v1.2.1

CONTRIBUTING.md

@@ -10,5 +10,14 @@ By participating in this project you agree to abide by its terms.
 Please follow [our guideline](.github/ISSUE_TEMPLATE/bug_report.md) to create an issue.
 
 ## PRs
-For the details about how to create a pull request, see the official document of [GitHub](https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/proposing-changes-to-your-work-with-pull-requests/creating-a-pull-request).
+For the details about how to create a pull request (PR), see the official document of [GitHub](https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/proposing-changes-to-your-work-with-pull-requests/creating-a-pull-request).
 
+If fixing a bug or adding a new feature to a `dcTensor` function, please add a `testthat` unit test code like `tests/testthat/test_[NEW_FEATURE].R`.
+
+Codes that take too long are commented out in `testthat.R`.
+This is because CRAN expects tests to be completed in a short period of time.
+If your test code takes more than 1 minute, please comment it out.
+
+Also add a command `test_file("testthat/test_[NEW_FEATURE].R")` in `tests/testthat.R`.
+
+The test code is performed in `Dockerfile` called by `.github/workflows/build_test_push.yml` (GitHub Actions) and you will see whether the test code has been finished without error after your PR. **Please do not leave the PR with errors.**

inst/NEWS

@@ -1,3 +1,7 @@
+VERSION 1.2.1
+------------------------
+   o Debugged tests
+
 VERSION 1.2.0
 ------------------------
    o {Binary,Ternary}-regularization terms are removed because of the calculation time issue

tests/testthat/test_NA_dNMF.R

@@ -1,12 +1,14 @@
-X <- dcTensor::toyModel("dNMF")
-X[sample(seq(length(X)), 0.1*length(X))] <- NA
+test_that("dNMF (NA)", {
+	X <- dcTensor::toyModel("dNMF")
+	X[sample(seq(length(X)), 0.1*length(X))] <- NA
 
-out1 <- dNMF(X, J=3, algorithm="Frobenius", num.iter=2)
-out2 <- dNMF(X, J=3, algorithm="KL", num.iter=2)
-out3 <- dNMF(X, J=3, algorithm="IS", num.iter=2)
-out4 <- dNMF(X, J=3, algorithm="Beta", Beta=-1, num.iter=2)
+	out1 <- dNMF(X, J=3, algorithm="Frobenius", num.iter=2)
+	out2 <- dNMF(X, J=3, algorithm="KL", num.iter=2)
+	out3 <- dNMF(X, J=3, algorithm="IS", num.iter=2)
+	out4 <- dNMF(X, J=3, algorithm="Beta", Beta=-1, num.iter=2)
 
-expect_equivalent(length(out1), 10)
-expect_equivalent(length(out2), 10)
-expect_equivalent(length(out3), 10)
-expect_equivalent(length(out4), 10)
+	expect_equivalent(length(out1), 6)
+	expect_equivalent(length(out2), 6)
+	expect_equivalent(length(out3), 6)
+	expect_equivalent(length(out4), 6)
+})

tests/testthat/test_NA_dNTD.R

@@ -1,35 +1,37 @@
-#
-# 3-order tensor
-#
-X <- dcTensor::toyModel("dNTF")
-X@data[sample(seq(length(X@data)), 0.1*length(X@data))] <- NA
+test_that("dNTD (NA)", {
+	#
+	# 3-order tensor
+	#
+	X <- dcTensor::toyModel("dNTF")
+	X@data[sample(seq(length(X@data)), 0.1*length(X@data))] <- NA
 
