photonai-graph 0.1.0

PHOTON Graph - Graph machine learning with photonai.

photonai-graph

Photon Graph is an extension for the PHOTON framework that allows for the use of machine learning based on graphs. Furthermore, the Graph Utilities contain a wide variety of functions that allow for the visualization and converting of graphs.

Documentation

You can find a detailed documentation here: https://wwu-mmll.github.io/photonai_graph/

Installation

To install photonai-graph create a dedicated conda/python environment and activate it. Then install photonai-graph via

pip install photonai-graph

To be able to use all modules of the toolbox you will still need to install tensorflow, dgl, pytorch and grakel according to your system configuration, for example with

pip install tensorflow
pip install torch
pip install dgl
pip install grakel

For graph embeddings the gem python package is needed, along with tensorflow. Please install tensorflow according to your system.

pip install nxt-gem
pip install tensorflow

For graph kernels the grakel package needs to be installed. You can install grakel via pip.

pip install git+https://github.com/ysig/GraKeL.git@cfd14e0543075308d201327ac778a48643f81095'

For graph neural networks pytorch and deep graph library are required. You can install them via pip

pip install torch
pip install dgl