Sample size em edits

[d-morrison]

Thank you for taking the time to submit a pull request!

To maximize the chances of acceptance:

• The title of your PR should briefly describe the change.
• The body of your PR should contain Fixes #issue-number (if relevant).
• Commit/merge messages to be included in NEWS.md should begin with -.
• Code should follow the tidyverse style guide.
• Documentation should use roxygen2, with Markdown syntax.
• Contributions should include unit tests (using testthat).

For more information see Contributing.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ All tests successful. No failed tests found.

[Copilot]

Pull request overview

This PR adds a new data-raw/ script to run parallelized seroincidence simulations for exploring sample-size behavior, and updates package metadata to include additional optional dependencies used by supporting workflows/scripts.

Changes:

• Add data-raw/sample-size.R to simulate cross-sectional serology data and summarize CI coverage across scenarios.
• Add furrr and svglite to Suggests in DESCRIPTION (for optional parallel mapping and SVG-based visual regression testing support).

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 6 comments.

File	Description
DESCRIPTION	Adds optional dependencies (furrr, svglite) needed by scripts/testing tools.
data-raw/sample-size.R	New simulation + summarization script for evaluating sample size / CI coverage under different incidence rates.

[d-morrison]

Addressed the review findings in data-raw/sample-size.R:

• Critical/deprecated API: replaced load_curve_params() → load_sr_params(), sim.cs() → sim_pop_data() (renamed n.smpl/age.rng/n.mc/renew.params/add.noise → n_samples/age_range/n_mcmc_samples/renew_params/add_noise), and est.incidence(curve_params=) → est_seroincidence(sr_params=) (fixes the missing-sr_params runtime error).
• snake_case: renamed results_100_0.1/results_100_0.01/final_table_100_0.1/ft_100_0.01 to underscore forms and updated all references.
• Logic fix: the λ=0.01 scenario now actually passes lambda = 0.01.
• Reproducibility: added set.seed() before the two unseeded runs.
• Minor: seq_along() instead of 1:length(), moved library(future) to the top, and declared future in Suggests.

This is an internal maintainer script in data-raw/ (not user-visible / not installed), so I applied the no changelog + no version increment labels for the version-check and changelog gates, and added gh to WORDLIST to fix spellcheck.

@claude review

---

Generated by Claude Code

[github-actions]

👀 Picked up by workflow run #26910326827. R setup runs first; Claude itself responds after that.

[d-morrison]

@claude review

---

Generated by Claude Code

[github-actions]

👀 Picked up by workflow run #26911728637. R setup runs first; Claude itself responds after that.

[d-morrison]

Working on this — paws off until I'm done.

---

Generated by Claude Code

[d-morrison]

ARD summary

This PR adds a maintainer-only simulation script (data-raw/sample-size.R) for exploring CI coverage of est_seroincidence() across sample sizes/incidence rates, plus furrr/future/svglite in Suggests. All findings from the original Copilot review (deprecated API calls, snake_case violations, a mislabeled λ scenario, missing seeds) were already Addressed in earlier commits on this branch and confirmed resolved by a follow-up Claude review on 2026-06-03. That review left two minor open items, both now addressed in this session:

Finding	Disposition
1:n_sim in furrr::future_map() is fragile when n_sim = 0 (reverse sequence)	Addressed — changed to seq_len(n_sim)
λ=0.1 / λ=0.01 coverage results computed but discarded; paste() called on a data frame produced an unreadable message	Addressed — all four coverage tables (n=50, n=100 λ=0.2/0.1/0.01) are now assigned and printed consistently via cat() + print()

Additionally, since this PR was opened in Feb 2025 and main had drifted 59 commits ahead:

• Merged main into this branch. One real conflict, in DESCRIPTION's Version: field — resolved by taking main's value.
• That left the dev version equal to (not greater than) main's, which fails the version-check CI gate, so bumped to 1.4.1.9001.

Pushed as commits 55b46060 (merge), a9b3ff7a (nits), 9ff9ed2c (version bump).

CI status: still running/pending as of last check — not yet confirmed fully green. Will need a follow-up check before this is mergeable.

---

Generated by Claude Code

[d-morrison]

Done for now — unclaimed.

