[codex] Polish RankQuant maintenance surfaces

[Fieldnote-Echo]

Summary

• Align BucketCode / RankQuantSpec with the current fixed-composition b=8 contract: b=8 is accepted when dim % 256 == 0 and rejected as non-uniform otherwise.
• Move the bench-only search_asymmetric_byte_lut helper behind a non-default bench-utils feature, and update the benchmark/docs/tests that intentionally use it.
• Rename checked_new_len to checked_new_count, add L2_NORMALISE_EPSILON, pin L2 threshold-edge behavior, and trim dense inline rationale comments in rank bucketing helpers.

Validation

• cargo test --locked
• cargo test --locked --all-features
• cargo clippy --locked --all-targets --all-features -- -D warnings
• cargo check --locked --features bench-utils --example bench_rank
• cargo fmt --check
• git diff --check

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 3fa2395218

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "Codex (@codex) review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "Codex (@codex) address that feedback".

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[Fieldnote-Echo]

/agentic_review

[qodo-code-review]

Looking for bugs?

Check back in a few minutes. Qodo's review agents are on it.

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

[github-actions]

Review Summary

Overall this is a well-structured maintenance PR. The bench-utils feature gating is clean, the API changes are correct, and the comment trimming in rank_to_bucket reduces noise without sacrificing clarity.

One outstanding concern from the previous review (P2) remains partially unaddressed — the SDE AVX-512 jobs still don't enable bench-utils, so the SIMD-parity tests are skipped under SDE. The parity tests are valuable for verifying that AVX-512 produces identical results to the scalar reference, and they should run in that environment.

P2: SDE jobs missing --features bench-utils

The avx512 job in .github/workflows/ci.yml:437-447 and release-avx512 in .github/workflows/release.yml:254-261 run:

cargo test
cargo test --features experimental

The SIMD-parity tests (rt2_asym_*_matches_scalar, byte_lut_huge_k_*, batched_subset_rerank_matches_scalar_reference_across_tiers, delta_e1_byte_lut_panics_on_b1_index) are all gated #[cfg(feature = "bench-utils")] and won't run without that feature flag. Adding --features bench-utils to both cargo test commands is a one-line change per command and ensures the AVX-512 kernels are verified against the byte-LUT scalar reference under SDE.

---

Was this automated review useful? React with 👍 or 👎 to this review to help us measure review quality.
Workflow run: https://github.com/Project-Navi/ordvec/actions/runs/27857100213

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.