Merge pull request #1204 from tidymodels/stricter-sparsevctrs-tests

Stricter sparsevctrs tests

NEWS.md

@@ -2,7 +2,7 @@
 
 * `fit_xy()` can now take dgCMatrix input for `x` argument (#1121).
 
-* `fit()` and `fit_xy()` can now take sparse tibbles as data values (#1165).
+* `fit_xy()` can now take sparse tibbles as data values (#1165).
 
 * `predict()` can now take dgCMatrix and sparse tibble input for `new_data` argument, and error informatively when model doesn't support it (#1167).
 

R/convert_data.R

@@ -48,6 +48,13 @@
     )
   }
 
+  if (is_sparse_tibble(data)) {
+    cli::cli_abort(
+      "Sparse data cannot be used with formula interface. Please use 
+      {.fn fit_xy} instead."
+    )
+  }
+
   if (remove_intercept) {
     data <- data[, colnames(data) != "(Intercept)", drop = FALSE]
   }

R/predict.R

@@ -470,6 +470,10 @@ prepare_data <- function(object, new_data) {
   if (allow_sparse(object) && inherits(new_data, "dgCMatrix")) {
     return(new_data)
   }
+  if (allow_sparse(object) && is_sparse_tibble(new_data)) {
+    new_data <- sparsevctrs::coerce_to_sparse_matrix(new_data)
+    return(new_data)
+  }
 
   fit_interface <- object$spec$method$fit$interface
   switch(

tests/testthat/_snaps/sparsevctrs.md

@@ -1,51 +1,75 @@
-# sparse tibble can be passed to `fit()
+# sparse tibble can be passed to `fit() - supported
+
+    Code
+      xgb_fit <- fit(spec, avg_price_per_room ~ ., data = hotel_data)
+    Condition
+      Error in `.convert_form_to_xy_fit()`:
+      ! Sparse data cannot be used with formula interface. Please use `fit_xy()` instead.
+
+# sparse tibble can be passed to `fit() - unsupported
 
     Code
       lm_fit <- fit(spec, avg_price_per_room ~ ., data = hotel_data[1:100, ])
     Condition
       Warning:
       `data` is a sparse tibble, but `linear_reg()` with engine "lm" doesn't accept that. Converting to non-sparse.
 
-# sparse matrix can be passed to `fit()
+# sparse matrix can be passed to `fit() - supported
+
+    Code
+      xgb_fit <- fit(spec, avg_price_per_room ~ ., data = hotel_data)
+    Condition
+      Error in `.convert_form_to_xy_fit()`:
+      ! Sparse data cannot be used with formula interface. Please use `fit_xy()` instead.
+
+# sparse matrix can be passed to `fit() - unsupported
 
     Code
       lm_fit <- fit(spec, avg_price_per_room ~ ., data = hotel_data[1:100, ])
     Condition
       Warning:
       `data` is a sparse tibble, but `linear_reg()` with engine "lm" doesn't accept that. Converting to non-sparse.
 
-# sparse tibble can be passed to `fit_xy()
+# sparse tibble can be passed to `fit_xy() - unsupported
 
     Code
       lm_fit <- fit_xy(spec, x = hotel_data[1:100, -1], y = hotel_data[1:100, 1])
     Condition
       Warning:
       `x` is a sparse tibble, but `linear_reg()` with engine "lm" doesn't accept that. Converting to non-sparse.
 
-# sparse matrices can be passed to `fit_xy()
+# sparse matrices can be passed to `fit_xy() - unsupported
 
     Code
       lm_fit <- fit_xy(spec, x = hotel_data[1:100, -1], y = hotel_data[1:100, 1])
     Condition
       Error in `fit_xy()`:
       ! `x` is a sparse matrix, but `linear_reg()` with engine "lm" doesn't accept that.
 
-# sparse tibble can be passed to `predict()
+# sparse tibble can be passed to `predict() - unsupported
 
     Code
       preds <- predict(lm_fit, sparse_mtcars)
     Condition
       Warning:
       `x` is a sparse tibble, but `linear_reg()` with engine "lm" doesn't accept that. Converting to non-sparse.
 
