[Data] Fix incorrect pending task size if outputs are empty

[bveeramani]

Why are these changes needed?

If an operator outputs empty blocks, then Ray Data thinks that the operator has 256 MiB of pending task outputs, even though it should be 0. For example:

import pyarrow as pa
output = pa.Table.from_pydict({"data": [None] * 128})
assert output.nbytes == 0, output.nbytes

The reason for the bug is because we check if average_bytes_per_output is truthy rather than if it's not None.

ray/python/ray/data/_internal/execution/interfaces/op_runtime_metrics.py

Lines 369 to 371 in 1f83fb4

	bytes_per_output = self.average_bytes_per_output
	if bytes_per_output is None:
	bytes_per_output = context.target_max_block_size

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 :(