#devtools::test("dae") context("canonical") cat("#### Test for source formation\n") test_that("sources", { skip_on_cran() library(dae) #((A*B)/C)*D ABCD.lay <- fac.gen(list(A = 3, B = 3, C = 3, D = 3)) ABCD.struct <- pstructure(~ ((A*B)/C)*D, labels = "sources", data =ABCD.lay) ABCD.dat <- as.data.frame(ABCD.struct) testthat::expect_equal(names(ABCD.struct$Q), ABCD.dat$sources) testthat::expect_equal(ABCD.dat$sources, c("A","B","A#B","C[A:B]","D","A#D","B#D","A#B#D","C#D[A:B]")) ABCD.struct <- pstructure(~ ((A*B)/C)*D, omit.projectors = TRUE, labels = "terms", data =ABCD.lay) ABCD.dat <- as.data.frame(ABCD.struct) testthat::expect_equal(ABCD.dat$df, c(2,2,4,18,2,4,4,8,36)) #Generalized factor crossed with others genfac1.lay <- fac.gen(list(S = 3, A = 3, D = 3)) gf1.struct <- pstructure(~ (S:A)*D, labels = "sources", data =genfac1.lay) testthat::expect_equal(unname(gf1.struct$sources), c("S:A","D","(S:A)#D")) #Generalized factor nested within another factor gf2.struct <- pstructure(~ D/(S:A), labels = "sources", data =genfac1.lay) testthat::expect_equal(unname(gf2.struct$sources), c("D","S:A[D]")) #Generalized factor nesting another factor gf3.struct <- pstructure(~ (S:A)/D, labels = "sources", data =genfac1.lay) testthat::expect_equal(unname(gf3.struct$sources), c("S:A","D[S:A]")) #Two generalized factor nested within another and each other genfac4.lay <- fac.gen(list(S = 3, A = 3, D = 3, B = 3, C = 3)) gf4.struct <- pstructure(~ D/(S:A)/(B:C), labels = "sources", data =genfac4.lay) testthat::expect_equal(unname(gf4.struct$sources), c("D","S:A[D]","B:C[D:S:A]")) #Two generalized factor nested within another and each other - add grand mean gf4gm.struct <- pstructure(~ D/(S:A)/(B:C), labels = "sources", grandMean = TRUE, data =genfac4.lay) testthat::expect_equal(unname(gf4gm.struct$sources), c("Mean", "D","S:A[D]","B:C[D:S:A]")) testthat::expect_equal(nrow(as.data.frame(gf4gm.struct, omit.marginality = TRUE)), 4) #Single term one.struct <- pstructure(~ D, labels = "sources", data =genfac4.lay) #Single term with grand mean one.struct <- pstructure(~ D, labels = "sources", grandMean = TRUE, data =genfac4.lay) testthat::expect_equal(unname(one.struct$terms), c("Mean", "D")) testthat::expect_equal(unname(one.struct$sources), c("Mean", "D")) testthat::expect_equal(nrow(as.data.frame(one.struct, omit.marginality = TRUE)), 2) }) cat("#### Test for designAnatomy for Thao designs\n") test_that("Thao", { skip_on_cran() library(dae) Br <- 4 Bc <- 4 Sr<-4 Sc<-4 n <- Sr*Sc*Br*Bc v<- 4 s<-4 ##### Poset 2f design 2 # Factor D and E are latinised using LCCD. LS1 <- factor(c(1:4, 4:1, 2,1,4,3,3,4,1,2 )) LS2 <- factor(c(1:4,2,1,4,3,3,4,1,2,4,3,2,1 )) LS3 <- factor(c(1:4, 3,4,1,2, 4:1, 2,1,4,3)) LS4 <- factor(c(1:4, 2,1,4,3, 3,4,1,2, 4:1)) LS4.CC4x4.2f.ran <- data.frame( A = factor(c(rep(c(1,2,1,2 ,2,1,2,1 ,1,2,1,2 ,2,1,2,1), each=16))), B = factor(c(rep(c(1,1,2,2 ,1,1,2,2 ,2,2,1,1 ,2,2,1,1), each=16))), D = factor(c(rep(LS1[1:4],each=4), rep(LS1[5:8],each=4),rep(LS1[9:12],each=4),rep(LS1[13:16],each=4) ,rep(LS2[1:4],each=4), rep(LS2[5:8],each=4),rep(LS2[9:12],each=4),rep(LS2[13:16],each=4) ,rep(LS3[1:4],each=4), rep(LS3[5:8],each=4),rep(LS3[9:12],each=4),rep(LS3[13:16],each=4) ,rep(LS4[1:4],each=4), rep(LS4[5:8],each=4),rep(LS4[9:12],each=4),rep(LS4[13:16],each=4))), E = factor(c(rep(LS1[1:4],times=4),rep(LS2[1:4],times=4),rep(LS3[1:4],times=4),rep(LS4[1:4],times=4) ,rep(LS1[5:8],times=4), rep(LS2[5:8],times=4), rep(LS3[5:8],times=4),rep(LS4[5:8],times=4) ,rep(LS1[9:12],times=4),rep(LS2[9:12],times=4),rep(LS3[9:12],times=4),rep(LS4[9:12],times=4) ,rep(LS1[13:16],times=4),rep(LS2[13:16],times=4),rep(LS3[13:16],times=4) ,rep(LS4[13:16],times=4)))) #generate layout LS4.CC4x4.2f.lay <- designRandomize(allocated=LS4.CC4x4.2f.ran, recipient = list(BigRows= Br, BigColumns=Bc, Rows=Sr, Columns=Sc), nested.recipients = list(Columns="BigColumns", Rows="BigRows"), seed = 550) #Compute anatomy LS4.CC4x4.2f.canon <- designAnatomy(list(unit = ~ (BigRows/Rows)*(BigColumns/Columns), trt = ~ A*B*D*E), data = LS4.CC4x4.2f.lay) testthat::expect_true(all(LS4.CC4x4.2f.canon$sources$unit == c("BigRows", "Rows[BigRows]", "BigColumns", "Columns[BigColumns]", "BigRows#BigColumns", "BigRows#Columns[BigColumns]", "Rows#BigColumns[BigRows]", "Rows#Columns[BigRows:BigColumns]"))) summary(LS4.CC4x4.2f.canon, which.criteria = c("aeff", "xeff", "eeff", "order")) #### Poset 3c, design 1 - an N-poset #Set up allocated factors LS4.CC4x4.3c.ran <- cbind(data.frame(A = factor(rep(c(1,1,2,2 ,1,1,2,2 ,2,2,1,1 ,2,2,1,1), each = 16)), B = factor(rep(c(1,2,1,2 ,2,1,2,1 ,1,2,1,2 ,2,1,2,1), each = 16))), fac.gen(list(D = 4, E = 4), times = 16)) #generate layout and analyze LS4.CC4x4.3c.lay <- designRandomize(allocated=LS4.CC4x4.3c.ran, recipient = list(BigRows= Br, BigColumns=Bc, Rows=Sr, Columns=Sc), nested.recipients = list(Rows="BigRows", Columns=c("BigRows","BigColumns")), seed = 550) #Compute anatomy LS4.CC4x4.3c.canon <- designAnatomy(list(unit = ~ (BigRows/Rows)*BigColumns +((BigColumns*BigRows)/Columns)/Rows, trt = ~ A*B*D*E), data =LS4.CC4x4.3c.lay) testthat::expect_true(all(LS4.CC4x4.3c.canon$sources$unit == c("BigRows", "Rows[BigRows]", "BigColumns", "BigRows#BigColumns", "Rows#BigColumns[BigRows]", "Columns[BigRows:BigColumns]", "Rows#Columns[BigRows:BigColumns]"))) summary(LS4.CC4x4.3c.canon, which.criteria = c("aeff", "order")) ### Poset 4a LS4x4.CC5x5.4a.ran <- cbind(data.frame(A = factor(rep(c(1,2,1,2 ,2,1,2,1 ,1,2,1,2 ,2,1,2,1), each = 16)), B = factor(rep(c(1,1,2,2 ,1,1,2,2 ,2,2,1,1 ,2,2,1,1), each = 16))), fac.gen(list(D = 4, E = 4), times = 16)) #generate layout and analyze LS4x4.CC5x5.4a.lay <- designRandomize(allocated=LS4x4.CC5x5.4a.ran, recipient = list(BigRows= Br, BigColumns=Bc, Rows=Sr, Columns=Sc), nested.recipients = list(Columns=c("BigRows","BigColumns"), Rows=c("BigRows","BigColumns")), seed = 690) #Compute anatomy LS4x4.CC5x5.4a.lcanon <- designAnatomy(list(unit = ~ (BigRows*BigColumns)/(Rows*Columns), trt = ~ A*B*D*E), data =LS4x4.CC5x5.4a.lay) testthat::expect_true(all(LS4x4.CC5x5.4a.lcanon$sources$unit == c("BigRows", "BigColumns", "BigRows#BigColumns", "Rows[BigRows:BigColumns]", "Columns[BigRows:BigColumns]", "Rows#Columns[BigRows:BigColumns]"))) summary(LS4x4.CC5x5.4a.lcanon, which.criteria = c("aeff", "order")) }) cat("#### Test for designAnatomy with sources and marginality using Cochran&Cox PBIBD2\n") test_that("PBIBD2_sources", { skip_on_cran() library(dae) #'# PBIBD(2) from p. 379 of Cochran and Cox (1957) Experimental Designs. 2nd edn Wiley, New York" #'## Input the design and randomize" Treatments <- factor(c(1,4,2,5, 2,5,3,6, 3,6,1,4, 4,1,5,2, 5,2,6,3, 6,3,4,1)) PBIBD2.lay <- designRandomize(allocated = Treatments, recipient = list(Blocks =6, Units = 4), nested.recipients = list(Units = "Blocks"), seed = 98177) #'## Test that projs.canon has been deprecated testthat::expect_warning(PBIBD2.canon <- projs.canon(formulae = list(unit = ~Blocks/Units, trt = ~ Treatments), which.criteria = c('aeff', 'xeff', 'eeff','order'), labels = "sources", data = PBIBD2.lay)) #'##By differencing PBIBD2D.canon <- designAnatomy(formulae = list(unit = ~Blocks/Units, trt = ~ Treatments), which.criteria = c('aeff', 'xeff', 'eeff','order'), labels = "sources", orthogonalize = "diff", data = PBIBD2.lay) summary(PBIBD2D.canon, which.criteria = c('aeff', 'xeff', 'eeff','order')) testthat::expect_equal(PBIBD2D.canon$Q[[1]]$Blocks$Treatments$adjusted$aefficiency, 0.25) testthat::expect_lt(abs(PBIBD2D.canon$Q[[1]]$'Units[Blocks]'$Treatments$adjusted$aefficiency - 0.8824), 1e-04) testthat::expect_equal(PBIBD2D.canon$Q[[2]]$'Blocks&Treatments', 2) testthat::expect_equal(PBIBD2D.canon$Q[[2]]$'Blocks&Residual', 3) testthat::expect_equal(PBIBD2D.canon$Q[[2]]$'Units[Blocks]&Treatments', 5) testthat::expect_equal(PBIBD2D.canon$Q[[2]]$'Units[Blocks]&Residual', 13) testthat::expect_lt(abs(PBIBD2D.canon$Q[[1]]$'Units[Blocks]'$Treatments$adjusted$aefficiency - 0.8824), 1e-04) testthat::expect_null(PBIBD2D.canon$aliasing$unit) testthat::expect_null(PBIBD2D.canon$aliasing$trt) PBIBD2.marg <- PBIBD2D.canon$marginality #'##By differencing with grand mean PBIBD2Dg.canon <- designAnatomy(formulae = list(unit = ~Blocks/Units, trt = ~ Treatments), grandMean = TRUE, labels = "sources", orthogonalize = "diff", which.criteria = c('aeff', 'xeff', 'eeff','order'), data = PBIBD2.lay) summary(PBIBD2Dg.canon, which.criteria = c('aeff', 'xeff', 'eeff','order')) testthat::expect_equal(PBIBD2Dg.canon$Q[[2]]$`Mean&Mean`, 1) #'##By eigenmethods PBIBD2E.canon <- designAnatomy(formulae = list(unit = ~Blocks/Units, trt = ~ Treatments), which.criteria = c('aeff', 'xeff', 'eeff','order'), labels = "sources", orthogonalize = "eigen", data = PBIBD2.lay) summary(PBIBD2E.canon, which.criteria = c('aeff', 'xeff', 'eeff','order')) testthat::expect_true(attr(PBIBD2E.canon, which = "labels") == "terms") testthat::expect_lt(abs(PBIBD2E.canon$Q[[1]]$'Blocks:Units'$Treatments$adjusted$aefficiency - 0.8824), 1e-04) #'##By eigenmethods with supplied marginality PBIBD2Em.canon <- designAnatomy(formulae = list(unit = ~Blocks/Units, trt = ~ Treatments), marginality = PBIBD2D.canon$marginality, which.criteria = c('aeff', 'xeff', 'eeff','order'), labels = "sources", orthogonalize = "eigen", data = PBIBD2.lay) summary(PBIBD2Em.canon, which.criteria = c('aeff', 'xeff', 'eeff','order')) testthat::expect_equal(length(PBIBD2Em.canon$marginality), 2) testthat::expect_true(attr(PBIBD2Em.canon, which = "labels") == "sources") #marginality list with trt NULL - same as supplying just unit marginality PBIBD2Em.canon <- designAnatomy(formulae = list(unit = ~Blocks/Units, trt = ~ Treatments), marginality = PBIBD2.marg, which.criteria = c('aeff', 'xeff', 'eeff','order'), labels = "sources", orthogonalize = "eigen", data = PBIBD2.lay) summary(PBIBD2Em.canon, which.criteria = c('aeff', 'xeff', 'eeff','order')) testthat::expect_equal(length(PBIBD2Em.canon$marginality), 2) testthat::expect_true(attr(PBIBD2Em.canon, which = "labels") == "sources") #full marginality list PBIBD2Em.canon <- designAnatomy(formulae = list(unit = ~Blocks/Units, trt = ~ Treatments), marginality = PBIBD2.marg, which.criteria = c('aeff', 'xeff', 'eeff','order'), labels = "sources", orthogonalize = "eigen", data = PBIBD2.lay) summary(PBIBD2Em.canon, which.criteria = c('aeff', 'xeff', 'eeff','order')) testthat::expect_equal(length(PBIBD2Em.canon$marginality), 2) testthat::expect_true(attr(PBIBD2Em.canon, which = "labels") == "sources") testthat::expect_true(all(names(PBIBD2Em.canon$marginality) == names(PBIBD2.marg))) testthat::expect_lt(abs(PBIBD2Em.canon$Q[[1]]$'Units[Blocks]'$Treatments$adjusted$aefficiency - 0.8824), 1e-04) #Add grand mean term PBIBD2Egm.canon <- designAnatomy(formulae = list(unit = ~Blocks/Units, trt = ~ Treatments), grandMean = TRUE, marginality = PBIBD2.marg, which.criteria = c('aeff', 'xeff', 'eeff','order'), labels = "sources", orthogonalize = "eigen", data = PBIBD2.lay) summary(PBIBD2Egm.canon, which.criteria = c('aeff', 'xeff', 'eeff','order')) testthat::expect_equal(length(PBIBD2Em.canon$marginality), 2) testthat::expect_true(attr(PBIBD2Em.canon, which = "labels") == "sources") testthat::expect_true(all(PBIBD2Egm.canon$terms$unit == c("Mean", "Blocks", "Blocks:Units"))) testthat::expect_true(all(PBIBD2Egm.canon$terms$trt == c("Mean", "Treatments"))) testthat::expect_true(all(names(PBIBD2Em.canon$marginality) == names(PBIBD2.marg))) testthat::expect_lt(abs(PBIBD2Em.canon$Q[[1]]$'Units[Blocks]'$Treatments$adjusted$aefficiency - 0.8824), 1e-04) #'## unique plot levels PBIBD2.lay$AUnits <- with(PBIBD2.lay, fac.combine(list(Blocks,Units))) PBIBD2A.canon <- designAnatomy(list(unit = ~Blocks + AUnits, trt = ~ Treatments), data = PBIBD2.lay, labels = "sources", which.criteria = c('aeff', 'xeff', 'eeff','order'), orthogonalize = "hybrid") summary(PBIBD2A.canon, which.criteria = c('aeff', 'xeff', 'eeff','order')) testthat::expect_lt(abs(PBIBD2A.canon$Q[[1]]$'AUnits[Blocks]'$Treatments$adjusted$aefficiency - 0.8824), 1e-04) }) cat("#### Test for Jarrett & Ruggiero example\n") test_that("JarrettRuggiero", { skip_on_cran() library(dae) #'## Jarrett & Ruggiero example jr.lay <- fac.gen(list(Set=7, Dye=2, Array=3)) jr.lay <- within(jr.lay, { Block <- factor(rep(1:7, each=6)) Plant <- factor(rep(c(1,2,3,2,3,1), times=7)) Sample <- factor(c(rep(c(2,1,2,2,1,1, 1,2,1,1,2,2), times=3), 2,1,2,2,1,1)) S1 <- Dye Treat <- factor(c(1,2,4,2,4,1, 2,3,5,3,5,2, 3,4,6,4,6,3, 4,5,7,5,7,4, 5,6,1,6,1,5, 6,7,2,7,2,6, 7,1,3,1,3,7), labels=c("A","B","C","D","E","F","G")) }) array.plot.trt.canon <- designAnatomy(formulae = list(array = ~ (Set:Array)*Dye, plot = ~ Block/Plant/Sample, trt = ~ Treat), labels = "sources", data = jr.lay) testthat::expect_equal(length(array.plot.trt.canon$Q[[3]]), 7) testthat::expect_equal(array.plot.trt.canon$Q[[3]]$`(Set:Array)#Dye&Sample[Block:Plant]`, 6) testthat::expect_lt(abs(array.plot.trt.canon$Q[[2]]$`(Set:Array)#Dye&Plant[Block]`$Treat$adjusted$aefficiency - 0.58333333), 1e-05) testthat::expect_lt(abs(array.plot.trt.canon$Q[[1]]$`(Set:Array)#Dye`$`Plant[Block]`$adjusted$aefficiency - 0.75), 1e-05) summ.default <-summary(array.plot.trt.canon) testthat::expect_equal(nrow(summ.default$decomp), 7) testthat::expect_equal(ncol(summ.default$decomp), 9) testthat::expect_false(is.null(attr(summ.default$aliasing, which = "title"))) summ.none <-summary(array.plot.trt.canon, which.criteria = "none") testthat::expect_equal(ncol(summ.none$decomp), 6) summ.all <-summary(array.plot.trt.canon, which.criteria = "all") testthat::expect_equal(ncol(summ.all$decomp), 13) summ.2 <-summary(array.plot.trt.canon, which.criteria =c("aefficiency", "order")) testthat::expect_equal(ncol(summ.2$decomp), 8) #with grand mean apt.gm.canon <- designAnatomy(formulae = list(array = ~ (Set:Array)*Dye, plot = ~ Block/Plant/Sample, trt = ~ Treat), grandMean = TRUE, labels = "sources", data = jr.lay) summary(apt.gm.canon) testthat::expect_equal(apt.gm.canon$Q[[3]]$`Mean&Mean&Mean`, 1) #test eigen with supplied marginality for tier 2 only plot.marg <- array.plot.trt.canon$marginality$plot apt.tier2.canon <- designAnatomy(formulae = list(array = ~ (Set:Array)*Dye, plot = ~ Block/Plant/Sample, trt = ~ Treat), marginality = list(plot = plot.marg), orthog = "eigen", labels = "sources", data = jr.lay) summary(apt.tier2.canon) testthat::expect_true(all(apt.tier2.canon$sources$array == c("Set:Array", "Dye", "Set:Array:Dye"))) testthat::expect_true(all(apt.tier2.canon$sources$plot == c("Block", "Plant[Block]", "Sample[Block:Plant]"))) testthat::expect_equal(length(apt.tier2.canon$marginality), 3) testthat::expect_equal(nrow(apt.tier2.canon$marginality$plot), 3) apt.tier2.gm.canon <- designAnatomy(formulae = list(array = ~ (Set:Array)*Dye, plot = ~ Block/Plant/Sample, trt = ~ Treat), marginality = list(plot = plot.marg), orthog = "eigen", grandMean = TRUE, labels = "sources", data = jr.lay) summary(apt.tier2.gm.canon) testthat::expect_true(all(apt.tier2.gm.canon$sources$array == c("Mean", "Set:Array", "Dye", "Set:Array:Dye"))) testthat::expect_true(all(apt.tier2.gm.canon$sources$plot == c("Mean", "Block", "Plant[Block]", "Sample[Block:Plant]"))) testthat::expect_equal(length(apt.tier2.gm.canon$marginality), 3) testthat::expect_equal(nrow(apt.tier2.gm.canon$marginality$plot), 3) testthat::expect_equal(apt.tier2.gm.canon$Q[[3]]$`Mean&Mean&Mean`, 1) }) cat("#### Test for Baby pseudoterm example\n") test_that("Baby", { skip_on_cran() library(dae) #'## Baby pseudoterm example pseudo.lay <- data.frame(pl = factor(1:12), ab = factor(rep(1:4, times=3)), a = factor(rep(1:2, times=6))) pseudo.canon <- designAnatomy(formulae = list(unit=~pl, trt=~a+ab), labels = "sources", data = pseudo.lay) summary(pseudo.canon) testthat::expect_equal(pseudo.canon$Q[[2]]$'pl&ab[a]', 2) pseudo.canon <- designAnatomy(formulae = list(unit=~pl, trt=~ab+a), labels = "sources", data = pseudo.lay) summ.hybrid <- summary(pseudo.canon) testthat::expect_equal(nrow(summ.hybrid$aliasing), 2) testthat::expect_equal(ncol(summ.hybrid$aliasing), 7) testthat::expect_equal(pseudo.canon$Q[[2]]$'pl&ab', 3) testthat::expect_equal(pseudo.canon$Q[[2]]$'pl&Residual', 8) pseudo.canon <- designAnatomy(formulae = list(unit=~pl, trt=~ab+a), orthogonalize = "diff", labels = "sources", data = pseudo.lay) summ.diff <- summary(pseudo.canon) testthat::expect_equal(nrow(summ.diff$aliasing), 