Accessing batch id inside a Minibatch

## ❓ Questions and Help

Hi! I am wondering if it is possible to access the batch ID of each node type when sampling a minibatch from the Graphbolt neighbor sampler.

For example, in Pyg, it's possible to get something like by using the NeighborLoader(batch_size = 32)
```
batch.batch_dict
{'drivers': tensor([ 0, 1, 2, ..., 30, 30, 6], device='cuda:0'), 'races': tensor([ 0, 0, 0, ..., 28, 27, 28], device='cuda:0'), 'constructors': tensor([ 0, 1, 1, ..., 30, 16, 21], device='cuda:0'), 'constructor_results': tensor([ 0, 0, 0, ..., 31, 31, 31], device='cuda:0'), 'constructor_standings': tensor([ 0, 0, 0, ..., 15, 28, 31], device='cuda:0'), 'results': tensor([ 0, 0, 0, ..., 28, 31, 31], device='cuda:0'), 'circuits': tensor([ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
..., 31, 2, 17, 24],
device='cuda:0'), 'standings': tensor([ 0, 0, 0, ..., 24, 24, 24], device='cuda:0'), 'qualifying': tensor([ 0, 0, 0, ..., 27, 27, 27], device='cuda:0')}
```
Here, the `batch_dict` can be used to identify the batch correspondence of a specific row and makes some feature construction much easier.

In Minibatch of Graphbolt's sampler, `batch.indices` is empty by default and `batch.input_nodes` can only give us information about the whole MFG graphs.