-# sparse matrices can be passed to `predict()
+# sparse matrices can be passed to `predict() - unsupported
 
     Code
       predict(lm_fit, sparse_mtcars)
     Condition
       Error in `predict()`:
       ! `x` is a sparse matrix, but `linear_reg()` with engine "lm" doesn't accept that.
 
+# sparse data work with xgboost engine
+
+    Code
+      xgb_fit <- fit(spec, avg_price_per_room ~ ., data = hotel_data)
+    Condition
+      Error in `.convert_form_to_xy_fit()`:
+      ! Sparse data cannot be used with formula interface. Please use `fit_xy()` instead.
+
 # to_sparse_data_frame() is used correctly
 
     Code

tests/testthat/helper-objects.R

@@ -28,7 +28,7 @@ is_tf_ok <- function() {
 # ------------------------------------------------------------------------------
 # For sparse tibble testing
 
-sparse_hotel_rates <- function() {
+sparse_hotel_rates <- function(tibble = FALSE) {
   # 99.2 sparsity
   hotel_rates <- modeldata::hotel_rates
 
@@ -49,5 +49,15 @@ sparse_hotel_rates <- function() {
   )
 
   res <- as.matrix(res)
-  Matrix::Matrix(res, sparse = TRUE)
+  res <- Matrix::Matrix(res, sparse = TRUE)
+
+  if (tibble) {
+    res <- sparsevctrs::coerce_to_sparse_tibble(res)
+
+    # materialize outcome
+    withr::local_options("sparsevctrs.verbose_materialize" = NULL)
+    res$avg_price_per_room <- res$avg_price_per_room[]
+  }
+
+  res
 }

tests/testthat/test-sparsevctrs.R

@@ -1,16 +1,23 @@
-test_that("sparse tibble can be passed to `fit()", {
+test_that("sparse tibble can be passed to `fit() - supported", {
   skip_if_not_installed("xgboost")
+  # Make materialization of sparse vectors throw an error
+  # https://r-lib.github.io/sparsevctrs/dev/reference/sparsevctrs_options.html
+  withr::local_options("sparsevctrs.verbose_materialize" = 3)
 
-  hotel_data <- sparse_hotel_rates()
-  hotel_data <- sparsevctrs::coerce_to_sparse_tibble(hotel_data)
+  hotel_data <- sparse_hotel_rates(tibble = TRUE)
 
   spec <- boost_tree() %>%
     set_mode("regression") %>%
     set_engine("xgboost")
-
-  expect_no_error(
-    lm_fit <- fit(spec, avg_price_per_room ~ ., data = hotel_data)
+
+  expect_snapshot(
+    error = TRUE,
+    xgb_fit <- fit(spec, avg_price_per_room ~ ., data = hotel_data)
   )
+})
+
+test_that("sparse tibble can be passed to `fit() - unsupported", {
+  hotel_data <- sparse_hotel_rates(tibble = TRUE)
 
   spec <- linear_reg() %>%
     set_mode("regression") %>%
@@ -21,19 +28,28 @@ test_that("sparse tibble can be passed to `fit()", {
   )
 })
 
-test_that("sparse matrix can be passed to `fit()", {
+test_that("sparse matrix can be passed to `fit() - supported", {
   skip_if_not_installed("xgboost")
-
+  # Make materialization of sparse vectors throw an error
+  # https://r-lib.github.io/sparsevctrs/dev/reference/sparsevctrs_options.html
+  withr::local_options("sparsevctrs.verbose_materialize" = 3)
+
   hotel_data <- sparse_hotel_rates()
-
+  
   spec <- boost_tree() %>%
     set_mode("regression") %>%
     set_engine("xgboost")
 
-  expect_no_error(
-    lm_fit <- fit(spec, avg_price_per_room ~ ., data = hotel_data)
+  expect_snapshot(
+    error = TRUE,
+    xgb_fit <- fit(spec, avg_price_per_room ~ ., data = hotel_data)
   )
 
