Thank you for reviewing our paper.

In this readme file we explain on how to obtain the datasets used in the experiments and how to run their code.

Datasets:
All datasets in our expeiments are freely and openly available, and are cited in the paper.
For convenience, we have attached Cora, Pubmed, CiteSeer, Wisconsin, Texas, Cornell and Chameleon datasets in this submission as it is a rather small datasets.

For the rest of the datasets, we use the dataloaders from the useful PyTorch-Geometric which implemented an automatic download of those datasets.

Files contents:

The main files that implement our method is it to be found in:
1. src/graphOps.py - there, we implement the code for graph operators like gradient and divergence.
2. src/graphNet.py - where we implement the networks we used for our experiments.
3. src/utils.py and src/utils_gcnii.py - utility files, some written by us and some by the code kindly shared by the authors of GCNII, which we use to compare our method with GCNII.
4. src/inits.py - initializations for the weights of the networks.

Experiments:
All the experiments that were carried are included in this submission. 
In order to run any of those experiments, simply type in the command line: python src/x.py where x is the file name of the experiment.
All the expriments codes below include both training and evaluation code.

The experiments are as follows:
1. src/edgeconv_cora.py - where we evaluate edge conv (DGCNN) on cora.
2. src/gcnii_modelnet10.py - where we evaluate GCNII on ModelNet-10.
3. src/pdegcn_semi.py - all our PDEGCN semi-supervised node classification experiments.
4. src/pdegcn_fully.py - all our PDEGCN fully-supervised node classification experiments.
5. src/pdegcn_ppi.py - the inductive learning on PPI using our PDEGCN.
6. src/pdegcn_modelnet10.py - shape classification on ModelNet-10 using our PDEGCN.
7. src/pdegcn_faust.py - dense shape correspondence on FAUST with our PDEGCN.

Code dependencies:
We use the following python packages:
-pytorch
-numpy
-pytorch-geometric
-sklearn
-optuna
-matplotlib
-trimesh
-scipy
-mpl_toolkits
