add testing for k_means ClusterR usage

Author: EmilHvitfeldt

tests/testthat/test-k_means-clusterR.R

@@ -0,0 +1,79 @@
+test_that("fitting", {
+  skip_if_not_installed("ClusterR")
+
+  set.seed(1234)
+  spec <- k_means(num_clusters = 3) %>%
+    set_engine("ClusterR")
+
+  expect_no_error(
+    res <- fit(spec, ~., mtcars)
+  )
+
+  expect_no_error(
+    res <- fit_xy(spec, mtcars)
+  )
+})
+
+test_that("predicting", {
+  skip_if_not_installed("ClusterR")
+
+  set.seed(1234)
+  spec <- k_means(num_clusters = 3) %>%
+    set_engine("ClusterR")
+
+  res <- fit(spec, ~., mtcars)
+
+  preds <- predict(res, mtcars[c(1:5), ])
+
+  expect_identical(
+    preds,
+    tibble::tibble(.pred_cluster = factor(paste0("Cluster_", c(1, 1, 1, 2, 2)),
+                                          paste0("Cluster_", 1:3)))
+  )
+})
+
+test_that("extract_centroids() works", {
+  skip_if_not_installed("ClusterR")
+
+  set.seed(1234)
+  spec <- k_means(num_clusters = 3) %>%
+    set_engine("ClusterR")
+
+  res <- fit(spec, ~., mtcars)
+
+  centroids <- extract_centroids(res)
+
+  expected <- vctrs::vec_cbind(
+    tibble::tibble(.cluster = factor(paste0("Cluster_", 1:3))),
+    tibble::as_tibble(res$fit$centroids)
+  )
+
+  expect_identical(
+    centroids,
+    expected
+  )
+})
+
+test_that("extract_cluster_assignment() works", {
+  skip_if_not_installed("ClusterR")
+
+  set.seed(1234)
+  spec <- k_means(num_clusters = 3) %>%
+    set_engine("ClusterR")
+
+  res <- fit(spec, ~., mtcars)
+
+  clusters <- extract_cluster_assignment(res)
+
+  exp_cluster <- res$fit$cluster
+  exp_cluster <- order(unique(exp_cluster))[exp_cluster]
+
+  expected <- vctrs::vec_cbind(
+    tibble::tibble(.cluster = factor(paste0("Cluster_", exp_cluster)))
+  )
+
+  expect_identical(
+    clusters,
+    expected
+  )
+})