2) testthat::expect_equal(ncol(summ.diff$aliasing), 7) testthat::expect_equal(pseudo.canon$Q[[2]]$'pl&ab', 3) testthat::expect_equal(pseudo.canon$Q[[2]]$'pl&Residual', 8) pseudo.canon <- designAnatomy(formulae = list(unit=~pl, trt=~ab+a), orthogonalize = "eigen", labels = "sources", data = pseudo.lay) summ.eigen <- summary(pseudo.canon) testthat::expect_equal(nrow(summ.eigen$aliasing), 1) testthat::expect_equal(ncol(summ.eigen$aliasing), 7) testthat::expect_equal(pseudo.canon$Q[[2]]$'pl&ab', 3) testthat::expect_equal(pseudo.canon$Q[[2]]$'pl&Residual', 8) }) cat("#### Test for pseudoreplicated N experiment\n") test_that("pseudoN", { skip_on_cran() library(dae) #'## A design with the same RCBD in each Area sameRCBD.lay <- designRandomize(allocated = fac.gen(list(Nitrogen = 3, 3, Varieties = 24)), recipient = list(Areas = 3, Blocks = 3, Plots = 24), nested.recipients = list(Plots = "Blocks"), seed = 6411) sameRCBD.lay <- within(sameRCBD.lay, { Plot <- fac.combine(list(Blocks,Plots)) }) testthat::expect_equal(nrow(sameRCBD.lay), 216) testthat::expect_equal(ncol(sameRCBD.lay), 6) #'### nested numbering of levels sameRCBD.canon <- designAnatomy(formulae = list(unit = ~ Areas*(Blocks/Plots), trt = ~ Nitrogen*Varieties), data = sameRCBD.lay, labels = "sources", grandMean = TRUE) testthat::expect_true(all(unlist(sameRCBD.canon$Q[[2]]) == c(1,2,2,23,46,4,46,92))) testthat::expect_equal(names(sameRCBD.canon$Q[[1]][4]), "Plots[Blocks]") #Test for use of label.swap summ.s <- summary(sameRCBD.canon, which = c('aeff','order')) testthat::expect_equal(summ.s$decomp$Source.unit[4], "Plots[Blocks]") summ.t <- summary(sameRCBD.canon, which = c('aeff','order'), labels.swap = TRUE) testthat::expect_equal(summ.t$decomp$Term.unit[4], "Blocks:Plots") testthat::expect_equal(names(sameRCBD.canon$Q[[1]][4]), "Plots[Blocks]") #'### unique numbering of levels sameRCBD.canon <- designAnatomy(formulae = list(unit = ~ Areas*(Blocks + Plot), trt = ~ Nitrogen*Varieties), labels = "sources", data = sameRCBD.lay) summary(sameRCBD.canon, which = c('aeff','order')) testthat::expect_true(all(unlist(sameRCBD.canon$Q[[2]]) == c(2,2,23,46,4,46,92))) #'## A design with different RCBDs in each Area diffRCBD.lay <- designRandomize(allocated = fac.gen(list(Nitrogen = 3, 3, Varieties = 24)), recipient = list(Areas = 3, Blocks = 3, Plots = 24), nested.recipients = list(Plots = "Blocks"), seed = 6467) diffRCBD.lay <- within(diffRCBD.lay, { Block <- fac.combine(list(Areas,Blocks)) Plot <- fac.combine(list(Areas,Blocks,Plots)) }) testthat::expect_equal(nrow(diffRCBD.lay), 216) testthat::expect_equal(ncol(diffRCBD.lay), 7) #'### nested numbering of levels diffRCBD.canon <- designAnatomy(formulae = list(unit = ~ Areas/Blocks/Plots, trt = ~ Nitrogen*Varieties), data = diffRCBD.lay, labels = "sources", grandMean = TRUE) summary(diffRCBD.canon, which = c('aeff','order')) testthat::expect_true(all(unlist(diffRCBD.canon$Q[[2]]) == c(1,2,6,23,46,138))) #'### unique numbering of levels diffRCBD.canon <- designAnatomy(formulae = list(unit = ~ Areas + Block + Plot, trt = ~ Nitrogen*Varieties), labels = "sources", data = diffRCBD.lay) summary(diffRCBD.canon, which = c('aeff','order')) testthat::expect_true(all(unlist(diffRCBD.canon$Q[[2]]) == c(2,6,23,46,138))) #'### Use eigen and diff instead of hybrid diffRCBD.canon <- designAnatomy(formulae = list(unit = ~ Areas + Block + Plot, trt = ~ Nitrogen*Varieties), data = diffRCBD.lay, labels = "sources", grandMean = TRUE, orthogonalize = c("eigen", "diff")) summary(diffRCBD.canon, which = c('aeff','order')) testthat::expect_true(all(unlist(diffRCBD.canon$Q[[2]]) == c(1,2,6,23,46,138))) }) cat("#### Test for Repeated LSD for Housewives example\n") test_that("Housewives", { skip_on_cran() library(dae) data("LSRepeatHwife.dat") Hwife.struct <- pstructure( ~ Month:Week + Month:Hwife + Month:Week:Hwife, labels = "sources", data = LSRepeatHwife.dat) testthat::expect_equal(nrow(Hwife.struct$aliasing), 2) testthat::expect_equal(ncol(Hwife.struct$aliasing), 10) Hwife.canon <- designAnatomy(formulae = list(units = ~ Month:Week + Month:Hwife + Month:Week:Hwife), labels = "sources", data = LSRepeatHwife.dat) summ <- summary(Hwife.canon) testthat::expect_equal(nrow(summ$aliasing), 2) testthat::expect_equal(ncol(summ$aliasing), 7) testthat::expect_equal(nrow(Hwife.canon$aliasing), 2) testthat::expect_equal(ncol(Hwife.canon$aliasing), 11) testthat::expect_equal(Hwife.canon$Q[[1]]$`Month:Week`, 7) testthat::expect_equal(Hwife.canon$Q[[1]]$`Month:Hwife`, 6) testthat::expect_equal(Hwife.canon$Q[[1]]$`Week#Hwife[Month]`, 18) testthat::expect_error(Hwife.diff.canon <- designAnatomy(formulae = list(units = ~ Month:Week + Month:Hwife + Month:Week:Hwife), labels = "sources", orthogonalize = "diff", data = LSRepeatHwife.dat)) }) cat("#### Test for Preece examples\n") test_that("Preece", { skip_on_cran() library(dae) #'## Preece examples with two sets of treatments in a BIBD #'### two tiers preece1.lay <- fac.gen(list(block=10, plot=3)) preece1.lay$T1 <- factor(c(1,3,4, 2,4,5, 3,5,1, 4,1,2, 5,2,3, 1,2,5, 2,3,1, 3,4,2, 4,5,3, 5,1,4)) preece1.lay$T2 <- factor(c(1,2,5, 2,3,1, 3,4,2, 4,5,3, 5,1,4, 1,4,3, 2,5,4, 3,1,5, 4,2,1, 5,3,2)) preece1.lay$T3 <- factor(c(1,4,3, 2,5,4, 3,1,5, 4,2,1, 5,3,2, 1,5,2, 2,1,3, 3,2,4, 4,3,5, 5,4,1)) preece1.canon <- designAnatomy(formulae = list(plot= ~ block/plot, trt= ~ T1+T2), labels = "sources", data = preece1.lay) summary(preece1.canon, which.criteria = c("aeff", "order")) testthat::expect_equal(length(preece1.canon$Q[[1]]), 2) testthat::expect_equal(length(preece1.canon$Q[[1]]$block), 3) testthat::expect_equal(length(preece1.canon$Q[[1]]$`plot[block]`), 3) testthat::expect_lt(abs(preece1.canon$Q[[1]]$block$T1$adjusted$aefficiency - 0.1666667), 1e-05) testthat::expect_lt(abs(preece1.canon$Q[[1]]$block$T2$adjusted$aefficiency - 0.0952), 1e-04) testthat::expect_lt(abs(preece1.canon$Q[[1]]$`plot[block]`$T1$adjusted$aefficiency - 0.8333333), 1e-05) testthat::expect_lt(abs(preece1.canon$Q[[1]]$`plot[block]`$T2$adjusted$aefficiency - 0.7619), 1e-04) testthat::expect_equal(preece1.canon$Q[[2]]$'block&T1', 4) testthat::expect_equal(preece1.canon$Q[[2]]$'block&T2', 4) testthat::expect_equal(preece1.canon$Q[[2]]$'block&Residual', 1) testthat::expect_equal(preece1.canon$Q[[2]]$'plot[block]&T1', 4) testthat::expect_equal(preece1.canon$Q[[2]]$'plot[block]&T2', 4) testthat::expect_equal(preece1.canon$Q[[2]]$'plot[block]&Residual', 12) preece2.canon <- designAnatomy(formulae = list(plot= ~ block/plot, trt= ~ T1+T3), labels = "sources", data = preece1.lay) summary(preece2.canon) testthat::expect_equal(length(preece2.canon$Q[[1]]), 2) testthat::expect_equal(length(preece2.canon$Q[[1]]$block), 2) testthat::expect_equal(length(preece2.canon$Q[[1]]$`plot[block]`), 3) testthat::expect_lt(abs(preece2.canon$Q[[1]]$block$T1$adjusted$aefficiency - 0.1666667), 1e-05) testthat::expect_lt(abs(preece2.canon$Q[[1]]$`plot[block]`$T1$adjusted$aefficiency - 0.8333333), 1e-05) testthat::expect_lt(abs(preece2.canon$Q[[1]]$`plot[block]`$T3$adjusted$aefficiency - 0.8571), 1e-04) testthat::expect_equal(preece2.canon$Q[[2]]$'block&T1', 4) testthat::expect_equal(preece2.canon$Q[[2]]$'block&Residual', 5) testthat::expect_equal(preece2.canon$Q[[2]]$'plot[block]&T1', 4) testthat::expect_equal(preece2.canon$Q[[2]]$'plot[block]&T3', 4) testthat::expect_equal(preece2.canon$Q[[2]]$'plot[block]&Residual', 12) }) cat("#### Test for Mostafa's green wall experiment in 2014\n") test_that("Mostafa", { skip_on_cran() library(dae) #Mostafa's green wall experiment in 2014 data(gwall.lay) options(width = 100, nwarnings = 150) set.daeTolerance(1e-05, 1e-06) pot.treat.canon <- designAnatomy(formulae = list(pot = ~ Rows*Cols, trt = ~ Species*Irrigation*Media + First/(SpeCarry*IrrCarry*MedCarry)), labels = "sources", data = gwall.lay, keep.order=TRUE) summary(pot.treat.canon) testthat::expect_equal(length(pot.treat.canon$Q[[2]]), 17) testthat::expect_equal(pot.treat.canon$Q[[2]]$`Rows#Cols&Residual`, 69) testthat::expect_lt(abs(pot.treat.canon$Q[[1]]$`Rows#Cols`$`Species#Irrigation#Media`$adjusted$aefficiency - 0.8240828), 1e-05) testthat::expect_lt(abs(pot.treat.canon$Q[[1]]$`Rows#Cols`$`SpeCarry#IrrCarry[First]`$adjusted$aefficiency - 0.8320062), 1e-05) trt.struct <- pstructure(formula = ~ Species*Irrigation*Media + First/(SpeCarry*IrrCarry*MedCarry), labels = "sources", data = gwall.lay) testthat::expect_true(all(names(trt.struct$Q) == c("Species","Irrigation","Species#Irrigation", "Media","Species#Media","Irrigation#Media", "Species#Irrigation#Media","First", "SpeCarry[First]","IrrCarry[First]", "SpeCarry#IrrCarry[First]"))) #use designAnatomy to call pstructure trt.canon <- designAnatomy(list(trt = ~ Species*Irrigation*Media + First/(SpeCarry*IrrCarry*MedCarry)), labels = "sources", data = gwall.lay) testthat::expect_true(all(names(trt.canon$Q) == c("Species","Irrigation","Species#Irrigation", "Media","Species#Media","Irrigation#Media", "Species#Irrigation#Media","First", "SpeCarry[First]","IrrCarry[First]", "SpeCarry#IrrCarry[First]"))) }) cat("#### Test for four-tiered corn example\n") test_that("corn", { skip_on_cran() library(dae) #'## Four-tiered corn experiment with 3 structures including pseudoterms #'### Randomize field factors to lab factors corn1.recip <- list(Intervals=18, ConPlate=36) corn1.nest <- list(ConPlate = "Intervals") corn1.alloc <- fac.gen(list(Sites=3, Blocks=2, Plots=3, Lots=36)) #corn1.alloc$Harvesters <- factor(1:3, each=36, times=6) corn1.lay <- designRandomize(allocated = corn1.alloc, recipient = corn1.recip, nested.recipients = corn1.nest, seed = 81505) #'### Randomize treatments to lab factors corn2.recip <- list(Intervals=18, Containers=9, Plates=4) corn2.nest <- list(Containers = "Intervals", Plates = c("Containers", "Intervals")) Treats <- factor(rep(1:9, each=4, times=18)) corn2.lay <- designRandomize(allocated = Treats, recipient = corn2.recip, nested.recipients = corn2.nest, seed = 543205) corn2.lay <- corn2.lay[-1] #'### Randomize Harvesters to field factors corn3.recip <- list(Sites=3, Blocks=2, Plots=3, Lots=36) corn3.nest <- list(Blocks = "Sites", Plots = c("Blocks", "Sites"), Lots = c("Plots", "Blocks", "Sites")) Harvesters <- factor(rep(1:3, each=36, times=6)) corn3.lay <- designRandomize(allocated = Harvesters, recipient = corn3.recip, nested.recipients = corn3.nest, seed = 135205) #'### Combine randomizations corn.lay <- cbind(corn1.lay, corn2.lay) corn.lay <- merge(corn.lay, corn3.lay) corn.lay <- corn.lay[c("Intervals", "Containers", "Plates", "ConPlate", "Sites", "Blocks", "Plots", "Lots", "Treats", "Harvesters")] corn.lay <- corn.lay[do.call(order, corn.lay), ] rownames(corn.lay) <- NULL #'### Make factors with unique levels corn.lay <- within(corn.lay, { AContainers <- fac.combine(list(Intervals, Containers)) APlates <- fac.combine(list(AContainers, Plates)) ABlocks <- fac.combine(list(Sites, Blocks)) APlots <- fac.combine(list(ABlocks, Plots)) ALots <- fac.combine(list(APlots, Lots)) }) #'## Check properties corn.canon <- designAnatomy(formulae = list(plate= ~ Intervals + AContainers + APlates, field= ~ Sites + ABlocks + APlots + ALots, trts= ~ Sites*Harvesters*Treats), labels = "sources", data = corn.lay) summary(corn.canon, which.criteria="aeff") testthat::expect_equal(corn.canon$Q[[3]]$`Intervals&APlots[Sites:ABlocks]&Residual`, 6) testthat::expect_equal(corn.canon$Q[[3]]$`AContainers[Intervals]&ALots[Sites:ABlocks:APlots]&Residual`, 72) testthat::expect_equal(corn.canon$Q[[3]]$`APlates[Intervals:AContainers]&ALots[Sites:ABlocks:APlots]`, 486) #Create an example with double Residual corn2Res.canon <- designAnatomy(formulae = list(plate= ~ AContainers + APlates, field= ~ Sites + ABlocks + ALots, fldtrts= ~ Harvesters, labtrts= ~ Treats), labels = "sources", data = corn.lay) summary(corn2Res.canon, which.criteria="none") testthat::expect_equal(corn2Res.canon$Q[[4]]$`AContainers&ALots[Sites:ABlocks]&Residual&Treats`, 8) testthat::expect_equal(corn2Res.canon$Q[[4]]$`AContainers&ALots[Sites:ABlocks]&Residual&Residual`, 146) }) cat("#### Test for Piepho example with pseudoreplication\n") test_that("Piepho", { skip_on_cran() library(dae) #'## Piepho example with pseudoreplication data('PiephoLSDRand') piepho.canon <- designAnatomy(formulae = list(lab= ~ Times/(Locations*Ovens), field= ~ Block/Plot/Sample, trt= ~ Harvest*Method), labels = "sources", data = Piepho_LSD_Rand) summary(piepho.canon, which.criteria="aeff") testthat::expect_equal(piepho.canon$Q[[3]]$"Times&Plot[Block]&Harvest", 3) testthat::expect_equal(piepho.canon$Q[[3]]$"Locations[Times]&Block", 2) testthat::expect_equal(piepho.canon$Q[[3]]$"Locations[Times]&Plot[Block]", 6) testthat::expect_equal(piepho.canon$Q[[3]]$"Ovens[Times]&Sample[Block:Plot]", 8) testthat::expect_equal(piepho.canon$Q[[3]]$"Locations#Ovens[Times]&Sample[Block:Plot]&Method",2) testthat::expect_equal(piepho.canon$Q[[3]]$"Locations#Ovens[Times]&Sample[Block:Plot]&Harvest#Method", 6) testthat::expect_equal(piepho.canon$Q[[3]]$"Locations#Ovens[Times]&Sample[Block:Plot]&Residual", 8) Piepho_LSD_Rand <- within(Piepho_LSD_Rand, { ALocations <- fac.combine(list(Times, Locations)) AOvens <- fac.combine(list(Times, Ovens)) APositions <- fac.combine(list(Times, Locations, Ovens)) APlot <- fac.combine(list(Block, Plot)) ASample <- fac.combine(list(Block, Plot, Sample)) }) piephoA.canon <- designAnatomy(formulae = list(lab= ~ Times + ALocations + AOvens + APositions, field= ~ Block + APlot + ASample, trt= ~ Harvest*Method), labels = "sources", data = Piepho_LSD_Rand) summary(piephoA.canon, which.criteria="aeff") testthat::expect_equal(piephoA.canon$Q[[3]]$"Times&APlot[Block]&Harvest", 3) testthat::expect_equal(piephoA.canon$Q[[3]]$"ALocations[Times]&Block", 2) testthat::expect_equal(piephoA.canon$Q[[3]]$"ALocations[Times]&APlot[Block", 6) testthat::expect_equal(piephoA.canon$Q[[3]]$"AOvens[Times]&ASample[Block:APlot]", 8) testthat::expect_equal(piephoA.canon$Q[[3]]$"APositions[Times]&ASample[Block:APlot]&Method",2) testthat::expect_equal(piephoA.canon$Q[[3]]$"APositions[Times]&ASample[Block:APlot]&Harvest#Method", 6) testthat::expect_equal(piephoA.canon$Q[[3]]$"APositions[Times]&ASample[Block:APlot]&Residual", 8) }) cat("#### FAME example with complicated pseudoterm structure\n") test_that("FAME", { skip_on_cran() library(dae) "FAME example with complicated pseudoterm structure" data('FAME') FAME[c("Int1","Int2","Int3","Int4","Int5","Int6")] <- lapply(FAME[c("Int1","Int2","Int3","Int4","Int5","Int6")], factor) fame.canon <- designAnatomy(formulae = list(lab= ~ Int1:Int2:Int3/Int4/Int5/Int6, field= ~ ((Block/Plot)*Depth)/Sample, trt= ~ Tillage*Depth*Method), labels = "sources", data = FAME) summary(fame.canon, which.criteria="aeff") testthat::expect_equal(length(fame.canon$Q[[2]]), 14) testthat::expect_equal(length(fame.canon$Q[[3]]), 20) testthat::expect_equal(fame.canon$Q[[3]]$'Int1:Int2:Int3&Sample[Block:Plot:Depth]&Residual', 1) testthat::expect_equal(fame.canon$Q[[3]]$'Int5[Int1:Int2:Int3:Int4]&Plot[Block]&Residual', 3) testthat::expect_equal(fame.canon$Q[[3]]$'Int5[Int1:Int2:Int3:Int4]&Sample[Block:Plot:Depth]&Residual', 3) testthat::expect_equal(fame.canon$Q[[3]]$'Int6[Int1:Int2:Int3:Int4:Int5]&Plot#Depth[Block]&Residual', 3) testthat::expect_equal(fame.canon$Q[[3]]$'Int6[Int1:Int2:Int3:Int4:Int5]&Sample[Block:Plot:Depth]&Residual', 6) #Form factors with unique levels FAME <- within(FAME, { Int2 <- fac.combine(list(Int1,Int2)) Int3 <- fac.combine(list(Int2,Int3)) Int4 <- fac.combine(list(Int3,Int4)) Int5 <- fac.combine(list(Int4,Int5)) Int6 <- fac.combine(list(Int5,Int6)) Plot <- fac.combine(list(Block,Plot)) Sample <- fac.combine(list(Plot,Depth,Sample)) }) fameA.canon <- designAnatomy(formulae = list(lab= ~ Int3 + Int4 + Int5 + Int6, field= ~ (Block + Plot)*Depth + Sample, trt= ~ Tillage*Depth*Method), labels = "sources", data = FAME) summary(fameA.canon, which.criteria="aeff") testthat::expect_equal(length(fameA.canon$Q[[2]]), 14) testthat::expect_equal(length(fameA.canon$Q[[3]]), 20) testthat::expect_equal(fameA.canon$Q[[3]]$'Int3&Sample[Block:Plot:Depth]&Residual', 1) testthat::expect_equal(fameA.canon$Q[[3]]$'Int5[Int3:Int4]&Plot[Block]&Residual', 3) testthat::expect_equal(fameA.canon$Q[[3]]$'Int5[Int3:Int4]&Sample[Block:Plot:Depth]&Residual', 3) testthat::expect_equal(fameA.canon$Q[[3]]$'Int6[Int3:Int4:Int5]&Plot#Depth[Block]&Residual', 