tests/testthat/_snaps/nearest_neighbor.md

In tidymodels/parsnip: A Common API to Modeling and Analysis Functions

updating

Code
  nearest_neighbor(neighbors = 5) %>% set_engine("kknn", scale = FALSE) %>%
    update(neighbors = tune(), scale = tune())
Output
  K-Nearest Neighbor Model Specification (unknown mode)

  Main Arguments:
    neighbors = tune()

  Engine-Specific Arguments:
    scale = tune()

  Computational engine: kknn

bad input

Code
  nearest_neighbor(mode = "reallyunknown")
Condition
  Error in `nearest_neighbor()`:
  ! "reallyunknown" is not a known mode for model `nearest_neighbor()`.

check_args() works

Code
  spec <- nearest_neighbor(neighbors = -1) %>% set_engine("kknn") %>% set_mode(
    "classification")
  fit(spec, class ~ ., hpc)
Condition
  Error in `fit()`:
  ! `neighbors` must be a whole number larger than or equal to 0 or `NULL`, not the number -1.

Code
  spec <- nearest_neighbor(weight_func = 2) %>% set_engine("kknn") %>% set_mode(
    "classification")
  fit(spec, class ~ ., hpc)
Condition
  Error in `fit()`:
  ! `weight_func` must be a single string or `NULL`, not the number 2.

tidymodels/parsnip documentation built on Feb. 19, 2025, 2:10 a.m.