Merge pull request #137 from DoubleML/m-fix46

MalteKurz · web-flow · commit dcd73a20cf20 · 2021-10-26T15:09:02.000+02:00
Unit test adaptions
diff --git a/.github/workflows/rcheck.yml b/.github/workflows/rcheck.yml
@@ -16,17 +16,18 @@ jobs:
   R-CMD-check:
     runs-on: ${{ matrix.config.os }}
 
-    name: ${{ matrix.config.os }} (R ${{ matrix.config.r }}; mlr3 ${{ matrix.config.mlr3 }})
+    name: ${{ matrix.config.os }} (R ${{ matrix.config.r }}; mlr3 ${{ matrix.config.mlr3 }}; cran_no_glmnet ${{ matrix.config.cran_no_glmnet }})
 
     strategy:
       fail-fast: false
       matrix:
         config:
-          - {os: windows-latest, r: 'release', mlr3: 'release'}
-          - {os: macOS-latest, r: 'release', mlr3: 'release'}
-          - {os: ubuntu-20.04, r: 'release', mlr3: 'release', rspm: "https://packagemanager.rstudio.com/cran/__linux__/focal/latest"}
-          - {os: ubuntu-20.04, r: 'devel', mlr3: 'release', rspm: "https://packagemanager.rstudio.com/cran/__linux__/focal/latest"}
-          - {os: ubuntu-20.04, r: 'devel', mlr3: 'dev', rspm: "https://packagemanager.rstudio.com/cran/__linux__/focal/latest"}
+          - {os: windows-latest, r: 'release', mlr3: 'release', cran_no_glmnet: 'false'}
+          - {os: macOS-latest, r: 'release', mlr3: 'release', cran_no_glmnet: 'false'}
+          - {os: ubuntu-20.04, r: 'release', mlr3: 'release', rspm: "https://packagemanager.rstudio.com/cran/__linux__/focal/latest", cran_no_glmnet: 'false'}
+          - {os: ubuntu-20.04, r: 'devel', mlr3: 'release', rspm: "https://packagemanager.rstudio.com/cran/__linux__/focal/latest", cran_no_glmnet: 'false'}
+          - {os: ubuntu-20.04, r: 'devel', mlr3: 'release', rspm: "https://packagemanager.rstudio.com/cran/__linux__/focal/latest", cran_no_glmnet: 'true'}
+          - {os: ubuntu-20.04, r: 'devel', mlr3: 'dev', rspm: "https://packagemanager.rstudio.com/cran/__linux__/focal/latest", cran_no_glmnet: 'false'}
 
     env:
       R_REMOTES_NO_ERRORS_FROM_WARNINGS: true
@@ -80,15 +81,30 @@ jobs:
         shell: Rscript {0}
 
       - name: Check
-        if: runner.os != 'macOS'
+        if: |
+          runner.os != 'macOS' &&
+          matrix.config.cran_no_glmnet == 'false'
         env:
           _R_CHECK_CRAN_INCOMING_REMOTE_: false
           NOT_CRAN: false
         run: rcmdcheck::rcmdcheck(args = c("--no-manual", "--as-cran"), error_on = "warning", check_dir = "check")
         shell: Rscript {0}
 
