Collaborative effort to fix R CMD Check

DESCRIPTION

@@ -6,7 +6,8 @@ Authors@R:c(person("Jan", "Meis", role = c("aut", "cre"), email = "meis@imbi.uni
   person("Martin", "Maechler", role = c("cph"), email = "maechler@stat.math.ethz.ch", comment = c(ORCID = "0000-0002-8685-9910", "Original author of monoSpl.c (from the 'stats' package).")))
 Description:
   Methods to evaluate the performance characteristics of
-  various point and interval estimators for optimal adaptive two-stage designs.
+  various point and interval estimators for optimal adaptive two-stage designs as described
+  in Meis et al. (2024) <doi:10.1002/sim.10020>.
   Specifically, this package is written to work with trial designs created by the 'adoptr' package
   (Kunzmann et al. (2021) <doi:10.18637/jss.v098.i09>; Pilz et al. (2021) <doi:10.1002/sim.8953>)).
   Apart from the a priori evaluation of performance characteristics, this package also allows for the
@@ -24,7 +25,8 @@ LazyData: true
 VignetteBuilder: knitr
 RoxygenNote: 7.2.3
 Depends:
-    R (>= 4.0.0)
+    R (>= 4.0.0),
+    adoptr
 Imports: 
     methods,
     stats,
@@ -36,7 +38,6 @@ Imports:
     latex2exp,
     forcats,
     future.apply,
-    pracma,
     progressr,
     Rdpack
 Suggests: 

NAMESPACE

@@ -29,7 +29,6 @@ export(NaiveCI)
 export(NaivePValue)
 export(NeymanPearsonOrderingCI)
 export(NeymanPearsonOrderingPValue)
-export(Normal)
 export(NormalPrior)
 export(OverestimationProbability)
 export(PValue)
@@ -43,7 +42,6 @@ export(ScoreTestOrderingPValue)
 export(SoftCoverage)
 export(StagewiseCombinationFunctionOrderingCI)
 export(StagewiseCombinationFunctionOrderingPValue)
-export(Student)
 export(TestAgreement)
 export(UniformPrior)
 export(Variance)
@@ -61,8 +59,8 @@ exportClasses(IntervalEstimator)
 exportClasses(PValue)
 exportClasses(PointEstimator)
 exportClasses(Statistic)
-exportClasses(TwoStageDesign)
 exportMethods(plot)
+import(adoptr)
 import(ggplot2)
 import(methods)
 importFrom(Rdpack,reprompt)
@@ -83,7 +81,6 @@ importFrom(ggpubr,theme_pubr)
 importFrom(grDevices,xy.coords)
 importFrom(graphics,plot.default)
 importFrom(latex2exp,TeX)
-importFrom(pracma,gaussLegendre)
 importFrom(progressr,progressor)
 importFrom(scales,percent)
 importFrom(stats,dchisq)

NEWS.md

@@ -1,3 +1,8 @@
+# adestr 0.5.2
+
+* Preparing new CRAN release: adoptr back on CRAN, adding the
+collaborative produced test cases and vignettes
+
 # adestr 0.5.1
 
 * Replaced raster graphics in vignettes with vector graphics.

R/adestr_package.R

@@ -11,7 +11,7 @@
 #'
 #' @docType package
 #' @name adestr
-#' @import methods
+#' @import methods adoptr
 #' @importFrom stats dnorm pnorm qnorm dt pt qt dchisq pchisq qchisq integrate uniroot var
 #' @importFrom cubature hcubature
 #' @importFrom Rdpack reprompt

R/fastmonoHFC.R

@@ -48,6 +48,9 @@ Myregularize.values <- function (x, y, ties, warn.collapsing = TRUE, na.rm = TRU
   }
   list(x = x, y = y, keptNA = keptNA, notNA = if (keptNA) ok)
 }
+
+# This is isn't used anywhere, but maybe someone needs it sometime
+#nocov start
 fastmonoH.FC_function <- function(x, y=NULL, ties = mean, extrapol = c("linear", "cubic")) {
   x <- Myregularize.values(x, y, ties, missing(ties))
   y <- x$y
@@ -104,6 +107,7 @@ fastmonoH.FC_function <- function(x, y=NULL, ties = mean, extrapol = c("linear",
     res
   }
 }
+#nocov end
 
 fastmonoH.FC_coefficients <- function(x, y=NULL, ties = mean, extrapol = c("linear", "cubic")){
   x <- Myregularize.values(x, y, ties, missing(ties))