-out1_1 <- dNTD(X, rank=c(1,2,3), algorithm="Frobenius", num.iter=2)
-out1_2 <- dNTD(X, rank=c(1,2,3), algorithm="Frobenius", init="dNMF", num.iter=2)
-out1_3 <- dNTD(X, rank=c(1,2,3), algorithm="Frobenius", init="Random", num.iter=2)
-out2 <- dNTD(X, rank=c(1,2,3), algorithm="KL", num.iter=2)
-out3 <- dNTD(X, rank=c(1,2,3), algorithm="IS", num.iter=2)
-out4 <- dNTD(X, rank=c(1,2,3), algorithm="Beta", num.iter=2)
+	out1_1 <- dNTD(X, rank=c(1,2,3), algorithm="Frobenius", num.iter=2)
+	out1_2 <- dNTD(X, rank=c(1,2,3), algorithm="Frobenius", init="dNMF", num.iter=2)
+	out1_3 <- dNTD(X, rank=c(1,2,3), algorithm="Frobenius", init="Random", num.iter=2)
+	out2 <- dNTD(X, rank=c(1,2,3), algorithm="KL", num.iter=2)
+	out3 <- dNTD(X, rank=c(1,2,3), algorithm="IS", num.iter=2)
+	out4 <- dNTD(X, rank=c(1,2,3), algorithm="Beta", num.iter=2)
 
-out_NTD2_1 <- dNTD(X, rank=c(2,3), modes=1:2, algorithm="Frobenius", num.iter=2)
-out_NTD2_2 <- dNTD(X, rank=c(3,4), modes=2:3, algorithm="Frobenius", num.iter=2)
-out_NTD2_3 <- dNTD(X, rank=c(4,6), modes=c(1,3), algorithm="Frobenius", num.iter=2)
+	out_NTD2_1 <- dNTD(X, rank=c(2,3), modes=1:2, algorithm="Frobenius", num.iter=2)
+	out_NTD2_2 <- dNTD(X, rank=c(3,4), modes=2:3, algorithm="Frobenius", num.iter=2)
+	out_NTD2_3 <- dNTD(X, rank=c(4,6), modes=c(1,3), algorithm="Frobenius", num.iter=2)
 
-out_NTD1_1 <- dNTD(X, rank=3, modes=1, algorithm="Frobenius", num.iter=2)
-out_NTD1_2 <- dNTD(X, rank=4, modes=2, algorithm="Frobenius", num.iter=2)
-out_NTD1_3 <- dNTD(X, rank=5, modes=3, algorithm="Frobenius", num.iter=2)
+	out_NTD1_1 <- dNTD(X, rank=3, modes=1, algorithm="Frobenius", num.iter=2)
+	out_NTD1_2 <- dNTD(X, rank=4, modes=2, algorithm="Frobenius", num.iter=2)
+	out_NTD1_3 <- dNTD(X, rank=5, modes=3, algorithm="Frobenius", num.iter=2)
 
-expect_equivalent(length(out1_1), 8)
-expect_equivalent(length(out1_2), 8)
-expect_equivalent(length(out1_3), 8)
-expect_equivalent(length(out2), 8)
-expect_equivalent(length(out3), 8)
-expect_equivalent(length(out4), 8)
+	expect_equivalent(length(out1_1), 6)
+	expect_equivalent(length(out1_2), 6)
+	expect_equivalent(length(out1_3), 6)
+	expect_equivalent(length(out2), 6)
+	expect_equivalent(length(out3), 6)
+	expect_equivalent(length(out4), 6)
 
-expect_equivalent(length(out_NTD2_1), 8)
-expect_equivalent(length(out_NTD2_2), 8)
-expect_equivalent(length(out_NTD2_3), 8)
+	expect_equivalent(length(out_NTD2_1), 6)
+	expect_equivalent(length(out_NTD2_2), 6)
+	expect_equivalent(length(out_NTD2_3), 6)
 
-expect_equivalent(length(out_NTD1_1), 8)
-expect_equivalent(length(out_NTD1_2), 8)
-expect_equivalent(length(out_NTD1_3), 8)
+	expect_equivalent(length(out_NTD1_1), 6)
+	expect_equivalent(length(out_NTD1_2), 6)
+	expect_equivalent(length(out_NTD1_3), 6)
+})

tests/testthat/test_NA_dNTF.R

@@ -1,19 +1,21 @@
-#
-# 3-order tensor
-#
-X <- dcTensor:::toyModel("dNTF")
-X@data[sample(seq(length(X@data)), 0.1*length(X@data))] <- NA
+test_that("dNTF (NA)", {
+	#
+	# 3-order tensor
+	#
+	X <- dcTensor:::toyModel("dNTF")
+	X@data[sample(seq(length(X@data)), 0.1*length(X@data))] <- NA
 
