tests/testthat/_snaps/ncl.md

In themis: Extra Recipes Steps for Dealing with Unbalanced Data

errors when not enough observations for neighbors

Code
  bake(prep(step_ncl(recipe(class ~ x + y, data = small), class, neighbors = 5,
  skip = FALSE)), new_data = NULL)
Condition
  Error in `step_ncl()`:
  Caused by error in `bake()`:
  ! Not enough observations to compute 5 nearest neighbors.
  i 3 observations were found, but 6 are needed.

bad data

Code
  prep(step_ncl(rec, x))
Condition
  Error in `step_ncl()`:
  Caused by error in `prep()`:
  ! `x` should be a factor variable.

Code
  prep(step_ncl(rec, class, id))
Condition
  Error in `step_ncl()`:
  Caused by error in `prep()`:
  ! The selector should select at most a single variable.

errors if character are present

Code
  prep(step_ncl(recipe(~., data = df_char), x))
Condition
  Error in `step_ncl()`:
  Caused by error in `prep()`:
  x All columns selected for the step should be double or integer.
  * 1 factor variable found: `y`

NA in response

Code
  prep(step_ncl(recipe(Status ~ Age, data = credit_data0), Status))
Condition
  Error in `step_ncl()`:
  Caused by error in `prep()`:
  ! Cannot have any missing values. NAs found in Status.

distance_with errors on non-numeric column

Code
  prep(step_ncl(recipe(class ~ ., data = df_mixed), class, distance_with = c(x,
    name)))
Condition
  Error in `step_ncl()`:
  Caused by error in `prep()`:
  x All columns selected for the step should be double or integer.
  * 1 factor variable found: `name`

bad distance arg for step_ncl()

Code
  bake(prep(step_ncl(recipe(class ~ x + y, data = circle_example), class,
  distance = "L2")), new_data = NULL)
Condition
  Error in `step_ncl()`:
  ! `distance` must be one of "euclidean", "cosine", "mahalanobis", "manhattan", "chebyshev", "squared_chord", "matusita", "hellinger", "bhattacharyya", "canberra", "soergel", "lorentzian", "jeffreys", "topsoe", "jensen-shannon", "jensen_difference", "taneja", or "kumar-johnson", not "L2".

bad args

Code
  step_ncl(recipe(~., data = mtcars), seed = TRUE)
Condition
  Error in `step_ncl()`:
  ! `seed` must be a whole number, not `TRUE`.

Code
  prep(step_ncl(recipe(class ~ x + y, data = circle_example), class, neighbors = -
  1))
Condition
  Error in `step_ncl()`:
  Caused by error in `prep()`:
  ! `neighbors` must be a whole number larger than or equal to 1, not the number -1.

Code
  prep(step_ncl(recipe(class ~ x + y, data = circle_example), class,
  threshold_clean = -1))
Condition
  Error in `step_ncl()`:
  Caused by error in `prep()`:
  ! `threshold_clean` must be a number larger than or equal to 0, not the number -1.

unused outcome levels are skipped with a warning (#238)

Code
  res <- bake(prep(step_ncl(recipe(class ~ x + y, data = circle_example), class)),
  new_data = NULL)
Condition
  Warning in `prep()`:
  Unused factor level "unused" in `class` was dropped.
  i  Level with zero observations is skipped when computing sampling targets.

bake method errors when needed non-standard role columns are missing

Code
  bake(trained, new_data = circle_example[, -3])
Condition
  Error in `step_ncl()`:
  ! The following required column is missing from `new_data`: class.

empty printing

Code
  rec
Message

  -- Recipe ----------------------------------------------------------------------

  -- Inputs 
  Number of variables by role
  outcome:    1
  predictor: 10

  -- Operations 
  * NCL based on: <none>

Code
  rec
Message

  -- Recipe ----------------------------------------------------------------------

  -- Inputs 
  Number of variables by role
  outcome:    1
  predictor: 10

  -- Training information 
  Training data contained 32 data points and no incomplete rows.

  -- Operations 
  * NCL based on: <none> | Trained

printing

Code
  print(rec)
Message

  -- Recipe ----------------------------------------------------------------------

  -- Inputs 
  Number of variables by role
  outcome:   1
  predictor: 2

  -- Operations 
  * NCL based on: class

Code
  prep(rec)
Message

  -- Recipe ----------------------------------------------------------------------

  -- Inputs 
  Number of variables by role
  outcome:   1
  predictor: 2

  -- Training information 
  Training data contained 400 data points and no incomplete rows.

  -- Operations 
  * NCL based on: class | Trained

themis documentation built on Aug. 2, 2026, 9:07 a.m.