3) testthat::expect_equal(fameA.canon$Q[[3]]$'Int6[Int3:Int4:Int5]&Sample[Block:Plot:Depth]&Residual', 6) }) cat("#### Test for Piracicaba Euc pulp example\n") test_that("EucPulp", { skip_on_cran() library(dae) "Piracicaba Euc pulp example- an example with no Residuals" data('EucPulp2x2') euc.canon <- designAnatomy(formulae = list(lab= ~ Runs/Positions, cook= ~ Cookings/Samples, trt= ~ ((Kinds*Ages)/Lots/Batches)*Times), labels = "sources", data = Euc_Pulp2x2) summary(euc.canon, which.criteria="aeff") testthat::expect_equal(length(euc.canon$Q[[3]]), 11) testthat::expect_equal(euc.canon$Q[[3]]$'Runs&Cookings&Batches[Kinds:Ages:Lots]', 36) testthat::expect_equal(euc.canon$Q[[3]]$'Positions[Runs]&Samples[Cookings]&Batches#Times[Kinds:Ages:Lots]', 180) }) cat("#### Test for Split plot with rows and columns in main and split-plots\n") test_that("SpliPlotRowsColumns", { skip_on_cran() library(dae) #'## Split plot with rows and columns in main and split-plots data('LS4.CC4x4.1a') split.recip <- LS4_CC4x4_1a[, 1:4] split.nest <- list(SubCols = "BigRows") split.trt <- LS4_CC4x4_1a[, 5:8] split.layout <- designRandomize(allocated = split.trt, recipient = split.recip, nested.recipients = split.nest, seed = 7154) #'### Check design split.canon <- designAnatomy(formulae = list(plot= ~ ((BigRows/SubCols)*BigCols)*SubRows, trts= ~ Soils*Varieties*TreatB*MF), labels = "sources", data = split.layout) summary(split.canon, which.criteria=c("aeff","ord")) testthat::expect_equal(length(split.canon$Q[[1]]), 11) testthat::expect_equal(split.canon$Q[[1]]$'SubCols[BigRows]'$'Soils#TreatB#MF'$adjusted$aefficiency, 0.25) testthat::expect_equal(split.canon$Q[[1]]$'SubCols#BigCols[BigRows]'$'Soils#TreatB#MF'$adjusted$aefficiency, 0.75) testthat::expect_equal(split.canon$Q[[1]]$'SubCols#SubRows[BigRows]'$'Soils#Varieties#TreatB#MF'$adjusted$aefficiency, 0.25) testthat::expect_equal(split.canon$Q[[1]]$'SubCols#BigCols#SubRows[BigRows]'$'Soils#Varieties#TreatB#MF'$adjusted$aefficiency, 0.75) testthat::expect_equal(split.canon$Q[[2]]$'SubCols#BigCols#SubRows[BigRows]&Residual', 72) Q.plot <- pstructure( ~ ((BigRows/SubCols)*BigCols)*SubRows, labels = "sources", data = split.layout) names (Q.plot) split.layout <- within(split.layout, { ASubCols <- fac.combine(list(BigRows, SubCols)) }) splitA.canon <- designAnatomy(formulae = list(plot= ~ ((BigRows + ASubCols)*BigCols)*SubRows, trts= ~ Soils*Varieties*TreatB*MF), labels = "sources", data = split.layout) summary(splitA.canon, which.criteria=c("aeff","ord")) testthat::expect_equal(length(splitA.canon$Q[[1]]), 11) testthat::expect_equal(splitA.canon$Q[[1]]$'ASubCols[BigRows]'$'Soils#TreatB#MF'$adjusted$aefficiency, 0.25) testthat::expect_equal(splitA.canon$Q[[1]]$'ASubCols#BigCols[BigRows]'$'Soils#TreatB#MF'$adjusted$aefficiency, 0.75) testthat::expect_equal(splitA.canon$Q[[1]]$'ASubCols#SubRows[BigRows]'$'Soils#Varieties#TreatB#MF'$adjusted$aefficiency, 0.25) testthat::expect_equal(splitA.canon$Q[[1]]$'ASubCols#BigCols#SubRows[BigRows]'$'Soils#Varieties#TreatB#MF'$adjusted$aefficiency, 0.75) testthat::expect_equal(splitA.canon$Q[[2]]$'ASubCols#BigCols#SubRows[BigRows]&Residual', 72) }) cat("#### Test for EXP249 - a two-phae, p-rep design\n") test_that("Exp249", { skip_on_cran() library(dae) data(file="Exp249.mplot.sys") Exp249.mplot.sys$Blocks <- factor(rep(1:6, each = 44)) #'## Expand design to rerandomize lines and to assign conditions to locations Exp249.recip <- list(Groups = 6, Columns = 4, Pairs = 11, Locations = 2) Exp249.nest <- list(Columns = c("Groups", "Pairs"), Locations = c("Groups", "Columns", "Pairs")) Exp249.alloc <- data.frame(Lines = factor(rep(Exp249.mplot.sys$Lines, each=2), levels=1:75), Checks = fac.recode(rep(Exp249.mplot.sys$Lines, each=2), newlevels=c(rep(3, 73), 1 , 2), labels = c("NAM","Scout","Gladius")), Conditions = factor(rep(1:2, times=264), labels = c('0 NaCl','100 NaCl'))) Exp249.lay <- designRandomize(allocated = Exp249.alloc, recipient = Exp249.recip, nested.recipients = Exp249.nest, seed = 51412) #'## Add second-phase factors #'## (to which the first-phase factors have been systematically allocated) Exp249.lay <- cbind(fac.gen(list(Lanes = 24, Positions = 2:23)), fac.gen(list(Zones = 6, Rows = 4, MainPosn = 11, Subplots = 2)), Exp249.lay) #'## Check design properties Exp249.canon <- designAnatomy(formulae = list(carts = ~(Zones*MainPosn)/Rows/Subplots, tables = ~(Groups*Pairs)/Columns/Locations, treats = ~(Checks + Lines) * Conditions), labels = "sources", data = Exp249.lay) summary(Exp249.canon) testthat::expect_equal(length(Exp249.canon$Q[[1]]), 5) testthat::expect_lt(abs(Exp249.canon$Q[[2]]$'Zones&Groups'$'Lines'$adjusted$aefficiency - 0.1498311), 1e-05) testthat::expect_lt(abs(Exp249.canon$Q[[2]]$'Rows[Zones:MainPosn]&Columns[Groups:Pairs]'$'Lines'$adjusted$aefficiency - 0.6639769), 1e-02) testthat::expect_lt(abs(Exp249.canon$Q[[2]]$'Subplots[Zones:MainPosn:Rows]&Locations[Groups:Pairs:Columns]'$'Conditions'$adjusted$aefficiency - 1), 1e-05) testthat::expect_equal(Exp249.canon$Q[[3]]$'Rows[Zones:MainPosn]&Columns[Groups:Pairs]&Residual', 124) testthat::expect_equal(Exp249.canon$Q[[3]]$'Subplots[Zones:MainPosn:Rows]&Locations[Groups:Pairs:Columns]&Residual', 189) #'## Add factors and variates for new analysis Exp249.lay <- within(Exp249.lay, { xMainPosn <- as.numfac(MainPosn) xMainPosn <- -(xMainPosn - mean(xMainPosn)) Mainplots <- fac.combine(list(Rows,MainPosn)) }) #'## Check properties if only