-out1_1 <- dNTF(X, rank=3, algorithm="Frobenius", num.iter=2)
-out1_2 <- dNTF(X, rank=3, algorithm="Frobenius", init="dNMF", num.iter=2)
-out1_3 <- dNTF(X, rank=3, algorithm="Frobenius", init="Random", num.iter=2)
-out2 <- dNTF(X, rank=3, algorithm="KL", num.iter=2)
-out3 <- dNTF(X, rank=3, algorithm="IS", num.iter=2)
-out4 <- dNTF(X, rank=3, algorithm="Beta", Beta=-1, num.iter=2)
+	out1_1 <- dNTF(X, rank=3, algorithm="Frobenius", num.iter=2)
+	out1_2 <- dNTF(X, rank=3, algorithm="Frobenius", init="dNMF", num.iter=2)
+	out1_3 <- dNTF(X, rank=3, algorithm="Frobenius", init="Random", num.iter=2)
+	out2 <- dNTF(X, rank=3, algorithm="KL", num.iter=2)
+	out3 <- dNTF(X, rank=3, algorithm="IS", num.iter=2)
+	out4 <- dNTF(X, rank=3, algorithm="Beta", Beta=-1, num.iter=2)
 
-expect_equivalent(length(out1_1), 8)
-expect_equivalent(length(out1_2), 8)
-expect_equivalent(length(out1_3), 8)
-expect_equivalent(length(out2), 8)
-expect_equivalent(length(out3), 8)
-expect_equivalent(length(out4), 8)
+	expect_equivalent(length(out1_1), 6)
+	expect_equivalent(length(out1_2), 6)
+	expect_equivalent(length(out1_3), 6)
+	expect_equivalent(length(out2), 6)
+	expect_equivalent(length(out3), 6)
+	expect_equivalent(length(out4), 6)
+})

tests/testthat/test_NA_dPLS.R

@@ -1,8 +1,9 @@
-X <- dcTensor::toyModel("dPLS_Easy")
-X[[1]][sample(seq(length(X[[1]])), 0.1*length(X[[1]]))] <- NA
-X[[2]][sample(seq(length(X[[2]])), 0.1*length(X[[2]]))] <- NA
-X[[3]][sample(seq(length(X[[3]])), 0.1*length(X[[3]]))] <- NA
-
-out <- dPLS(X, J=3, num.iter=2)
-expect_equivalent(length(out), 7)
+test_that("dPLS (NA)", {
+	X <- dcTensor::toyModel("dPLS_Easy")
+	X[[1]][sample(seq(length(X[[1]])), 0.1*length(X[[1]]))] <- NA
+	X[[2]][sample(seq(length(X[[2]])), 0.1*length(X[[2]]))] <- NA
+	X[[3]][sample(seq(length(X[[3]])), 0.1*length(X[[3]]))] <- NA
 
+	out <- dPLS(X, J=3, num.iter=2)
+	expect_equivalent(length(out), 6)
+})

tests/testthat/test_NA_dSVD.R

@@ -1,5 +1,7 @@
-X <- dcTensor::toyModel("dSVD")
-X[sample(seq(length(X)), 0.1*length(X))] <- NA
+test_that("dSVD (NA)", {
+	X <- dcTensor::toyModel("dSVD")
+	X[sample(seq(length(X)), 0.1*length(X))] <- NA
 