+})
+
+test_that("sparse matrix can be passed to `fit() - unsupported", {
+  hotel_data <- sparse_hotel_rates()
+
   spec <- linear_reg() %>%
     set_mode("regression") %>%
     set_engine("lm")
@@ -43,19 +59,25 @@ test_that("sparse matrix can be passed to `fit()", {
   )
 })
 
-test_that("sparse tibble can be passed to `fit_xy()", {
+test_that("sparse tibble can be passed to `fit_xy() - supported", {
   skip_if_not_installed("xgboost")
-
-  hotel_data <- sparse_hotel_rates()
-  hotel_data <- sparsevctrs::coerce_to_sparse_tibble(hotel_data)
+  # Make materialization of sparse vectors throw an error
+  # https://r-lib.github.io/sparsevctrs/dev/reference/sparsevctrs_options.html
+  withr::local_options("sparsevctrs.verbose_materialize" = 3)
+
+  hotel_data <- sparse_hotel_rates(tibble = TRUE)
 
   spec <- boost_tree() %>%
     set_mode("regression") %>%
     set_engine("xgboost")
 
   expect_no_error(
-    lm_fit <- fit_xy(spec, x = hotel_data[, -1], y = hotel_data[, 1])
+    xgb_fit <- fit_xy(spec, x = hotel_data[, -1], y = hotel_data[, 1])
   )
+})
+
+test_that("sparse tibble can be passed to `fit_xy() - unsupported", {
+  hotel_data <- sparse_hotel_rates(tibble = TRUE)
 
   spec <- linear_reg() %>%
     set_mode("regression") %>%
@@ -66,8 +88,11 @@ test_that("sparse tibble can be passed to `fit_xy()", {
   )
 })
 
-test_that("sparse matrices can be passed to `fit_xy()", {
+test_that("sparse matrices can be passed to `fit_xy() - supported", {
   skip_if_not_installed("xgboost")
+  # Make materialization of sparse vectors throw an error
+  # https://r-lib.github.io/sparsevctrs/dev/reference/sparsevctrs_options.html
+  withr::local_options("sparsevctrs.verbose_materialize" = 3)
 
   hotel_data <- sparse_hotel_rates()
 
@@ -76,8 +101,12 @@ test_that("sparse matrices can be passed to `fit_xy()", {
     set_engine("xgboost")
 
   expect_no_error(
-    lm_fit <- fit_xy(spec, x = hotel_data[, -1], y = hotel_data[, 1])
+    xgb_fit <- fit_xy(spec, x = hotel_data[, -1], y = hotel_data[, 1])
   )
+})
+
+test_that("sparse matrices can be passed to `fit_xy() - unsupported", {
+  hotel_data <- sparse_hotel_rates()
 
   spec <- linear_reg() %>%
     set_mode("regression") %>%
@@ -89,11 +118,13 @@ test_that("sparse matrices can be passed to `fit_xy()", {
   )
 })
 
-test_that("sparse tibble can be passed to `predict()", {
+test_that("sparse tibble can be passed to `predict() - supported", {
   skip_if_not_installed("ranger")
+  # Make materialization of sparse vectors throw an error
+  # https://r-lib.github.io/sparsevctrs/dev/reference/sparsevctrs_options.html
+  withr::local_options("sparsevctrs.verbose_materialize" = 3)
 
-  hotel_data <- sparse_hotel_rates()
-  hotel_data <- sparsevctrs::coerce_to_sparse_tibble(hotel_data)
+  hotel_data <- sparse_hotel_rates(tibble = TRUE)
 
   spec <- rand_forest(trees = 10) %>%
     set_mode("regression") %>%
@@ -104,6 +135,10 @@ test_that("sparse tibble can be passed to `predict()", {
   expect_no_error(
     predict(tree_fit, hotel_data)
   )
+})
+
+test_that("sparse tibble can be passed to `predict() - unsupported", {
+  hotel_data <- sparse_hotel_rates(tibble = TRUE)
 
   spec <- linear_reg() %>%
     set_mode("regression") %>%
@@ -120,8 +155,11 @@ test_that("sparse tibble can be passed to `predict()", {
   )
 })
 