R/fisher_information.R

@@ -1,3 +1,4 @@
+#nocov start
 fisher_information <- function(design, mu, sigma, two_armed) {
   sapply(mu, \(x)
     integrate_over_sample_space(
@@ -23,3 +24,4 @@ make_inverse_fisher_information_table <- function(designs, mu, sigma, two_armed)
   }
   do.call(rbind, res_list)
 }
+#nocov end

R/twostagedesign_with_cache.R

@@ -1,116 +1,3 @@
-### Remove some of this once adoptr is back on CRAN ###
-
-#' Re-export of two-stage design class
-#'
-#' This is a re-export of the \code{TwoStageDesign} class from the
-#' \code{adoptr} \insertCite{kunzmann2021adoptr}{adestr} package.
-#'
-#' This function is currently re-exported here to resolve CRAN conflicts.
-#' For details, please refer to the paper \insertCite{kunzmann2021adoptr}{adestr}
-#' and the original github repository \url{https://github.com/kkmann/adoptr}.
-#'
-#' @slot n1 (numeric) first-stage sample size.
-#' @slot c1f (numeric) first-stage futility boundary.
-#' @slot c1e (numeric) first-stage early efficacy boundary.
-#' @slot n2_pivots (numeric) vector containing the values of the n2 spline function.
-#' @slot c2_pivots (numeric) vector containing the values of the second-stage
-#' rejection boundary spline c2
-#' @slot x1_norm_pivots (numeric) vector containing the x-axis (z-sclae) points
-#' for the n2 and c2 splines
-#' @slot weights (numeric) vector containing integration weights
-#' @slot tunable (logical) vector determining whether desing paramters are to be optimized
-#'
-#' @seealso The original implementation of the adoptr package by Kevin Kunzmann and
-#' Maximilian Pilz is available at \url{https://github.com/kkmann/adoptr}.
-#'
-#' @exportClass TwoStageDesign
-#' @importFrom pracma gaussLegendre
-setClass("TwoStageDesign", representation(
-  n1        = "numeric",
-  c1f       = "numeric",
-  c1e       = "numeric",
-  n2_pivots = "numeric",
-  c2_pivots = "numeric",
-  x1_norm_pivots = "numeric",
-  weights   = "numeric",
-  tunable   = "logical"
-))
-TwoStageDesign <- function(n1, c1f, c1e, n2_pivots, c2_pivots, order = length(c2_pivots)) {
-  if (order != length(c2_pivots)) {
-    stop("order needs to be same length as c2_pivots")
-  }
-  if (length(n2_pivots) != length(c2_pivots) && length(n2_pivots)!=1) {
-    stop("n2_pivots needs to be the same length as c2_pivots or of length 1.")
-  }
-  if (length(n2_pivots)==1L)
-    n2_pivots <- rep(n2_pivots, order)
-  glr <- gaussLegendre(order, -1, 1)
-  tunable        <- c("n1" = TRUE, "c1f" = TRUE, "c1e" = TRUE, "n2_pivots" = TRUE, "c2_pivots" = TRUE,
-                      "x1_norm_pivots" = FALSE, "weights" = FALSE, "tunable" = FALSE)
-  new("TwoStageDesign", n1 = n1, c1f = c1f, c1e = c1e, n2_pivots = n2_pivots,
-      c2_pivots = c2_pivots, x1_norm_pivots = glr$x, weights = glr$w,
-      tunable = tunable)
-}
-setClass("DataDistribution", representation(
-  two_armed = "logical")
-)
-setClass("Normal", contains = "DataDistribution")
-#' Normally distributed data with known variance
-#'
-#' This function creates an object representing the distributional assumptions
-#' of the data: normally distributed outcomes sample from a trial with
-#' one or two arms (depending on the value of the parameter \code{two_armed}),
-#' under the assumption of known variance.
-#'
-#' @param two_armed (logical) determines whether one or two-armed trials are assumed.
-#'
-#' @export
-#' @returns an object of class \code{Normal}. This object encodes the distributional
-#' assumptions of the data for usage in the functions
-#' \code{\link{evaluate_estimator}} and \code{\link{analyze}}.
-#' @examples
-#' evaluate_estimator(
-#'   score = MSE(),
-#'   estimator = SampleMean(),
-#'   data_distribution = Normal(FALSE),
-#'   design = get_example_design(),
-#'   mu = c(0, 0.3, 0.6),
-#'   sigma = 1,
-#'   exact = FALSE
-#' )
-#'
-#'
-Normal <- function(two_armed = TRUE) new("Normal", two_armed = two_armed)
-setClass("Student", contains = "DataDistribution")
-
-#' Normally distributed data with unknown variance
-#'
-#' This function creates an object representing the distributional assumptions
-#' of the data: normally distributed outcomes sample from a trial with
-#' one or two arms (depending on the value of the parameter \code{two_armed}),
-#' under the assumption of known variance.
-#'
-#' @param two_armed (logical) determines whether one or two-armed trials are assumed.
-#' @returns an object of class \code{Student}. This object encodes the distributional
-#' assumptions of the data for usage in the functions
-#' \code{\link{evaluate_estimator}} and \code{\link{analyze}}.
-#'
-#' @export
-#' @examples
-#' evaluate_estimator(
-#'   score = MSE(),
-#'   estimator = SampleMean(),
-#'   data_distribution = Student(FALSE),
-#'   design = get_example_design(),
-#'   mu = c(0, 0.3, 0.6),
-#'   sigma = 1,
-#'   exact = FALSE
-#' )
-#'
-Student <- function(two_armed = TRUE) new("Student", two_armed = two_armed)
-n1 <- function(design, round = FALSE) if (round) round(design@n1) else design@n1
-### end of remove ###
-
 setClass(
   "TwoStageDesignWithCache",
   contains = "TwoStageDesign",

