[core][cgraph] Export classes related to NCCL communicator

[ruisearch42]

Why are these changes needed?

These classes need to exported to implement external communicators.
See vllm-project/vllm#19816

Related issue number

Checks

• I've signed off every commit(by using the -s flag, i.e., git commit -s) in this PR.
• I've run scripts/format.sh to lint the changes in this PR.
• I've included any doc changes needed for https://docs.ray.io/en/master/.
  • I've added any new APIs to the API Reference. For example, if I added a
    method in Tune, I've added it in doc/source/tune/api/ under the
    corresponding .rst file.
• I've made sure the tests are passing. Note that there might be a few flaky tests, see the recent failures at https://flakey-tests.ray.io/
• Testing Strategy
  • Unit tests
  • Release tests
  • This PR is not tested :(

[Copilot]

Pull Request Overview

This PR exports classes related to the NCCL communicator to support external communicators. The changes include:

• Updating the recv function documentation in nccl_group.py with new parameters (shape, dtype, allocator).
• Exporting AcceleratorContext and TorchTensorAllocator in init.py as part of the public API.

Reviewed Changes

Copilot reviewed 2 out of 3 changed files in this pull request and generated no comments.

File	Description
python/ray/experimental/channel/nccl_group.py	Updated the function documentation to include new parameters for tensor reception.
python/ray/experimental/channel/init.py	Modified imports and export list to include new accelerator-related classes.

Comments suppressed due to low confidence (2)

python/ray/experimental/channel/nccl_group.py:201

• Ensure that the function signature has been updated to include the new parameters 'shape', 'dtype', and 'allocator', and that these parameters are correctly handled in the function logic.

            shape: The shape of the tensor to receive.

python/ray/experimental/channel/init.py:14

• Verify that exposing 'AcceleratorContext' and 'TorchTensorAllocator' aligns with the overall module design and that they are implemented appropriately to maintain a consistent public API.

from ray.experimental.channel.accelerator_context import AcceleratorContext

[stephanie-wang]

Sorry for the delay in review.

Can you add a test to show example usage? E.g., create a torch.distributed group manually and then wrap it with the Ray interfaces?