PROBE-NXG-HIST-1: nested-mask histogram reproduces the partition-point rank on 94 572 real rows

[AdaWorldAPI]

What this is

Plan nexgen-mask-histogram-thresholds-v1 §5 step 2 — the first gate before any of D-NXG-4..12 leaves PROPOSAL. Operator: "merge 1176 and start PROBE-NXG-HIST-1" (#1176 merged as a7520a15).

crates/lance-graph-planner/examples/probe_nxg_hist_1.rs — builds one 16-band nested-mask histogram from a REAL in-tree column (|sample| of data/tts-cascade/tts_real_output.wav, 94 572 rows of 16-bit speech) using only named T1 primitives (gt_i32_to_mask, mask_andnot, mask_ternlog::<AND_ANDNOT2>, popcount_batch_u64) and asserts three pre-registered claims at the END of the run.

CARGO_PROFILE_DEV_DEBUG=0 cargo run -p lance-graph-planner --example probe_nxg_hist_1

Result (first run, verbatim)

rows N = 94572, mask words = 1478, bands B = 16
bucket popcounts: [5913, 5925, 5900, 5907, 5908, 5914, 5918, 5903, 5915, 5906, 5911, 5913, 5910, 5910, 5908, 5911]
C1 nested=true disjoint=true sum=94572 (N=94572)
C2 rank mismatches = 0 / 94572; non-empty buckets = 16/16; max bucket = 5925
C3 named==andnot: true; AND3 differs where lower band non-empty: true
C1 PASS · C2 PASS · C3 PASS.

Disable runs, red-then-green:

• bucket via AND3 instead of AND_ANDNOT2 → C1 disjoint=false, sum=715 262; C2 88 659 mismatches; C3 false; exit 101
• partition_point(|b| b <= v) → C2 192 mismatches (rows exactly on a boundary); exit 101
• restored → green

What the probe corrected

Plan §3 room 3 claimed σ is recoverable from the histogram via popcount-weighted bucket midpoints. Measured: σ_hist 4 244.9 vs σ_direct 3 785.8, ratio 1.121 — quantile buckets are equal-mass, the top bucket (9 299..20 634) is a heavy tail, midpoints over-weight it. Regraded in E-NXG-17: needs per-bucket means; stays an observation. Also: normalized entropy over freshly re-quantiled buckets is 0.99999964 by construction, so D-NXG-9's rollover timer must read FIXED boundaries after the distribution moves.

Honest scope

• Audio, not a facet column — no facet fixture exists in-tree (2026-09-05 harvest). Mechanism is value-type-agnostic; facet rerun is the next arm.
• No library code changed. Example only.
• popcount_batch_u64 exists but is a scalar count_ones sum, not the POPCNT lane its comment promises (D-NXG-2 half-closed).

Board

E-NXG-17 prepended; AGENT_LOG entry; STATUS_BOARD D-NXG-1/2/4 rows; plan §5 step 2 GREEN. Supersession index regenerated last (no diff). One pre-existing AGENT_LOG.md:481 citation exposed by the prepend was re-anchored by heading per the citation-decay gate's own instruction (second commit).

Gates run locally

clippy -D warnings on the example: clean · fmt: clean · append-only: OK (9 files) · citation-decay --since origin/main: 0 new · plan-dids: nothing to check

🤖 Generated with Claude Code

https://claude.ai/code/session_016WkNBjHc2e3zuyz9i8qJEv

---

Generated by Claude Code

[coderabbitai]

Warning

Review limit reached

Next included review available in 23 minutes.

Check out review usage here.

View limit details

Limit details: You’ve used the included review currently available. Your 68 included PR review attempts over the past 7 days set your current allowance at 1 review per hour.

Your organization has reached its usage spending cap. Adjust your spending cap in the billing tab.

Learn how review limits work.

Review configuration:

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Essentials

Run ID: 552a821d-7741-4bcf-b96b-afa3c97179c0

📥 Commits

Reviewing files that changed from the base of the PR and between a7520a1 and 466f310.

📒 Files selected for processing (6)

• .claude/board/AGENT_LOG.md
• .claude/board/EPIPHANIES.md
• .claude/board/STATUS_BOARD.md
• .claude/board/TECH_DEBT.md
• .claude/nexgen/plans/nexgen-mask-histogram-thresholds-v1.md
• crates/lance-graph-planner/examples/probe_nxg_hist_1.rs

---

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

[cursor]

Bugbot couldn't run - usage limit reached

Bugbot is counted against Cursor usage for this user or team, and this run hit a usage or spend limit.

A user or team admin can review and increase usage limits in the Cursor dashboard.

(requestId: serverGenReqId_6b16e420-b54f-4495-8d59-c61195431e6a)

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, you can upgrade your account or add credits to your account and enable them for code reviews in your settings.