---

Generated by Claude Code

[d-morrison]

Working on this (resolving the conflict with main, addressing the two remaining nits, and tidying stale threads) --- paws off until I'm done.

---

Generated by Claude Code

[d-morrison]

Merged main into the branch and drove this to a clean state:

• Resolved the main conflict (DESCRIPTION Version: only) and bumped the dev version to 1.4.1.9005 so it stays above main's 1.4.1.9004 (version-check parity).
• Addressed the two remaining review nits in data-raw/sample-size.R: corrected the stale # Loop through each of the 100 results comment (the loop is sample-size-agnostic), and renamed the lambda() helper to get_lambda() so it no longer shadows the lambda value used throughout the script.
• Resolved all 6 stale inline threads from the original Copilot review --- the deprecated-API and dotted-object-name items were all superseded by later commits (object names already use underscores; lint CI is green).

Review verdict was already clean. Leaving the merge decision to a human.

---

Generated by Claude Code

[d-morrison]

Heads-up: after the merge + version bump I pushed, lint-changed-files will fail on data-raw/sample-size.R --- but that's the repo-wide lintr 3.4.0 indentation_linter regression (#558), not a defect in this PR. This PR is otherwise clean and is blocked pending the lint fix in #558.

---

Generated by Claude Code

[claude]

Code Review (confirming pass)

• Diff current HEAD (888b1617) against the last-reviewed commit (a9b3ff7, prior verdict)
• Verify no regressions were introduced by the main merge
• Re-check function signatures cited in the script against current R/ source (hallucination check)
• Confirm CI status

---

Summary

The only changes since the 2026-06-30 review are a merge of main into the branch and a DESCRIPTION version bump (1.4.1.9004 → 1.4.1.9005) for version-check parity. data-raw/sample-size.R itself was also improved beyond that review: both optional nits it raised are now fixed —

• the stale # Loop through each of the 100 results comment is corrected to describe the loop generically, and
• the lambda() helper (which shadowed the lambda value used elsewhere) is renamed to get_lambda().

That prior verdict — all critical/blocking findings resolved, script in good shape — still stands, and the remaining nits from that round are now also closed out.

Independently re-verified for this pass (not just trusting changelog):

• est_seroincidence(pop_data=, sr_params=, noise_params=, lambda_start=, build_graph=, verbose=, antigen_isos=, ...) — matches current signature in R/est_seroincidence.R.
• sim_pop_data(lambda=, n_samples=, age_range=, antigen_isos=, n_mcmc_samples=, renew_params=, add_noise=, curve_params=, noise_limits=, format=, ...) — matches current signature in R/sim_pop_data.R.
• load_sr_params(file_path, antigen_isos = NULL) — matches current signature in R/load_sr_params.R.
• No dotted object names remain; all identifiers are valid snake_case.
• set.seed() precedes all four simulate_seroincidence() calls; seq_len()/seq_along() used throughout (no reverse-sequence risk).
• furrr, future, svglite correctly declared in Suggests; data-raw/ remains .Rbuildignored.
• PR carries the no changelog label, so the missing NEWS.md bullet is expected/exempt.

CI status: All required checks are green — lint-changed-files, Spellcheck, version-check, docs-check, Check Changelog Action, test-coverage, and all R-CMD-check matrix jobs (ubuntu/macos/windows release) are ✅. (Note: an earlier comment on this thread flagged a possible lint-changed-files failure from a repo-wide lintr regression — that check is currently passing, so it's not blocking.)

No new issues found.

Verdict

Ready for merge

[github-actions]

Claude finished review — View run

---

Posted a confirming review on PR #369: comment.

Summary: Since the last review round (2026-06-30), the only changes are a main merge and a version bump — the two optional nits from that review (stale comment, lambda() shadowing) have since been fixed by the author. I independently re-verified the est_seroincidence(), sim_pop_data(), and load_sr_params() calls against current R/ source (no hallucinated APIs), confirmed no dotted object names remain, and checked that all CI checks are currently green.

Verdict: Ready for merge

[github-actions]

💰 Cost: $1.0235 (review) — run

[github-actions]

📖 https://UCD-SERG.github.io/serocalculator/preview/pr369/
Preview documentation for this PR (at commit 888b161)