Recreate Graphs as from Example from SQL database

[xavierallem]

I have a SQL database of my own where I have data relating to people meeting and the infection, I want to create a Infection dataset just like the example, Here are my questions

1. Is there a Infectious dataset example for reference in example? if not then how? since it has temporal data
2. How can I create the dataset just like Infectious dataset?
3. Where Can I learn Information about how exactly the nodes, graph is created in the dataset?

[rusty1s]

The InfectionDataset in PyG is a synthetic dataset to use to evaluate XAI methods, I am not sure if it fits your needs. How you want to create your infection dataset depends a bit on your needs and the use-case you have in mind. I assume you want to treat this as a binary classification task?

[xavierallem]

Hi,
I wanted I wanted to test the infectious_ct1 dataset from Link, Unfortunately I thought this was already implemented in pyg dataset as both were termed infectious. Is there an example similar to the infection_ct1? I wanted to create a neural network with similar dataset of infectious_ct1.

[rusty1s]

I am not sure if this works, but you may wanna try

from torch_geometric.datasets import TUDataset

dataset = TUDataset(root, name="infectious_ct1")

Does that work?

[xavierallem]

current_dir = os.getcwd()
path = osp.join(current_dir, '../Data/INFECTIOUS')
dataset = TUDataset(path, name='infectious_ct1',use_node_attr=True,use_edge_attr=True).shuffle()

using this for python notebook. Got the same error

[rusty1s]

Progress towards this will be tracked here: #9351.