-out <- dSVD(X, J=3, num.iter=2)
-expect_equivalent(length(out), 7)
+	out <- dSVD(X, J=3, num.iter=2)
+	expect_equivalent(length(out), 6)
+})

tests/testthat/test_NA_djNMF.R

@@ -1,14 +1,16 @@
-X <- dcTensor:::toyModel("dsiNMF_Hard")
-X[[1]][sample(seq(length(X[[1]])), 0.1*length(X[[1]]))] <- NA
-X[[2]][sample(seq(length(X[[2]])), 0.1*length(X[[2]]))] <- NA
-X[[3]][sample(seq(length(X[[3]])), 0.1*length(X[[3]]))] <- NA
+test_that("djNMF (NA)", {
+	X <- dcTensor:::toyModel("dsiNMF_Hard")
+	X[[1]][sample(seq(length(X[[1]])), 0.1*length(X[[1]]))] <- NA
+	X[[2]][sample(seq(length(X[[2]])), 0.1*length(X[[2]]))] <- NA
+	X[[3]][sample(seq(length(X[[3]])), 0.1*length(X[[3]]))] <- NA
 
-out1 <- djNMF(X, J=3, algorithm="Frobenius", num.iter=2)
-out2 <- djNMF(X, J=3, algorithm="KL", num.iter=2)
-out3 <- djNMF(X, J=3, algorithm="IS", num.iter=2)
-out4 <- djNMF(X, J=3, algorithm="PLTF", p=1, num.iter=2)
+	out1 <- djNMF(X, J=3, algorithm="Frobenius", num.iter=2)
+	out2 <- djNMF(X, J=3, algorithm="KL", num.iter=2)
+	out3 <- djNMF(X, J=3, algorithm="IS", num.iter=2)
+	out4 <- djNMF(X, J=3, algorithm="PLTF", p=1, num.iter=2)
 
-expect_equivalent(length(out1), 13)
-expect_equivalent(length(out2), 13)
-expect_equivalent(length(out3), 13)
-expect_equivalent(length(out4), 13)
+	expect_equivalent(length(out1), 7)
+	expect_equivalent(length(out2), 7)
+	expect_equivalent(length(out3), 7)
+	expect_equivalent(length(out4), 7)
+})

tests/testthat/test_NA_dsiNMF.R

@@ -1,14 +1,16 @@
-X <- dcTensor::toyModel("dsiNMF_Easy")
-X[[1]][sample(seq(length(X[[1]])), 0.1*length(X[[1]]))] <- NA
-X[[2]][sample(seq(length(X[[2]])), 0.1*length(X[[2]]))] <- NA
-X[[3]][sample(seq(length(X[[3]])), 0.1*length(X[[3]]))] <- NA
+test_that("dsiNMF (NA)", {
+	X <- dcTensor::toyModel("dsiNMF_Easy")
+	X[[1]][sample(seq(length(X[[1]])), 0.1*length(X[[1]]))] <- NA
+	X[[2]][sample(seq(length(X[[2]])), 0.1*length(X[[2]]))] <- NA
+	X[[3]][sample(seq(length(X[[3]])), 0.1*length(X[[3]]))] <- NA
 
-out1 <- dsiNMF(X, J=3, algorithm="Frobenius", num.iter=2)
-out2 <- dsiNMF(X, J=3, algorithm="KL", num.iter=2)
-out3 <- dsiNMF(X, J=3, algorithm="IS", num.iter=2)
-out4 <- dsiNMF(X, J=3, algorithm="PLTF", p=1, num.iter=2)
+	out1 <- dsiNMF(X, J=3, algorithm="Frobenius", num.iter=2)
+	out2 <- dsiNMF(X, J=3, algorithm="KL", num.iter=2)
+	out3 <- dsiNMF(X, J=3, algorithm="IS", num.iter=2)
+	out4 <- dsiNMF(X, J=3, algorithm="PLTF", p=1, num.iter=2)
 