       - name: Check
-        if: runner.os == 'macOS'
+        if: matrix.config.cran_no_glmnet == 'true'
+        env:
+          _R_CHECK_CRAN_INCOMING_REMOTE_: false
+          _R_CHECK_FORCE_SUGGESTS_: false
+          NOT_CRAN: false
+        run:  |
+          remove.packages("glmnet")
+          rcmdcheck::rcmdcheck(args = c("--no-manual", "--as-cran", "--ignore-vignettes"), build_args = c("--no-build-vignettes"), error_on = "warning", check_dir = "check")
+        shell: Rscript {0}
+
+      - name: Check
+        if: |
+          runner.os == 'macOS' &&
+          matrix.config.cran_no_glmnet == 'false'
         env:
           _R_CHECK_CRAN_INCOMING_REMOTE_: false
         run: rcmdcheck::rcmdcheck(args = c("--no-manual"), error_on = "warning", check_dir = "check")
diff --git a/tests/testthat/test-double_ml_iivm.R b/tests/testthat/test-double_ml_iivm.R
@@ -7,17 +7,17 @@ lgr::get_logger("mlr3")$set_threshold("warn")
 on_cran = !identical(Sys.getenv("NOT_CRAN"), "true")
 if (on_cran) {
   test_cases = expand.grid(
-    learner = "cv_glmnet",
+    learner = "rpart",
     dml_procedure = "dml2",
     score = "LATE",
-    trimming_threshold = c(0),
+    trimming_threshold = c(1e-5),
     stringsAsFactors = FALSE)
 } else {
   test_cases = expand.grid(
     learner = "cv_glmnet",
     dml_procedure = c("dml1", "dml2"),
     score = "LATE",
-    trimming_threshold = c(0),
+    trimming_threshold = c(1e-5),
     stringsAsFactors = FALSE)
 }
 
@@ -35,7 +35,8 @@ patrick::with_parameters_test_that("Unit tests for IIVM:",
       ml_g = learner_pars$ml_g$clone(),
       ml_m = learner_pars$ml_m$clone(),
       ml_r = learner_pars$ml_r$clone(),
-      dml_procedure = dml_procedure, score = score)
+      dml_procedure = dml_procedure, score = score,
+      trimming_threshold = trimming_threshold)
     theta = iivm_hat$coef
     se = iivm_hat$se
 
@@ -45,7 +46,8 @@ patrick::with_parameters_test_that("Unit tests for IIVM:",
       n_folds = 5, smpls = iivm_hat$smpls,
       all_preds = iivm_hat$all_preds,
       score = score,
-      bootstrap = "normal", n_rep_boot = n_rep_boot)$boot_coef
+      bootstrap = "normal", n_rep_boot = n_rep_boot,
+      trimming_threshold = trimming_threshold)$boot_coef
 
