borg_register_hooks: Register BORG Hooks

In BORG: Bounded Outcome Risk Guard for Model Evaluation

Register BORG Hooks

Description

Registers BORG validation hooks that automatically check data dependencies when using common ML framework functions. This is an experimental feature.

Usage

borg_register_hooks(
  frameworks = c("rsample", "caret", "mlr3"),
  action = c("error", "warn", "message")
)

Arguments

frameworks	Character vector. Which frameworks to hook into. Options: "rsample", "caret", "mlr3". Default: all available.
action	Character. What to do when dependencies detected: "error" (block), "warn" (warn but proceed), "message" (info only).

Details

This function uses R's trace mechanism to add BORG checks to framework functions. The hooks are session-specific and do not persist.

To remove hooks, use borg_unregister_hooks().

Value

Invisible NULL. Called for side effect.

Examples

if (requireNamespace("rsample", quietly = TRUE)) {
  # Register hooks for rsample
  borg_register_hooks("rsample")

  # Now vfold_cv() will check for dependencies
  spatial_data <- data.frame(
    lon = runif(50), lat = runif(50), response = rnorm(50)
  )
  options(borg.check_data = spatial_data)
  options(borg.check_coords = c("lon", "lat"))

  # Remove hooks
  borg_unregister_hooks()
}

BORG documentation built on March 20, 2026, 5:09 p.m.