pyg-team/pytorch_geometric

Increment of available per-dataset information

Open

#7848 opened on Aug 4, 2023

View on GitHub
 (3 comments) (0 reactions) (0 assignees)Python (19,985 stars) (3,514 forks)batch import
1 - Priority P1datasetdocumentationfeaturehelp wanted

Description

🚀 The feature, motivation and pitch

This is a request that is related to the amount of information available per dataset in PyG. I have noticed that some dataset papers have started to increase the amount of information available, and include several graph properties which might be interesting for researchers that seek to benchmark algorithms. One example of such paper is "A Critical Look at the Evaluation of GNNs under Heterophily: Are We Really Making Progress?", where they include properties such as graph diameter, average degree... image I believe that adding this or similar information can be helpful and adding it bit by bit to the dataset cheatsheet could be good.

Alternatives

Another version of this, could be the the properties from Network Repository, which slightly differ from the ones used in the previous paper. image

Additional context

I found myself developing a layer that worked better under specific graph properties, and I believe people can benefit from this information being more readily accesible

Contributor guide

Tech stack
pythonpytorch
Domain
documentationdata
Issue type
feature
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
half day
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
fresh
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
clear
Prerequisites
PythonPyTorch Geometricgraph properties
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
70
Research direction
Investigate the current dataset cheatsheet in PyG documentation (likely in docs/ or examples/). Look at the Dataset base class in torch geometric/data/dataset.py to understand how to add computed properties like graph diameter, average degree, etc. Consider implementing a utility function to compute these properties and integrate them into the dataset metadata. Check the referenced paper for the specific list of properties.