     set.seed(3141)
     double_mliivm_obj = DoubleMLIIVM$new(
diff --git a/tests/testthat/test-double_ml_iivm_user_score.R b/tests/testthat/test-double_ml_iivm_user_score.R
@@ -23,8 +23,8 @@ score_fct = function(y, z, d, g0_hat, g1_hat, m_hat, r0_hat,
 on_cran = !identical(Sys.getenv("NOT_CRAN"), "true")
 if (on_cran) {
   test_cases = expand.grid(
-    learner = "regr.glmnet",
-    learner_m = "classif.glmnet",
+    learner = "regr.rpart",
+    learner_m = "classif.rpart",
     dml_procedure = "dml2",
     trimming_threshold = c(0),
     stringsAsFactors = FALSE)
diff --git a/tests/testthat/test-double_ml_irm.R b/tests/testthat/test-double_ml_irm.R
@@ -7,7 +7,7 @@ lgr::get_logger("mlr3")$set_threshold("warn")
 on_cran = !identical(Sys.getenv("NOT_CRAN"), "true")
 if (on_cran) {
   test_cases = expand.grid(
-    learner = "cv_glmnet",
+    learner = "rpart",
     dml_procedure = "dml1",
     score = "ATTE",
     trimming_threshold = 0,
diff --git a/tests/testthat/test-double_ml_irm_user_score.R b/tests/testthat/test-double_ml_irm_user_score.R
@@ -22,18 +22,18 @@ score_fct = function(y, d, g0_hat, g1_hat, m_hat, smpls) {
 on_cran = !identical(Sys.getenv("NOT_CRAN"), "true")
 if (on_cran) {
   test_cases = expand.grid(
-    learner = "regr.glmnet",
-    learner_m = "classif.glmnet",
+    learner = "regr.rpart",
+    learner_m = "classif.rpart",
     dml_procedure = "dml2",
-    trimming_threshold = 0,
+    trimming_threshold = 1e-5,
     stringsAsFactors = FALSE)
   test_cases[".test_name"] = apply(test_cases, 1, paste, collapse = "_")
 } else {
   test_cases = expand.grid(
     learner = "regr.glmnet",
     learner_m = "classif.glmnet",
     dml_procedure = c("dml1", "dml2"),
-    trimming_threshold = c(0, 0.01),
+    trimming_threshold = c(1e-5, 0.01),
     stringsAsFactors = FALSE)
   test_cases[".test_name"] = apply(test_cases, 1, paste, collapse = "_")
 }
diff --git a/tests/testthat/test-double_ml_pliv.R b/tests/testthat/test-double_ml_pliv.R
@@ -7,7 +7,7 @@ lgr::get_logger("mlr3")$set_threshold("warn")
 on_cran = !identical(Sys.getenv("NOT_CRAN"), "true")
 if (on_cran) {
   test_cases = expand.grid(
-    learner = "regr.glmnet",
+    learner = "regr.lm",
     dml_procedure = "dml1",
     score = "partialling out",
     stringsAsFactors = FALSE)
diff --git a/tests/testthat/test-double_ml_pliv_one_way_cluster.R b/tests/testthat/test-double_ml_pliv_one_way_cluster.R
@@ -5,7 +5,7 @@ lgr::get_logger("mlr3")$set_threshold("warn")
 on_cran = !identical(Sys.getenv("NOT_CRAN"), "true")
 if (on_cran) {
   test_cases = expand.grid(
-    learner = "regr.glmnet",
+    learner = "regr.lm",
     dml_procedure = "dml1",
     score = "partialling out",
     stringsAsFactors = FALSE)
diff --git a/tests/testthat/test-double_ml_pliv_partial_functional_initializer.R b/tests/testthat/test-double_ml_pliv_partial_functional_initializer.R
@@ -5,7 +5,7 @@ lgr::get_logger("mlr3")$set_threshold("warn")
 on_cran = !identical(Sys.getenv("NOT_CRAN"), "true")
 if (on_cran) {
   test_cases = expand.grid(
-    learner = "regr.cv_glmnet",
+    learner = "regr.lm",
     dml_procedure = "dml2",
     score = "partialling out",
     stringsAsFactors = FALSE)
diff --git a/tests/testthat/test-double_ml_pliv_two_way_cluster.R b/tests/testthat/test-double_ml_pliv_two_way_cluster.R
@@ -7,7 +7,7 @@ lgr::get_logger("mlr3")$set_threshold("warn")
 on_cran = !identical(Sys.getenv("NOT_CRAN"), "true")
 if (on_cran) {
   test_cases = expand.grid(
-    learner = "regr.glmnet",
+    learner = "regr.lm",
     dml_procedure = "dml1",
     score = "partialling out",
     stringsAsFactors = FALSE)
diff --git a/tests/testthat/test-double_ml_pliv_user_score.R b/tests/testthat/test-double_ml_pliv_user_score.R
@@ -18,7 +18,7 @@ score_fct = function(y, z, d, g_hat, m_hat, r_hat, smpls) {
 on_cran = !identical(Sys.getenv("NOT_CRAN"), "true")
 if (on_cran) {
   test_cases = expand.grid(
-    learner = "regr.glmnet",
+    learner = "regr.lm",
     dml_procedure = "dml2",
     stringsAsFactors = FALSE)
 } else {
diff --git a/tests/testthat/test-double_ml_plr.R b/tests/testthat/test-double_ml_plr.R
@@ -7,7 +7,7 @@ lgr::get_logger("mlr3")$set_threshold("warn")
 on_cran = !identical(Sys.getenv("NOT_CRAN"), "true")
 if (on_cran) {
   test_cases = expand.grid(
-    learner = "regr.cv_glmnet",
+    learner = "regr.lm",
     dml_procedure = "dml2",
     score = "partialling out",
     stringsAsFactors = FALSE)
diff --git a/tests/testthat/test-double_ml_plr_classifier.R b/tests/testthat/test-double_ml_plr_classifier.R
@@ -7,8 +7,8 @@ lgr::get_logger("mlr3")$set_threshold("warn")
 on_cran = !identical(Sys.getenv("NOT_CRAN"), "true")
 if (on_cran) {
   test_cases = expand.grid(
-    g_learner = c("regr.cv_glmnet", "classif.cv_glmnet"),
-    m_learner = "classif.cv_glmnet",
+    g_learner = c("regr.rpart", "classif.rpart"),
+    m_learner = "classif.rpart",
     dml_procedure = "dml2",
     score = "partialling out",
     stringsAsFactors = FALSE)
diff --git a/tests/testthat/test-double_ml_plr_export_preds.R b/tests/testthat/test-double_ml_plr_export_preds.R
@@ -4,15 +4,25 @@ library("mlr3learners")
 
