fix(streaming): reset wrapper sample and cycle counters on reset_state_dict

[truecallerabreham]

Before submitting

• Was this discussed/agreed via a Github issue? (no need for typos and docs improvements)
• Did you read the contributor guideline, Pull Request section?
• Did you make sure to update the docs?
• Did you write any new necessary tests?

What does this PR do?

Fixes #927.

Resuming CombinedStreamingDataset or ParallelStreamingDataset from a mid-epoch checkpoint no longer leaks num_samples_yielded or num_cycles into subsequent epochs.

When StreamingDataLoader.load_state_dict() restored a mid-epoch checkpoint and finished the resumed epoch, the next epoch called self.dataset.reset_state_dict(), which reset child StreamingDataset instances but left self._num_samples_yielded on _BaseStreamingDatasetWrapper and self._num_cycles on ParallelStreamingDataset populated from the checkpoint. In addition, load_state_dict() and the dataset iterators held direct references to the restored counter lists, so in-place increments mutated the loaded state dict in single-worker mode (num_workers=0).

• Clear self._num_samples_yielded = None in _BaseStreamingDatasetWrapper.reset_state_dict() and self._num_cycles = None in ParallelStreamingDataset.reset_state_dict().
• Deep-copy num_samples_yielded and num_cycles in load_state_dict() and _CombinedDatasetIterator / _ParallelDatasetIterator so in-place counter updates do not mutate the caller's state dict.
• Add regression tests in tests/streaming/test_combined.py and tests/streaming/test_parallel.py verifying multi-epoch iteration and subsequent checkpoint resumption after a mid-epoch restore.

PR review

Anyone in the community is free to review the PR once the tests have passed.

Did you have fun?

Make sure you had fun coding 🙃

[deependujha]

thanks :)

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

❌ Patch coverage is 91.66667% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 82%. Comparing base (bb96b0e) to head (5843d9b).
❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@          Coverage Diff          @@
##            main    #928   +/-   ##
=====================================
  Coverage     82%     82%           
=====================================
  Files         70      70           
  Lines      16153   16160    +7     
=====================================
+ Hits       13258   13264    +6     
- Misses      2895    2896    +1     

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.