fix(scaling): bound top-1 CSR loser streams and payload budget

[seonghobae]

Why

Advances #731. Top-1 CSR materialization used unbounded list(losers), leaked ordinary iterator exception text, and did not share the ranking CSR byte ceiling with full rankings.

What

• Bound loser streams to n - 1 with a stable overlong error
• Normalize outer/inner iteration failures (non-reflective); propagate process-control
• Enforce MAX_RANKING_CSR_BYTES on winner+loser+start uint64 payload
• Validate-then-allocate exact-size handoff arrays
• RED→GREEN tests, APA doctoring, changelog fragment

Supersedes draft #731 once green.

[coderabbitai]

Warning

Review limit reached

@opencode-agent[bot], you've reached your PR review limit, so we couldn't start this review.

Next review available in: 29 minutes

You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository.

How can I continue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.

How do review limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro Plus

Run ID: 8caf1ffd-8660-428f-b135-fdd0a67b84f3

📥 Commits

Reviewing files that changed from the base of the PR and between bb30b19 and a9ec6a4.

📒 Files selected for processing (5)

• CHANGELOG.md
• docs/changelog.d/731-top1-csr-input-bounds.md
• docs/doctoring/top1_csr_input_bounds.md
• python/fast_mlsirm/scaling.py
• tests/test_scaling_top1_input_bounds.py

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[seonghobae]

@opencode-agent address

Exact bounded integration handoff for PR #732. Before writing, refetch source head ab9c7aabb83e5a0f04fb40e623affc03121138f5, resolve protected main (currently bb30b196d2f83df5117a6bebf5e9680faf18c841 after #728), and abort if either moved or another writer appears. The branch is now one protected-main commit behind and mergeable=false; its prior CI/Security/SAST success is predecessor-base evidence only.

Reconcile current protected main non-destructively while preserving the accepted ATA semantic-preflight change from #728 and this PR's unique five-file top-1 CSR slice. Do not force-push or rewrite fail-first lineage. Regenerate CHANGELOG.md only through the repository-authoritative fragment renderer after reconciliation. Re-run focused top-1 CSR bounds/parity tests, full Python 3.12/3.14, Rust/PyO3/package/reinstall/GPU-existing/fuzz, Security Scan and SAST on one unchanged final head. CodeRabbit is currently rate-limited on the predecessor head; do not spam same-head requests. Once the head changes after legitimate reconciliation, request one fresh review when available. Keep the branch merge candidate only if live-base mergeability returns, all exact-head gates are green, and zero valid unresolved findings remain.