NEWS

ukbflow 0.3.4 (2026-04-08)

Released: April 2026

Bug Fixes

All test-integration-*.R files now call skip_on_cran() at file scope, preventing integration tests from running on CRAN. Resolves CHECK ERROR on r-devel-linux-x86_64-fedora-clang and r-devel-linux-x86_64-fedora-gcc reported by CRAN on 2026-04-08.
RNGkind() is now explicitly set before set.seed() in integration tests that use simulated data, ensuring cross-platform RNG reproducibility.

ukbflow 0.3.3 (2026-04-07)

Released: April 2026

Bug Fixes

grs_validate() — added skip_if_not_installed("pROC") guard to all logistic-path tests so the suite passes cleanly when pROC is absent

Improvements

Removed hardcoded default paths across all modules; examples refactored to use ops_toy() data where possible — runnable examples unwrapped, long-running examples (>5s) wrapped in \donttest{}, network-dependent examples remain in \dontrun{}
VignetteBuilder field corrected in DESCRIPTION

CI

R CMD check split into four targeted workflows: dev (3 jobs), main (6 jobs), release (8 jobs, --as-cran), and weekly (scheduled every Saturday 21:00 CST, full CRAN-like matrix)
Added _R_CHECK_DEPENDS_ONLY_ matrix to verify Suggests packages are correctly declared

Documentation

Rd files regenerated across all modules after example and parameter updates

ukbflow 0.3.2

Released: March 26, 2026

Bug Fixes

ops_snapshot() — removed Unicode character Δ from documentation to fix LaTeX PDF manual build on CRAN
test-plot.R — added skip_on_cran() to plot_tableone() rendering tests that caused a 20-minute hang on Windows CRAN

Improvements

plot_forest() and plot_tableone() examples are now fully runnable (removed \dontrun{} wrapper; save = FALSE throughout)

Documentation

README.md — updated Codecov badge URL; fixed CONTRIBUTING.md and README_zh.md links to full GitHub URLs
Added inst/WORDLIST with Biobank to suppress false-positive spelling NOTE on CRAN

ukbflow 0.3.1

Released: March 25, 2026

Bug Fixes

derive_followup() — coerce date columns to IDate before pmin() to avoid type mismatch
install.Rmd — corrected vignette cross-references

Improvements

All modules hardened with consistent .assert_*() input validation helpers
All cli::cli_abort() calls now use call = NULL for cleaner error messages
plot_tableone() — auto-coerce data.table input to data.frame

Documentation

Man pages updated across all modules (auth, fetch, extract, job, decode, derive, ops, grs, assoc, plot)
All vignettes reviewed and corrected (output format, param contracts, example accuracy)
get-started.Rmd and README updated for accuracy
pkgdown reference index and vignette menu completed

Tests

Test suites overhauled for all modules: added input validation coverage, integration-style stubs, and edge cases

ukbflow 0.3.0

Released: March 13, 2026

New Features

Operations (ops_*)

ops_withdraw() — exclude UKB withdrawn participants from a cohort data.table by EID

Snapshot — Column Tracking

ops_snapshot() gains column-tracking helpers: cols(), diff(), remove(), and set_safe_cols()

Visualisation Enhancements

plot_tableone() — new png_scale, pdf_width, and pdf_height parameters for fine-grained output control

Bug Fixes

fetch_file() — enforce RAP-only guard; updated tests
grs_score() — fix -icmd argument format; skip script upload if file already exists on RAP
GRS pipeline — updated chr split threshold: chromosomes 1–16 use large instances, 17–22 use standard

Documentation

Added roxygen2 documentation for ops_withdraw()
Unit tests added for ops_withdraw()

Internal

Added broom to DESCRIPTION Imports and ops_setup() dependency check
Updated package logo (new hex sticker design)
Integration tests requiring RAP environment are now skipped in local CI

ukbflow 0.2.0

Released: March 10, 2026

New Features

Operations (ops_*)

ops_setup() — check and report the local environment (R, dx-toolkit, dxpy) health
ops_toy() — generate synthetic UKB-style cohort or forest-plot data for testing and demos; includes GRS columns and cancer self-report fields
ops_na() — summarise missing-value rates per column with threshold-based filtering and cli progress feedback
ops_snapshot() — record and display a history of dataset row/column counts across pipeline steps

