which_paddock_row(), used internally by trace_asco() to place
primary_infection_foci in the paddock. Its row-lookup formula assumed
x varied fastest through paddock's rows, but data.table::CJ() (how
paddock is built) actually varies y fastest. As a result,
primary_infection_foci supplied as explicit, asymmetric x/y
coordinates (i.e. anything other than "centre" or a symmetric location)
could be seeded at the transposed coordinate instead of the one supplied.
which_paddock_row() now looks up rows with a data.table join instead
of a hand-rolled formula, so it no longer depends on assumptions about
paddock row order. This changes simulation output for runs that used
non-"centre", asymmetric primary_infection_foci coordinates -- results
from prior versions with such inputs should be treated with caution and
ideally re-run.newM_weather, the package's internal
test/example weather fixture, clarifying that it is not exported or
user-accessible.random_integer_from_real()),
stochastic tests are seeded and use tighter, statistically-justified
tolerances instead of loose bounds, and expect_is() has been migrated
to expect_type()/expect_s3_class() throughout.Clarification on input arguments in format_weather().
Typo fixes and other enhancements in documentation.
Add note in README pertaining to Rcpp errors that may occur if an older version of Rcpp is locally installed.
Initial CRAN release.
New R-package to simulate the spread of chickpea disease, Ascochyta blight.
NEWS.md file to track changes to the package.