linear trend fitted Exp249.canon <- designAnatomy(formulae = list(cart = ~Zones/Mainplots/Subplots, treat = ~xMainPosn + (Checks + Lines) * Conditions), data = Exp249.lay, labels = "sources", orthogonalize = c("diff", "eigenmethods")) summ <- summary(Exp249.canon) testthat::expect_equal(nrow(summ$aliasing), 2) testthat::expect_equal(ncol(summ$aliasing), 7) testthat::expect_equal(length(Exp249.canon$Q[[1]]), 3) testthat::expect_lt(abs(Exp249.canon$Q[[1]]$'Zones'$'Lines'$adjusted$aefficiency - 0.1499827), 1e-05) testthat::expect_lt(abs(Exp249.canon$Q[[1]]$'Mainplots[Zones]'$'Lines'$adjusted$aefficiency - 0.987877), 1e-05) testthat::expect_lt(abs(Exp249.canon$Q[[1]]$'Subplots[Zones:Mainplots]'$'Conditions'$adjusted$aefficiency - 1), 1e-05) testthat::expect_equal(Exp249.canon$Q[[2]]$'Mainplots[Zones]&Residual', 183) testthat::expect_equal(Exp249.canon$Q[[2]]$'Subplots[Zones:Mainplots]&Residual', 189) }) cat("#### Test for EXP249A - a two-phae, p-rep design\n") test_that("Exp249_All", { skip_on_cran() library(dae) data(file="Exp249.mplot.sys") Exp249.mplot.sys$Blocks <- factor(rep(1:6, each = 44)) #'## Expand design to rerandomize lines and to assign conditions to locations Exp249.recip <- list(Groups = 6, Columns = 4, Pairs = 11, Locations = 2) Exp249.nest <- list(Columns = c("Groups", "Pairs"), Locations = c("Groups", "Columns", "Pairs")) Exp249.alloc <- data.frame(Lines = factor(rep(Exp249.mplot.sys$Lines, each=2), levels=1:75), Checks = fac.recode(rep(Exp249.mplot.sys$Lines, each=2), newlevels=c(rep(3, 73), 1 , 2), labels = c("NAM","Scout","Gladius")), Conditions = factor(rep(1:2, times=264), labels = c('0 NaCl','100 NaCl'))) Exp249.lay <- designRandomize(allocated = Exp249.alloc, recipient = Exp249.recip, nested.recipients = Exp249.nest, seed = 51412) Exp249.lay <- cbind(fac.gen(list(Lanes = 24, Positions = 2:23)), fac.gen(list(Zones = 6, Rows = 4, MainPosn = 11, Subplots = 2)), Exp249.lay) Exp249.lay <- within(Exp249.lay, { xMainPosn <- as.numfac(MainPosn) xMainPosn <- -(xMainPosn - mean(xMainPosn)) Mainplots <- fac.combine(list(Rows,MainPosn)) AMainplots <- fac.combine(list(Zones, Mainplots)) ASubplots <- fac.combine(list(AMainplots,Subplots)) }) set.daeTolerance(1e-07,1e-07) Exp249A.canon <- designAnatomy(formulae = list(cart = ~Zones + AMainplots + ASubplots, treat = ~xMainPosn + (Checks + Lines) * Conditions), labels = "sources", data = Exp249.lay) summary(Exp249A.canon) testthat::expect_equal(length(Exp249A.canon$Q[[1]]), 3) testthat::expect_lt(abs(Exp249A.canon$Q[[1]]$'Zones'$'Lines[Checks]'$adjusted$aefficiency - 0.1499827), 1e-04) testthat::expect_lt(abs(Exp249A.canon$Q[[1]]$'AMainplots[Zones]'$'Lines[Checks]'$adjusted$aefficiency - 0.987877), 1e-05) testthat::expect_lt(abs(Exp249A.canon$Q[[1]]$'ASubplots[Zones:AMainplots]'$'Conditions'$adjusted$aefficiency - 1), 1e-05) testthat::expect_equal(Exp249A.canon$Q[[2]]$'AMainplots[Zones]&Residual', 183) testthat::expect_equal(Exp249A.canon$Q[[2]]$'ASubplots[Zones:AMainplots]&Residual', 189) }) cat("#### Test for Brien and Payne 3-tier sensory experiment\n") test_that("Sensory3tier", { skip_on_cran() library(dae) #Three-tier sensory experiment data("Need3.dat") #Do short names version names(Need3.dat)[match(c("Occasions", "Intervals", "Sittings", "Positions", "Judges", "Squares", "Rows", "Columns", "Halfplots"), names(Need3.dat))] <- c("Occ", "Int", "Sit", "Pos", "Jud", "Sqr", "Row", "Col", "HPlot") #'## Complete decomposition Eval.Field.Treat.canon <- designAnatomy(list(eval=~ ((Occ/Int/Sit)*Jud)/Pos, field=~ (Row*(Sqr/Col))/HPlot, treats=~ Trellis*Method), labels = "sources", data=Need3.dat) summary(Eval.Field.Treat.canon, which.criteria =c("aefficiency", "order")) testthat::expect_equal(names(Eval.Field.Treat.canon$Q[[1]]), c("Occ","Int[Occ]","Sit[Occ:Int]","Jud","Occ#Jud","Int#Jud[Occ]", "Sit#Jud[Occ:Int]","Pos[Occ:Int:Sit:Jud]")) testthat::expect_equal(names(Eval.Field.Treat.canon$Q[[2]]), c("Occ&Sqr","Int[Occ]","Sit[Occ:Int]&Col[Sqr]","Sit[Occ:Int]&Residual", "Jud","Occ#Jud","Int#Jud[Occ]&Row","Int#Jud[Occ]&Row#Sqr", "Int#Jud[Occ]&Residual","Sit#Jud[Occ:Int]&Col[Sqr]", "Sit#Jud[Occ:Int]&Row#Col[Sqr]","Sit#Jud[Occ:Int]&Residual", "Pos[Occ:Int:Sit:Jud]&HPlot[Row:Sqr:Col]", "Pos[Occ:Int:Sit:Jud]&Residual")) testthat::expect_equal(names(Eval.Field.Treat.canon$Q[[3]]), c("Occ&Sqr","Int[Occ]","Sit[Occ:Int]&Col[Sqr]&Trellis", "Sit[Occ:Int]&Col[Sqr]&Residual","Sit[Occ:Int]&Residual", "Jud","Occ#Jud","Int#Jud[Occ]&Row","Int#Jud[Occ]&Row#Sqr", "Int#Jud[Occ]&Residual","Sit#Jud[Occ:Int]&Col[Sqr]&Trellis", "Sit#Jud[Occ:Int]&Col[Sqr]&Residual","Sit#Jud[Occ:Int]&Row#Col[Sqr]&Trellis", "Sit#Jud[Occ:Int]&Row#Col[Sqr]&Residual","Sit#Jud[Occ:Int]&Residual", "Pos[Occ:Int:Sit:Jud]&HPlot[Row:Sqr:Col]&Method", "Pos[Occ:Int:Sit:Jud]&HPlot[Row:Sqr:Col]&Trellis#Method", "Pos[Occ:Int:Sit:Jud]&HPlot[Row:Sqr:Col]&Residual", "Pos[Occ:Int:Sit:Jud]&Residual")) testthat::expect_equal(Eval.Field.Treat.canon$Q[[3]]$'Sit#Jud[Occ:Int]&Col[Sqr]&Trellis', 3) testthat::expect_equal(Eval.Field.Treat.canon$Q[[3]]$'Sit#Jud[Occ:Int]&Col[Sqr]&Residual', 3) testthat::expect_equal(Eval.Field.Treat.canon$Q[[3]]$'Sit#Jud[Occ:Int]&Row#Col[Sqr]&Trellis', 3) testthat::expect_equal(Eval.Field.Treat.canon$Q[[3]]$'Sit#Jud[Occ:Int]&Row#Col[Sqr]&Residual', 9) summ <- summary(Eval.Field.Treat.canon, which.criteria =c("aefficiency", "order")) testthat::expect_true(all(abs(summ$decomp[11:14,7] - c(0.07407407,0.66666667,0.88888889,1.00000000)) < 1e-06)) })