master-thesis

This repo contains all code for the implementation done in Master's Thesis.

Source of datasets

1. Graph Generation:

• DBLP and IMDB : https://github.com/lingchen0331/HGEN
• PubMed : https://github.com/yangji9181/HNE/tree/master/Data

2. Node Feature Generation:

• DBLP : https://pytorch-geometric.readthedocs.io/en/latest/generated/torch_geometric.datasets.DBLP.html
• IMDB : https://pytorch-geometric.readthedocs.io/en/latest/generated/torch_geometric.datasets.IMDB.html
• PubMed : https://github.com/yangji9181/HNE/tree/master/Data

Instructions to run code

1. To run all jupyter notebooks create a conda environment
   conda create --name myenv --file requirements.txt
   conda activate myenv
   jupyter notebook
   

2. To run code for diffusion_graph_gen and diffusion_node_feature_gen in approach_3, follow the installation steps of DiGress: https://github.com/cvignac/DiGress

3. Use the .dat files from the datasets for Graph Generation wherever .dat files are present in code, e.g. {dataset}_link.dat, {dataset}_node.dat