-test_that("sparse matrices can be passed to `predict()", {
+test_that("sparse matrices can be passed to `predict() - supported", {
   skip_if_not_installed("ranger")
+  # Make materialization of sparse vectors throw an error
+  # https://r-lib.github.io/sparsevctrs/dev/reference/sparsevctrs_options.html
+  withr::local_options("sparsevctrs.verbose_materialize" = 3)
 
   hotel_data <- sparse_hotel_rates()
 
@@ -134,6 +172,10 @@ test_that("sparse matrices can be passed to `predict()", {
   expect_no_error(
     predict(tree_fit, hotel_data)
   )
+})
+
+test_that("sparse matrices can be passed to `predict() - unsupported", {
+  hotel_data <- sparse_hotel_rates()
 
   spec <- linear_reg() %>%
     set_mode("regression") %>%
@@ -151,6 +193,9 @@ test_that("sparse matrices can be passed to `predict()", {
 
 test_that("sparse data work with xgboost engine", {
   skip_if_not_installed("xgboost")
+  # Make materialization of sparse vectors throw an error
+  # https://r-lib.github.io/sparsevctrs/dev/reference/sparsevctrs_options.html
+  withr::local_options("sparsevctrs.verbose_materialize" = 3)
 
   spec <- boost_tree() %>%
     set_mode("regression") %>%
@@ -159,35 +204,38 @@ test_that("sparse data work with xgboost engine", {
   hotel_data <- sparse_hotel_rates()
 
   expect_no_error(
-    tree_fit <- fit_xy(spec, x = hotel_data[, -1], y = hotel_data[, 1])
+    xgb_fit <- fit_xy(spec, x = hotel_data[, -1], y = hotel_data[, 1])
    )
-
+   
    expect_no_error(
-     predict(tree_fit, hotel_data)
+     predict(xgb_fit, hotel_data)
    )
 
-  hotel_data <- sparsevctrs::coerce_to_sparse_tibble(hotel_data)
+  hotel_data <- sparse_hotel_rates(tibble = TRUE)
 
-
-  expect_no_error(
-    tree_fit <- fit(spec, avg_price_per_room ~ ., data = hotel_data)
+  expect_snapshot(
+    error = TRUE,
+    xgb_fit <- fit(spec, avg_price_per_room ~ ., data = hotel_data)
   )
 
   expect_no_error(
-    predict(tree_fit, hotel_data)
+    predict(xgb_fit, hotel_data)
   )
 
   expect_no_error(
-   tree_fit <- fit_xy(spec, x = hotel_data[, -1], y = hotel_data[, 1])
+    xgb_fit <- fit_xy(spec, x = hotel_data[, -1], y = hotel_data[, 1])
   )
 
   expect_no_error(
-    predict(tree_fit, hotel_data)
+    predict(xgb_fit, hotel_data)
   )
 })
 
 test_that("to_sparse_data_frame() is used correctly", {
   skip_if_not_installed("xgboost")
+  # Make materialization of sparse vectors throw an error
+  # https://r-lib.github.io/sparsevctrs/dev/reference/sparsevctrs_options.html
+  withr::local_options("sparsevctrs.verbose_materialize" = 3)
 
   local_mocked_bindings(
     to_sparse_data_frame = function(x, object) {
@@ -228,6 +276,9 @@ test_that("to_sparse_data_frame() is used correctly", {
 
 test_that("maybe_sparse_matrix() is used correctly", {
   skip_if_not_installed("xgboost")
+  # Make materialization of sparse vectors throw an error
+  # https://r-lib.github.io/sparsevctrs/dev/reference/sparsevctrs_options.html
+  withr::local_options("sparsevctrs.verbose_materialize" = 3)
 
   local_mocked_bindings(
     maybe_sparse_matrix = function(x) {