Is there a way to use graph transforms with dataloader in DGL?

Question

I'm doing a Graph Neural Network (GNN) project with DGL library.

I’d like to apply transformations and augmentations to the batched subgraphs sampled via a DGL dataloader class. Is there a way to do so?

In the docs they show you how to transform the full graph, not the batches like in PyTorch Geometric which has a transform parameter in its dataloaders.