-expect_equivalent(length(out1), 10)
-expect_equivalent(length(out2), 10)
-expect_equivalent(length(out3), 10)
-expect_equivalent(length(out4), 10)
+	expect_equivalent(length(out1), 6)
+	expect_equivalent(length(out2), 6)
+	expect_equivalent(length(out3), 6)
+	expect_equivalent(length(out4), 6)
+})

tests/testthat/test_dNMF.R

@@ -1,24 +1,26 @@
-X <- dcTensor::toyModel("dNMF")
+test_that("dNMF", {
+	X <- dcTensor::toyModel("dNMF")
 
-# Normal Test
-out1 <- dNMF(X, J=3, algorithm="Frobenius", num.iter=2)
-out2 <- dNMF(X, J=3, algorithm="KL", num.iter=2)
-out3 <- dNMF(X, J=3, algorithm="IS", num.iter=2)
-out4 <- dNMF(X, J=3, algorithm="Beta", num.iter=2)
+	# Normal Test
+	out1 <- dNMF(X, J=3, algorithm="Frobenius", num.iter=2)
+	out2 <- dNMF(X, J=3, algorithm="KL", num.iter=2)
+	out3 <- dNMF(X, J=3, algorithm="IS", num.iter=2)
+	out4 <- dNMF(X, J=3, algorithm="Beta", num.iter=2)
 
-expect_equivalent(length(out1), 10)
-expect_equivalent(length(out2), 10)
-expect_equivalent(length(out3), 10)
-expect_equivalent(length(out4), 10)
+	expect_equivalent(length(out1), 6)
+	expect_equivalent(length(out2), 6)
+	expect_equivalent(length(out3), 6)
+	expect_equivalent(length(out4), 6)
 
-# Mask Test
-M <- kFoldMaskTensor(X, k=5)
-out1_M <- dNMF(X, M=M[[1]], J=3, algorithm="Frobenius", num.iter=2)
-out2_M <- dNMF(X, M=M[[1]], J=3, algorithm="KL", num.iter=2)
-out3_M <- dNMF(X, M=M[[1]], J=3, algorithm="IS", num.iter=2)
-out4_M <- dNMF(X, M=M[[1]], J=3, algorithm="Beta", num.iter=2)
+	# Mask Test
+	M <- kFoldMaskTensor(X, k=5)
+	out1_M <- dNMF(X, M=M[[1]], J=3, algorithm="Frobenius", num.iter=2)
+	out2_M <- dNMF(X, M=M[[1]], J=3, algorithm="KL", num.iter=2)
+	out3_M <- dNMF(X, M=M[[1]], J=3, algorithm="IS", num.iter=2)
+	out4_M <- dNMF(X, M=M[[1]], J=3, algorithm="Beta", num.iter=2)
 
-expect_equivalent(length(out1_M), 10)
-expect_equivalent(length(out2_M), 10)
-expect_equivalent(length(out3_M), 10)
-expect_equivalent(length(out4_M), 10)
+	expect_equivalent(length(out1_M), 6)
+	expect_equivalent(length(out2_M), 6)
+	expect_equivalent(length(out3_M), 6)
+	expect_equivalent(length(out4_M), 6)
+})

tests/testthat/test_dNTD.R

@@ -1,11 +1,13 @@
-X <- dcTensor::toyModel("dNTF")
+test_that("dNTD", {
+	X <- dcTensor::toyModel("dNTF")
 
-out1 <- dNTD(X, rank=c(3,3,3), algorithm="Frobenius", num.iter=2)
-out2 <- dNTD(X, rank=c(3,3,3), algorithm="KL", num.iter=2)
-out3 <- dNTD(X, rank=c(3,3,3), algorithm="IS", num.iter=2)
-out4 <- dNTD(X, rank=c(3,3,3), algorithm="Beta", num.iter=2)
+	out1 <- dNTD(X, rank=c(3,3,3), algorithm="Frobenius", num.iter=2)
+	out2 <- dNTD(X, rank=c(3,3,3), algorithm="KL", num.iter=2)
+	out3 <- dNTD(X, rank=c(3,3,3), algorithm="IS", num.iter=2)
+	out4 <- dNTD(X, rank=c(3,3,3), algorithm="Beta", num.iter=2)
 
