test-functions_migration.R

# tests
source("param.r")

test_that("stats computations",{
	DistClasses<-makeDistClasses(maps$X,maps$Y,genIntervals)
	DistClasses<-makeDistClassesWithStreets(maps$X,maps$Y,genIntervals, blockIndex)
	nbpairs<-(length(maps$X)^2-length(maps$X))/2
	nbDistOk<-count(DistClasses$dists<max(genIntervals) & DistClasses$dists>0)
	nbpairs<-(length(maps$X)^2-length(maps$X))/2
	nbDistOk<-count(DistClasses$dists<max(genIntervals) & DistClasses$dists>0)
	expect_true(nbDistOk<nbpairs)
	expect_equal(sum(DistClasses$classSize),nbDistOk) # all dist ok are in bins
	expect_equal(count(DistClasses$dists>=max(genIntervals)), count(DistClasses$CClassIndex== -1)) # the -1 correspond to pairs out of dist range
	expect_true(all.equal(DistClasses$classSize,varioTest$n))
	expect_true(all.equal(DistClasses$classSize, DistClasses$classSizeSB + DistClasses$classSizeAS))

	vario1<-variogFromIndices(DistClasses$CClassIndex,maps$infest1,DistClasses$classSize) # variog for init
	# print(vario1)
	expect_true(all.equal(vario1$variog,varioTest$v))
	expect_true(all.equal(vario1$sdvariog,varioTest$sd))
})
test_that("probmatR and probmatC equivalent",{ 	
	threshold <- 2000
	### R
	dist_mat <- nearest.dist(x=sp, y=NULL, method="euclidian", delta=threshold, upper=NULL);          
	dist_mat <- as.matrix(dist_mat)
	probMatR <- generate_prob_mat(halfDistJ, halfDistH, useDelta, delta, rateHopInMove, rateSkipInMove, rateJumpInMove, threshold, sp, dist_mat, SB, FALSE)
	cumulProbMatR <- generate_prob_mat(halfDistJ, halfDistH, useDelta, delta, rateHopInMove, rateSkipInMove, rateJumpInMove, threshold, sp, dist_mat, SB, TRUE)

	### C
	DistClasses<-makeDistClassesWithStreets(maps$X,maps$Y,genIntervals, blockIndex)
	probMatC <- generate_prob_mat_C(halfDistJ, halfDistH, useDelta, delta, rateHopInMove, rateSkipInMove, rateJumpInMove, DistClasses$dists, blockIndex, L, FALSE)
	cumulProbMatC <- generate_prob_mat_C(halfDistJ, halfDistH, useDelta, delta, rateHopInMove, rateSkipInMove, rateJumpInMove, DistClasses$dists, blockIndex, L, TRUE)

	### tests
	expect_true(all.equal(cumulProbMatR, cumulProbMatC))
	expect_true(all.equal(probMatR, probMatC))
})

test_that("getPosteriorMaps (and multiGilStats behind the scene)",{
	source("param.r")
	source("maps_basic_regression.R")# for maps$X, maps$Y, maps$blockIndex
	set.seed(1)
	# and maps$infest3
	DistClasses<-makeDistClassesWithStreets(maps$X,maps$Y,genIntervals, maps$blockIndex)
	cumulProbMat <- generate_prob_mat_C(halfDistJ, halfDistH, useDelta, delta, rateHopInMove, rateSkipInMove, rateJumpInMove, DistClasses$dists, blockIndex, cumul=TRUE)

	thetas<-rnorm(10,mean=rateMove,sd=0.01)
	thetas<-thetas[which(thetas>0)]
	test<-getPosteriorMaps(maps,thetas,cumulProbMat=cumulProbMat,maps$infest3,nbit=2*52,repByTheta=10)
	expect_true(all.equal(test,maps$test))

	# par(mfrow=c(1,3))
	# plot_reel(maps$X,maps$Y,maps$infest3,base=0,top=1)
	# plot_reel(maps$X,maps$Y,maps$test,base=0,top=1)
	# plot_reel(maps$X,maps$Y,test,base=0,top=1)
	# dump("maps",file="maps_basic_regression.R")
})