Build the constraint matrix and recipe

Usage

build_constraint_matrix(
  .data,
  .covariates,
  constraints,
  exposure_type,
  sampling_weights = NULL,
  call = rlang::caller_env()
)

.data: The data frame.
.covariates: A character vector of covariate column names.
constraints: A balance_terms specification, or NULL for the default first-moment constraints.
exposure_type: One of "binary", "categorical", or "continuous".
sampling_weights: Optional sampling weights. When supplied, numeric columns are standardized to weighted mean zero and unit weighted standard deviation rather than the unweighted sample scale.
call: The calling environment, used to build the error's call so a constraint error names the user-facing function.

A list with matrix and recipe.