 lgr::get_logger("mlr3")$set_threshold("warn")
 
-test_cases = expand.grid(
-  g_learner = "regr.cv_glmnet",
-  m_learner = "regr.cv_glmnet",
-  dml_procedure = "dml2",
-  score = "partialling out",
-  stringsAsFactors = FALSE)
+on_cran = !identical(Sys.getenv("NOT_CRAN"), "true")
+if (on_cran) {
+  test_cases = expand.grid(
+    g_learner = "regr.rpart",
+    m_learner = "regr.rpart",
+    dml_procedure = "dml2",
+    score = "partialling out",
+    stringsAsFactors = FALSE)
+} else {
+  test_cases = expand.grid(
+    g_learner = c("regr.rpart", "regr.lm"),
+    m_learner = c("regr.rpart", "regr.lm"),
+    dml_procedure = "dml2",
+    score = "partialling out",
+    stringsAsFactors = FALSE)
+}
 test_cases[".test_name"] = apply(test_cases, 1, paste, collapse = "_")
 
-patrick::with_parameters_test_that("Unit tests for PLR with classifier for ml_m:",
+patrick::with_parameters_test_that("Unit tests for for the export of predictions:",
   .cases = test_cases, {
     n_folds = 3
 
@@ -31,6 +41,7 @@ patrick::with_parameters_test_that("Unit tests for PLR with classifier for ml_m:
     double_mlplr_obj$fit(store_predictions = TRUE)
 
     set.seed(3141)
+    Xnames = names(df)[names(df) %in% c("y", "d", "z") == FALSE]
     indx = (names(df) %in% c(Xnames, "y"))
     data = df[, indx]
     task = mlr3::TaskRegr$new(id = "ml_g", backend = data, target = "y")
@@ -42,6 +53,7 @@ patrick::with_parameters_test_that("Unit tests for PLR with classifier for ml_m:
     preds_g = as.data.table(resampling_pred$prediction())
     data.table::setorder(preds_g, "row_ids")
 
+    Xnames = names(df)[names(df) %in% c("y", "d", "z") == FALSE]
     indx = (names(df) %in% c(Xnames, "d"))
     data = df[, indx]
     task = mlr3::TaskRegr$new(id = "ml_m", backend = data, target = "d")
diff --git a/tests/testthat/test-double_ml_plr_loaded_mlr3learner.R b/tests/testthat/test-double_ml_plr_loaded_mlr3learner.R
@@ -25,8 +25,8 @@ patrick::with_parameters_test_that("Unit tests for PLR:",
     set.seed(3141)
 
     # load learner by name
-    learner_name = "regr.cv_glmnet"
-    params = list("s" = "lambda.min", "family" = "gaussian", "nfolds" = 5)
+    learner_name = "regr.rpart"
+    params = list("cp" = 0.01, "minsplit" = 20)
 
     set.seed(123)
     double_mlplr = DoubleMLPLR$new(
@@ -59,7 +59,7 @@ patrick::with_parameters_test_that("Unit tests for PLR:",
     boot_theta = double_mlplr$boot_coef
 