Bug Fixes

All cli::cli_abort() calls now pass call = NULL to suppress internal call-stack noise in error messages
ops_toy(): added cancer self-report fields (p20001, p20006) and corrected sr_codes → text label mapping

Documentation

New vignette: Smoking and Lung Cancer — End-to-End Analysis (vignette("smoking_lung_cancer"))
New vignette: ops_ Series* covering setup, toy data, NA summary, and snapshots
pkgdown site now auto-deploys via GitHub Actions (CI-managed docs/)

Internal

Resolved R CMD check NOTEs: added importFrom(stats, rnorm, runif), importFrom(utils, object.size), and pct_na to globalVariables()
Added builds/ to .Rbuildignore

ukbflow 0.1.0

Released: March 6, 2026

Initial release of ukbflow — a RAP-native R workflow for UK Biobank analysis.

New Features

Connection

auth_login() — authenticate to RAP via dx-toolkit token
auth_logout() — revoke current session
auth_status() — check current login state
auth_list_projects() — list accessible RAP projects
auth_select_project() — set active RAP project

Data Access

fetch_ls() / fetch_tree() — browse RAP project file structure
fetch_file() — download files from RAP to local
fetch_url() — generate pre-signed download URLs
fetch_metadata() — retrieve UKB field metadata (field.tsv, encoding.tsv)
fetch_field() — download the UKB field dictionary file
extract_ls() — list available UKB datasets on RAP
extract_pheno() — synchronously extract phenotype fields from a RAP dataset
extract_batch() — submit a DNAnexus table-exporter job to extract phenotype fields
job_wait() — poll job status until completion
job_status() / job_ls() / job_path() / job_result() — monitor and locate RAP jobs

Data Processing

decode_values() — convert integer codes to human-readable labels
decode_names() — rename p{field}_i{instance}_a{array} columns to descriptive names
derive_missing() — recode UKB informative-missing labels to NA
derive_covariate() — standardise common covariates (age, BMI, TDI, etc.)
derive_cut() — create ordered factor variables from numeric cutpoints
derive_hes() — derive disease phenotypes from Hospital Episode Statistics (ICD-10)
derive_cancer_registry() — derive cancer phenotypes from UK cancer registry (ICD-10)
derive_death_registry() — derive phenotypes from death registry (primary + secondary causes)
derive_first_occurrence() — derive phenotypes from UKB First Occurrence fields
derive_selfreport() — derive disease phenotypes from self-reported illness codes
derive_icd10() — merge multi-source ICD-10 case definitions across all registers
derive_case() — merge arbitrary multi-source case definitions
derive_followup() — compute follow-up time with competing event support
derive_timing() — classify prevalent vs. incident cases
derive_age() — compute age at event

Association Analysis

assoc_linear() — linear regression with automatic three-model framework
assoc_logistic() — logistic regression with automatic three-model framework
assoc_coxph() — Cox proportional hazards with automatic three-model framework
assoc_coxph_zph() — test proportional hazards assumption
assoc_competing() — Fine-Gray competing risks regression
assoc_subgroup() — subgroup analysis with interaction likelihood-ratio test
assoc_trend() — dose-response trend analysis across ordered categories
assoc_lag() — landmark / lag-time sensitivity analysis

Genomic Risk Score (GRS)

grs_check() — validate and reformat a GWAS summary statistics weight file
grs_bgen2pgen() — submit parallel BGEN → PGEN conversion jobs on RAP
grs_score() — submit distributed plink2 GRS scoring jobs on RAP
grs_standardize() — standardise GRS columns to mean = 0, SD = 1
grs_validate() — validate GRS distribution and association with known risk factors

Visualization

plot_forest() — publication-ready forest plots with customisable CI columns and p-value formatting
plot_tableone() — Table 1 baseline characteristics figure