CRAN resubmission; no changes to the API or behaviour.
getting-started vignette with a full walkthrough
of the pipeline on pollen_demo, covering gaps vs. missing values, the
classed gap error, ISO week/year boundaries, custom seasons, imputation
flags and min_obs, moving-average defaults, and within-group lags.stats (part of base R) is imported alongside cli
and rlang.Maintenance and packaging fixes; no user-facing API changes.
<user>
placeholder with the repository owner).*.Rcheck/, source tarball, README.html)
from version control and added them to .gitignore.Authors@R field and citation metadata (inst/CITATION,
CITATION.cff), removing a name suffix that was embedded in the family name.R CMD check now passes with no errors, warnings, or notes.First release.
complete_daily_grid() inserts rows for absent calendar days, within group.assign_iso_week() appends ISO 8601 week and year.assign_season() appends a configurable season label (meteorological,
astronomical, or custom pollen-season breaks).impute_weekly_mean() fills missing daily values with the ISO-week mean and
flags what it filled.build_moving_average() appends trailing (or centred) windowed means.apply_lag() appends time-indexed lags.pollen_demo: synthetic daily pollen counts from two sites over one spring
season, containing both gaps and missing values.Two decisions with cross-cutting consequences were made before implementation
and are recorded in inst/DESIGN-DECISIONS.md:
strftime()'s %V/%G, whose behaviour has varied by platform.
Output is identical on every operating system.apply_lag() and build_moving_average() raise an error on a
gapped daily grid rather than completing it silently. Call
complete_daily_grid() first.Every default in this package changes what an exposure variable means —
moving-average alignment, window inclusivity, min_obs, the season definition,
the treatment of gaps. Changing a default is therefore a breaking change to
the estimand, even when it is not a breaking change to the API, and will be
documented here as such.