here you can find some guides that will help you get up to speed in using libmg in your work.

Installation

libmg can be installed with pip (recommended) or from source. In this section you will find the prerequisites and the means to verify your installation.

How to define a dataset

A \(\mu\mathcal{G}\) model is run on a given dataset. Here you will find how to define a dataset of graphs.

How to define functions

A \(\mu\mathcal{G}\) is built by assembling the basic terms from \(\psi\), \(\phi\) and \(\sigma\) functions. Here you will find how these functions can be defined.

How to define a compiler

The compiler for \(\mu\mathcal{G}\) expressions must be instantiated given the functions that are to be used and some configuration information. Here you will find what kind of configuration is needed.

How to create, run, and optimize models

Creating a \(\mu\mathcal{G}\) model consists in compiling a \(\mu\mathcal{G}\) expression. The model can then receive graphs in input using the adequate dataset loader. Models can also be traced by the compiler using dummy data.

Visualization procedures

Here you will find information on how to visualize graphs and model outputs on your web browser.