Convert to visualizer

as_visualizer() inspects its input and instantiates the matching visualizer class. All common arguments are validated first so that incompatible combinations fail fast with informative messages before any expensive rendering work is scheduled.

Usage

as_visualizer(
  x,
  type = "auto",
  x1_limits = NULL,
  x2_limits = NULL,
  padding = 0,
  n_points = 100L,
  y_pred = NULL,
  y_true = NULL,
  input_type = "auto",
  y_curves = "both",
  learner = NULL,
  hypothesis = NULL,
  domain = NULL,
  retrain = TRUE
)

# S3 method for class 'Task'
as_visualizer(
  x,
  type = "auto",
  x1_limits = NULL,
  x2_limits = NULL,
  padding = 0,
  n_points = 100L,
  y_pred = NULL,
  y_true = NULL,
  input_type = "auto",
  y_curves = "both",
  learner = NULL,
  hypothesis = NULL,
  domain = NULL,
  retrain = TRUE
)

# S3 method for class 'Hypothesis'
as_visualizer(
  x,
  type = "auto",
  x1_limits = NULL,
  x2_limits = NULL,
  padding = 0,
  n_points = 100L,
  y_pred = NULL,
  y_true = NULL,
  input_type = "auto",
  y_curves = "both",
  learner = NULL,
  hypothesis = NULL,
  domain = NULL,
  retrain = TRUE
)

# S3 method for class 'Objective'
as_visualizer(
  x,
  type = "auto",
  x1_limits = NULL,
  x2_limits = NULL,
  padding = 0,
  n_points = 100L,
  y_pred = NULL,
  y_true = NULL,
  input_type = "auto",
  y_curves = "both",
  learner = NULL,
  hypothesis = NULL,
  domain = NULL,
  retrain = TRUE
)

# S3 method for class 'LossFunction'
as_visualizer(
  x,
  type = "auto",
  x1_limits = NULL,
  x2_limits = NULL,
  padding = 0,
  n_points = 1000L,
  y_pred = NULL,
  y_true = NULL,
  input_type = "auto",
  y_curves = "both",
  learner = NULL,
  hypothesis = NULL,
  domain = NULL,
  retrain = TRUE
)

# S3 method for class 'list'
as_visualizer(
  x,
  type = "auto",
  x1_limits = NULL,
  x2_limits = NULL,
  padding = 0,
  n_points = 1000L,
  y_pred = NULL,
  y_true = NULL,
  input_type = "auto",
  y_curves = "both",
  learner = NULL,
  hypothesis = NULL,
  domain = NULL,
  retrain = TRUE
)

Arguments

x: (any)
Object to convert to a visualizer.
type: (character(1))
The type of visualization: "auto" (default), "1d", "2d", or "surface". If "auto", automatically chooses between 1D and 2D ggplot2 visualizations based on the number of features/dimensions. Use "surface" for interactive plotly surface plots (2D inputs only, Models and Objectives only).
x1_limits: (numeric(2))
The x1 limits.
x2_limits: (numeric(2))
The x2 limits.
padding: (numeric(1))
A margin that is added to x1limits and x2limits. The x1 margin is calculated by max(x1lmits) - min(x1limits) * padding.
n_points: (integer(1))
The number of generated point per dimension. Note that a grid of npoints^2 values is generated and evaluated by objective$eval(x) to plot the surface.
y_pred: (numeric()) Predicted values (for loss function visualizations).
y_true: (numeric()) True values (for loss function visualizations).
input_type: (character(1)) One of "auto" (default), "score", or "probability". Passed through to the loss visualizer.
y_curves: (character(1)) Which response curve(s) to draw when input_type = "probability". One of "both", "y1", or "y0".
learner: (mlr3::Learner)
The learner to train the model with.
hypothesis: (Hypothesis|NULL) Optional hypothesis object that provides predictions instead of a learner. Supply either learner or hypothesis, but not both.
domain: (list|NULL) Named list giving axis limits per predictor when no Task data is available, e.g., list(x = c(-1, 1)) for 1D or list(x = c(-1,1), y = c(-1,1)) for 2D.
retrain: (logical(1)) If a learner is supplied, whether it should be (re)trained on the task (TRUE, default) or reused as-is (FALSE). If FALSE but the learner is untrained, a warning is issued and training is performed.

Value

An object inheriting from a Visualizer class (e.g. VisualizerModel, VisualizerObj, or VisualizerLossFuns) depending on the input and selected type.

Details

By default (type = "auto"), as_visualizer() chooses a ggplot2 backend for 1D and 2D inputs. For 2D models or objectives you can opt into interactive plotly surfaces via type = "surface". All inputs are checked by a shared validation helper that enforces mutual exclusivity of learner/hypothesis, ensures limits are finite and ordered, and rejects parameters that are irrelevant for the chosen input.

Supported inputs

Validation

Common argument validation is centralized so that:

unused parameters (e.g. 2D limits supplied for 1D losses) raise an error,
limits supplied manually are finite, strictly ordered, and match the dimensionality,
high grid resolutions for interactive surfaces trigger explicit warnings, and
probability inputs for loss visualizations are checked for consistency of input_type, y_pred, y_true, and y_curves.

Examples

obj = obj("TF_branin")
vis = as_visualizer(obj, type = "2d")

# High resolutions on surfaces warn before plotting
tryCatch(
  as_visualizer(obj, type = "surface", n_points = 256L),
  warning = function(w) message("Warning: ", conditionMessage(w))
)
#> Warning: Requested surface grid with n_points = 256; values above 200 can cause slow interactive rendering.