search

dmlc / dgl

Python package built to ease deep learning on graph, on top of existing DL frameworks.
http://dgl.ai
Apache License 2.0
13.46k stars 3.01k forks source link

[Feature] provide various data loaders #1220

Closed zheng-da closed 2 years ago

zheng-da commented 4 years ago

🚀 Feature

DGL should provide data loaders for common data sources, including graph databases, relational databases, no-sql databases, cloud filesystesm, etc. Here are a few examples.

Motivation

If we want DGL to be part of a pipeline in the industry, DGL needs to load data from these common data storage.

businessabstraction commented 4 years ago

Please consider loading from Sparql databases as well (Stardog, Neptune if used for RDF, Allegro, etc) Happy to provide more info Sparql (https://en.wikipedia.org/wiki/SPARQL) is effectively a graph pattern query language. However, RDF databases offer advantage of standard entities and some deterministic deductive reasoning. Some percentage of sites prefer to use Semantic Grpah (Sparql) databases rather than Property Graph databases for integration

classicsong commented 4 years ago

Please consider loading from Sparql databases as well (Stardog, Neptune if used for RDF, Allegro, etc) Happy to provide more info Sparql (https://en.wikipedia.org/wiki/SPARQL) is effectively a graph pattern query language. However, RDF databases offer advantage of standard entities and some deterministic deductive reasoning. Some percentage of sites prefer to use Semantic Grpah (Sparql) databases rather than Property Graph databases for integration

Currently, DGL's plan is to support CSV-like data format in its DataLoader first, and provide some stuff to support automatically building the dgl graph. We are not planing to directly support querying the graph database using SPARQL or Gremlin inside DGL. An standalone tool can be used such as https://github.com/awslabs/amazon-neptune-tools/tree/master/neptune-export for Neptune.

davidshumway commented 2 years ago

dgl.data.rdf appears to be relevant toward providing this feature (https://github.com/dmlc/dgl/blob/195f99362d883f8b6d131b70a7868a537e55b786/python/dgl/data/rdf.py)

github-actions[bot] commented 2 years ago

This issue has been automatically marked as stale due to lack of activity. It will be closed if no further activity occurs. Thank you

github-actions[bot] commented 2 years ago

This issue is closed due to lack of activity. Feel free to reopen it if you still have questions.

OysterQAQ commented 1 year ago

Is there any progress on this matter?