A recent PR, https://github.com/dmlc/dgl/pull/7727 removed the public function gb_convert_single_dgl_partition from dgl.distributed.partition. In GraphStorm we were planning to use that function to implement distributed GB partition conversion on our end.
Are there plans to implement distributed GB conversion on the DGL side? If not, can we bring back the function with the same signature and add to the public interface of the dgl.distributed.partition module (through its __init__.py)
🐛 Bug
A recent PR, https://github.com/dmlc/dgl/pull/7727 removed the public function
gb_convert_single_dgl_partitionfromdgl.distributed.partition. In GraphStorm we were planning to use that function to implement distributed GB partition conversion on our end.Are there plans to implement distributed GB conversion on the DGL side? If not, can we bring back the function with the same signature and add to the public interface of the
dgl.distributed.partitionmodule (through its__init__.py)Ping @Rhett-Ying