README.Rmd

@@ -34,6 +34,12 @@ Currently, it works for designs objects which were produced by the
 R-package `adoptr`, which calculates optimal design parameters adaptive
 two-stage designs.
 
+An introductory vignette covering common usecases is given at [https://jan-imbi.github.io/adestr/articles/Introduction.html](https://jan-imbi.github.io/adestr/articles/Introduction.html).
+
+The test [https://jan-imbi.github.io/adestr/articles/Introduction.html](https://jan-imbi.github.io/adestr/articles/Introduction.html).
+
+<!-- reference implementation verlinken -->
+
 ## Installation
 
 You can install the development version of adestr by typing
@@ -43,16 +49,7 @@ remotes::install_github("https://github.com/jan-imbi/adestr")
 ```
 into your R console.
 
-## Information for reviewers
-
-The scripts to reproduce the results from the paper can be found in the
-`/data/code/` directory of this repository. The results themselves are
-located in the `/data/` directory.
-
-The easiest way to inspect the results is to [clone this repository](https://docs.github.com/en/repositories/creating-and-managing-repositories/cloning-a-repository).
-
-
-## General example for usage of the package
+## Small introductory example
 
 Here is a quick example showing the capabilities of `adestr`.
 First, load `adestr`:
@@ -103,3 +100,8 @@ analyze(
  design = get_example_design()
 )
 ```
+
+Please refer to [https://jan-imbi.github.io/adestr/articles/Introduction.html](https://jan-imbi.github.io/adestr/articles/Introduction.html) for a more detailed introduction.
+
+
+

tests/testthat/setup_designs.R

@@ -1,7 +1,7 @@
 # adoptr designs used for testing
 designad <- get_example_design()
 ### Put GroupSequentialDesign( once adoptr is back on CRAN ###
-designgs <- TwoStageDesign(
+designgs <- GroupSequentialDesign(
   n1 = 29.53980042851903320411,
   c1f = 0.8563037186428685831885,
   c1e = 2.211178640465977007779,