Merge branch 'master' into 0.2.0

R/CompositeScore.R

@@ -29,7 +29,7 @@ setClass("CompositeConditionalScore",
 #' @template dotdotdot
 #'
 #' @return an object of class \code{CompositeConditionalScore} or
-#'   \code{CompositeUnconditionalScore} depending of the class of the scores used
+#'   \code{CompositeUnconditionalScore} depending on the class of the scores used
 #'   in \code{expr}
 #'
 #' @seealso \link{Scores}
@@ -41,7 +41,7 @@ setClass("CompositeConditionalScore",
 #' # linear combination:
 #' composite({ess - 50*power})
 #'
-#' # control flow
+#' # control flow (e.g. for and while loops)
 #' composite({
 #'   res <- 0
 #'   for (i in 1:3) {

R/adoptr.R

@@ -1,9 +1,9 @@
 #' Adaptive Optimal Two-Stage Designs
 #'
 #' The \pkg{adoptr} package provides functionality to explore custom optimal
-#' two-stage designs for one or two-arm superiority tests.
+#' two-stage designs for one- or two-arm superiority tests.
 #' Currently, only (asymptotically) normal test statistics are supported.
-#' \pkg{adoptr} is makes heavy use of the S4 class system.
+#' \pkg{adoptr} makes heavy use of the S4 class system.
 #' A good place to start learning about it can be found
 #' \href{http://adv-r.had.co.nz/OO-essentials.html}{here}.
 #'

tests/testthat/test_ConditionalScores.R

@@ -1,11 +1,15 @@
 context("test sample size")
 
+
+
+
 design <- TwoStageDesign(25, 0, 2, rep(40.0, 5), rep( 2, 5), 5L)
 dist   <- Normal()
 prior  <- ContinuousPrior(function(x) rep(1/10, length(x)), c(-4, 6))
 z1     <- seq(-1, 3, .1)
 
 
+
 test_that("conditional sample size maps to actual sample size", {
 
     css  <- ConditionalSampleSize()
@@ -14,44 +18,44 @@ test_that("conditional sample size maps to actual sample size", {
 
     expect_equal(
         evaluate(css, design, z1),
-        n(design, z1)
-    )
+        n(design, z1),
+        tolerance = 1e-4, scale = 1)
 
     expect_equal(
         evaluate(ess1, design),
-        evaluate(ess2, design)
-    )
+        evaluate(ess2, design),
+        tolerance = 1e-4, scale = 1)
 
 
 }) # end 'conditional sample size maps to actual sample size'
 
 
 
+
+
 context("Test Conditional Power")
 
+
+
 test_that("Conditional Power is monotonous", {
 
     # Define two simple designs
-    design1 <- TwoStageDesign(25, 0, 2, rep(40.0, 5), seq(2.0, 0.0, length.out = 5))
-    design2 <- TwoStageDesign(25, 0, 2, rep(60.0, 5), seq(2.0, 0.0, length.out = 5))
-
+    design1     <- TwoStageDesign(25, 0, 2, rep(40.0, 5), seq(2.0, 0.0, length.out = 5))
+    design2     <- TwoStageDesign(25, 0, 2, rep(60.0, 5), seq(2.0, 0.0, length.out = 5))
     dist        <- Normal()
     null        <- PointMassPrior(.0, 1)
     alternative <- PointMassPrior(.4, 1)
-
-    cp <- ConditionalPower(dist, alternative)
+    cp          <- ConditionalPower(dist, alternative)
 
     # Conditional Power is monotonously increasing in n
     expect_gt(
         evaluate(cp, design2, 1),
-        evaluate(cp, design1, 1)
-    )
+        evaluate(cp, design1, 1))
 
     # Conditional Power is monotonously increasing in z1
     expect_gt(
         evaluate(cp, design1, 1.5),
-        evaluate(cp, design1, 0.5)
-    )
+        evaluate(cp, design1, 0.5))
 
 }) # end "Conditional Power is monotonous"
 
@@ -64,13 +68,14 @@ test_that("Conditional power has correct values outside continuation region",{
 
     expect_equal(
         evaluate(cp, design1, -1),
-        0.0
-    )
+        0.0,
+        tolerance = 1e-4, scale = 1)
 
     expect_equal(
         evaluate(cp, design1, 3),
-        1.0
-    )
+        1.0,
+        tolerance = 1e-4, scale = 1)
+
 }) # end 'Conditional power has correct values outside continuation region'
 
 
@@ -82,14 +87,19 @@ test_that("Power works",{
 
     expect_equal(
         evaluate(pow1, design),
-        evaluate(pow2, design)
-    )
+        evaluate(pow2, design),
+        tolerance = 1e-4, scale = 1)
 
 }) # end 'Power works'
 
 
+
+
+
 context("Test class ConditionalScore")
 
+
+
 test_that("Conditional scores are vectorized in z1", {
 
     css    <- ConditionalSampleSize()