Pipeline encode stage#408
Merged
jeromekelleher merged 2 commits intoMay 19, 2026
Merged
Conversation
Introduce _EncodedChunk dataclass and _encoded_chunks generator method on FormatEncoder. _restart now wires self._iterator to that generator instead of variant_chunks directly; _advance pops _EncodedChunk items and publishes them. No behaviour change — _encode_chunk still runs on the consumer thread synchronously; this only sets the seam for a later PrefetchIterator wrap to push the encode work one chunk ahead.
Promote _PrefetchIterator from retrieval.py to utils.PrefetchIterator (public; same single-worker / one-deep / drain-on-close contract). Wrap FormatEncoder._encoded_chunks in utils.PrefetchIterator so that _encode_chunk runs on a background prefetch worker while the consumer drains the previous chunk's bytes — the encode pipeline now stacks on top of the existing StreamReader block-prefetch and the inner variant_chunks PrefetchIterator. Tests: - Move TestPrefetchIterator from test_retrieval.py to test_utils.py and rename to drop the private prefix. - Add TestReadahead to test_format_encoder.py covering: iterator-type assertion, encode runs on the vcztools-prefetch thread, worker exception surfaces on the consumer's read(), close joins the worker, no thread leaks across 100 construct/close cycles. - Replace _FakeEncoder's one-shot _encode_raise_once trigger in test_restart_after_failed_advance_emits_restart_not_init with a chunk-content-based _fail_if_first_variant trigger so the failure hits the chunk the consumer is actually waiting for (the prior flag could be consumed silently by a drained in-flight prefetch). Benchmark (wide_bench.vcz, _DiscardSink drain) is within noise — encode is the bottleneck and the synthetic drain is trivial, so producer/ consumer overlap has nothing to recover. Real workloads with a non- trivial drain (file/network) should benefit.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.