[d-morrison]

🤖 Generated with Claude Code

https://claude.ai/code/session_019xfc9qrViJfWgYsyX5QcEX

Closes nothing on its own; it is step 1 of sds#617.

Why

sds and lds write predictor vectors as \vpredvar.
Their odds-multiplier results (sds#614, lds#2430) therefore spell out \diff \vpredvar about 25 times, where rme writes \diffvx.
Ezra asked why those PRs didn't use a shorthand as rme does.
No \diffvpredvar existed, so this PR adds one.

Changes

• macros.qmd: \def\diffvpredvar{\diff\vpredvar}, next to \vpredvar. It builds on \diff and \vpredvar, as \diffvx builds on \diff and \vx.
• interpretations.tsv: one row. cut -f1 interpretations.tsv | sort | uniq -d prints nothing.
• notation-choices.qmd: the deviation-and-difference section now names \diffvpredvar as the \vpredvar-notation twin of \diffvx.
• data/symbols-all.csv: regenerated with python3 tools/gen_symbols.py.

Checks run locally

• python3 tools/gen_symbols.py --check exits 0.
• python3 tools/check_symbols_mathjax.py (mathjax-full 3) typesets 1208 symbols with 0 failures.
• A PDF test page, which includes macros.qmd through include-in-header and uses $\diffvpredvar_i^2$, $\oddsfactorf{\diffvpredvar}$ and $\dprod{\vcoef}{\diffvpredvar}$, rendered with pdflatex with no errors or warnings. I deleted the page afterwards rather than committing it.

Next

Once this PR merges, the sds and lds latex-macros pins move past it, and the odds-multiplier subfiles switch to \diffvpredvar (sds#617).

---

Generated by Claude Code

[d-morrison]

/review

Posted by Claude Code (AI agent) --- not written by a human.

---

Generated by Claude Code

[d-morrison]

/review

Pushed 77f8701: it applies the line-break suggestion. On the second point, I checked the downstream claim: sds and lds both write predictors as \vpredvar (for example sds _subfiles/logistic-inference/_thm-odds-multiplier.qmd).

Posted by Claude Code (AI agent) --- not written by a human.

---

Generated by Claude Code

[github-actions]

👀 Claude Review Dispatched — run reviewing PR #426 at 77f8701.

[github-actions]

Claude finished review — View run

---

I fetched PR #426 live just now. It is open, not a draft and not merged, and its head is 77f87010cf965cf817f34f93cda725af2555e6bc. That is the commit I reviewed, so the review below still applies.

The one new commit, 77f8701, only re-breaks the \diffvpredvar note at notation-choices.qmd:546. The rendered text is unchanged apart from splitting one sentence into two. No math or macro definitions changed. The prior review (2026-11-13, commit 501963a, comment) found no blocking problems, and that verdict still stands.

One non-blocking point remains: the claim that sds and lds write predictor vectors as \vpredvar is unverified, because those repositories aren't checked out here.

Verdict

Ready for merge

Structured Review Data (JSON)

{
  "schema_version": "1.1",
  "reviewer": "claude",
  "commit_sha": "77f87010cf965cf817f34f93cda725af2555e6bc",
  "verdict": "CLEAN",
  "findings": [],
  "detailed_assessment": "The new commit only re-wraps the diffvpredvar note at notation-choices.qmd:546 into clause-level lines; rendered content is unchanged and no macro or table row changed.",
  "holistic_assessment": "The earlier checks on the macro definition, interpretations row, generated symbols file and notation page still hold; only the unverified downstream-usage wording remains as a non-blocking point."
}

Reviewed commit: 77f8701

[github-actions]

💰 Cost: $0.1834 (review) — run

[github-actions]

PR Preview Action v1.8.2
🚀 Preview available at: https://Morrison-Lab.github.io/macros/pr-preview/pr-426/

_{Built to branch gh-pages at 2026-11-13 05:28 UTC.}

[d-morrison]

Cross-harness review (agy Flash) on 77f8701: VERDICT: CLEAN

\diffvpredvar is defined as \diff\vpredvar, mirroring \diffvx; both of those and \vpredvar exist on macros main. The csv row sorts between \difflogodds and \diffvx, the tsv row sits in the predictor-variable group, the two description strings are identical, and the notation-choices prose is accurate.

Posted by Claude Code (AI agent) --- not written by a human.

[d-morrison]

Annotation dispositions for head 77f8701:

I read the annotations of all 14 passing runs on this head (the run ids are the check-run ids on 77f8701).

• Warning, false positive: "Node.js 20 is deprecated. The following actions target Node.js 20 but are being forced to run on Node.js 24", on the three check runs 131709035274, 131709034967 and 131709034594. It is GitHub's notice about the pinned versions of third-party actions in the shared workflows; this PR changes no workflow file (it edits macros.qmd, the two symbol tables and notation-choices.qmd), so nothing here can fix or cause it.
• Routine notices, grouped: "The ubuntu-latest label will migrate to Ubuntu 26 beginning October 19, 2026", one on each of the 14 runs: 131709756835, 131709756828, 131709663591, 131709217415, 131709156039, 131709117515, 131709036499, 131709036107, 131709035274, 131709034967, 131709034594, 131709026743, 131709026514 and 131709026168. A GitHub-wide runner-image notice, not tied to this diff.
• No failure annotations on this head.

Posted by Claude Code (AI agent) --- not written by a human.