Hi there,
I noticed that some comments in class DistGraph of ./python/dgl/distributed/dist_graph.py say that:
“For heterogeneous graphs, users need to convert them into DGL graphs with one node type and one edge type and store the actual node types and edge types as node data and edge data.”
Does it still stand true if we want to save externally heterogeneous graphs into DGL files (local or HDFS) in a distributed way? If users have to do that manually, can someone show me an example as to how we can achieve that? Or is there any intelligent way taking care of the auto–conversion as the code/community evolves over time? Thanks
Are below statements still true for DistGraph, say "DistGraph
currently only supports graphs with only one node type and one edge type."? Thanks
Note
----
``DistGraph`` currently only supports graphs with only one node type and one edge type.
For heterogeneous graphs, users need to convert them into DGL graphs with one node type and
one edge type and store the actual node types and edge types as node data and edge data.
One similar question was posted by me in here:
I think I have answered the question in the other post. Let’s discuss the questions in this post: Partition graph with deprecated reshuffle = False - #3 by superhaiou