-expect_equivalent(length(out1), 8)
-expect_equivalent(length(out2), 8)
-expect_equivalent(length(out3), 8)
-expect_equivalent(length(out4), 8)
+	expect_equivalent(length(out1), 6)
+	expect_equivalent(length(out2), 6)
+	expect_equivalent(length(out3), 6)
+	expect_equivalent(length(out4), 6)
+})

tests/testthat/test_dNTF.R

@@ -1,11 +1,13 @@
-X <- dcTensor::toyModel("dNTF")
+test_that("dNTF", {
+	X <- dcTensor::toyModel("dNTF")
 
-out1 <- dNTF(X, rank=3, algorithm="Frobenius", num.iter=2)
-out2 <- dNTF(X, rank=3, algorithm="KL", num.iter=2)
-out3 <- dNTF(X, rank=3, algorithm="IS", num.iter=2)
-out4 <- dNTF(X, rank=3, algorithm="Beta", num.iter=2)
+	out1 <- dNTF(X, rank=3, algorithm="Frobenius", num.iter=2)
+	out2 <- dNTF(X, rank=3, algorithm="KL", num.iter=2)
+	out3 <- dNTF(X, rank=3, algorithm="IS", num.iter=2)
+	out4 <- dNTF(X, rank=3, algorithm="Beta", num.iter=2)
 
-expect_equivalent(length(out1), 8)
-expect_equivalent(length(out2), 8)
-expect_equivalent(length(out3), 8)
-expect_equivalent(length(out4), 8)
+	expect_equivalent(length(out1), 6)
+	expect_equivalent(length(out2), 6)
+	expect_equivalent(length(out3), 6)
+	expect_equivalent(length(out4), 6)
+})

tests/testthat/test_dPLS.R

@@ -1,5 +1,7 @@
-X <- dcTensor::toyModel("dPLS_Easy")
+test_that("dPLS", {
+	X <- dcTensor::toyModel("dPLS_Easy")
 
-out <- dPLS(X, J=3, num.iter=2)
+	out <- dPLS(X, J=3, num.iter=2)
 
-expect_equivalent(length(out), 7)
+	expect_equivalent(length(out), 6)
+})

tests/testthat/test_dSVD.R

@@ -1,5 +1,7 @@
-X <- dcTensor::toyModel("dSVD")
+test_that("dSVD", {
+	X <- dcTensor::toyModel("dSVD")
 
-out <- dSVD(X, J=3, num.iter=2)
+	out <- dSVD(X, J=3, num.iter=2)
 
-expect_equivalent(length(out), 7)
+	expect_equivalent(length(out), 6)
+})

tests/testthat/test_djNMF.R

@@ -1,11 +1,13 @@
-X <- dcTensor::toyModel("dsiNMF_Easy")
+test_that("djNMF", {
+	X <- dcTensor::toyModel("dsiNMF_Easy")
 
-out1 <- djNMF(X, J=3, algorithm="Frobenius", num.iter=2)
-out2 <- djNMF(X, J=3, algorithm="KL", num.iter=2)
-out3 <- djNMF(X, J=3, algorithm="IS", num.iter=2)
-out4 <- djNMF(X, J=3, algorithm="PLTF", num.iter=2)
+	out1 <- djNMF(X, J=3, algorithm="Frobenius", num.iter=2)
+	out2 <- djNMF(X, J=3, algorithm="KL", num.iter=2)
+	out3 <- djNMF(X, J=3, algorithm="IS", num.iter=2)
+	out4 <- djNMF(X, J=3, algorithm="PLTF", num.iter=2)
 
-expect_equivalent(length(out1), 13)
-expect_equivalent(length(out2), 13)
-expect_equivalent(length(out3), 13)
-expect_equivalent(length(out4), 13)
+	expect_equivalent(length(out1), 7)
+	expect_equivalent(length(out2), 7)
+	expect_equivalent(length(out3), 7)
+	expect_equivalent(length(out4), 7)
+})