     set.seed(123)
-    loaded_learner = mlr3::lrn("regr.cv_glmnet", "s" = "lambda.min", "family" = "gaussian", "nfolds" = 5)
+    loaded_learner = mlr3::lrn("regr.rpart", "cp" = 0.01, "minsplit" = 20)
     double_mlplr_loaded = DoubleMLPLR$new(
       data = data_plr$dml_data,
       ml_g = loaded_learner,
@@ -78,7 +78,7 @@ patrick::with_parameters_test_that("Unit tests for PLR:",
     boot_theta_loaded = double_mlplr$boot_coef
 
     set.seed(123)
-    semiloaded_learner = mlr3::lrn("regr.cv_glmnet")
+    semiloaded_learner = mlr3::lrn("regr.rpart")
     double_mlplr_semiloaded = DoubleMLPLR$new(
       data = data_plr$dml_data,
       ml_g = semiloaded_learner,
diff --git a/tests/testthat/test-double_ml_plr_nonorth.R b/tests/testthat/test-double_ml_plr_nonorth.R
@@ -15,7 +15,7 @@ non_orth_score = function(y, d, g_hat, m_hat, smpls) {
 on_cran = !identical(Sys.getenv("NOT_CRAN"), "true")
 if (on_cran) {
   test_cases = expand.grid(
-    learner = "regr.cv_glmnet",
+    learner = "regr.lm",
     dml_procedure = "dml1",
     score = c(non_orth_score),
     n_folds = c(3),
diff --git a/tests/testthat/test-double_ml_plr_p_adjust.R b/tests/testthat/test-double_ml_plr_p_adjust.R
@@ -7,7 +7,7 @@ lgr::get_logger("mlr3")$set_threshold("warn")
 on_cran = !identical(Sys.getenv("NOT_CRAN"), "true")
 if (on_cran) {
   test_cases = expand.grid(
-    learner = "regr.cv_glmnet",
+    learner = "regr.rpart",
     dml_procedure = "dml1",
     score = "partialling out",
     method = c("romano-wolf"),
@@ -26,8 +26,7 @@ test_cases[".test_name"] = apply(test_cases, 1, paste, collapse = "_")
 
 patrick::with_parameters_test_that("Unit tests for PLR:",
   .cases = test_cases, {
-    ml_m = lrn(learner, s = "lambda.min")
-    ml_g = lrn(learner, s = "lambda.min")
+    learner_pars = get_default_mlmethod_plr(learner)
 
     n_rep_boot = 498
 
@@ -55,8 +54,8 @@ patrick::with_parameters_test_that("Unit tests for PLR:",
       y_col = "y",
       d_cols = colnames(X)[1:k])
     double_mlplr_obj = DoubleMLPLR$new(data_ml,
-      ml_g = ml_g,
-      ml_m = ml_m,
+      ml_g = learner_pars$ml_g$clone(),
+      ml_m = learner_pars$ml_m$clone(),
       dml_procedure = dml_procedure,
       n_folds = n_folds,
       score = score,
diff --git a/tests/testthat/test-double_ml_plr_set_samples.R b/tests/testthat/test-double_ml_plr_set_samples.R
@@ -7,7 +7,7 @@ lgr::get_logger("mlr3")$set_threshold("warn")
 on_cran = !identical(Sys.getenv("NOT_CRAN"), "true")
 if (on_cran) {
   test_cases = expand.grid(
-    learner = "regr.cv_glmnet",
+    learner = "regr.rpart",
     dml_procedure = "dml2",
     score = "partialling out",
     n_folds = c(2),
diff --git a/tests/testthat/test-double_ml_plr_user_score.R b/tests/testthat/test-double_ml_plr_user_score.R
@@ -17,7 +17,7 @@ score_fct = function(y, d, g_hat, m_hat, smpls) {
 on_cran = !identical(Sys.getenv("NOT_CRAN"), "true")
 if (on_cran) {
   test_cases = expand.grid(
-    learner = "regr.glmnet",
+    learner = "regr.lm",
     dml_procedure = "dml1",
     n_folds = c(3),
     n_rep = c(2),
