Evaluate component values in predictor expressions

In predictor expressions, name_eval(...) can be used to evaluate the effect of a component called "name".

Usage

component_eval(
  main,
  group = NULL,
  replicate = NULL,
  weights = NULL,
  .state = NULL
)

Arguments

main, group, replicate, weights

Specification of where to evaluate a component. The four inputs are passed on to the joint bru_mapper for the component, as

list(mapper = list(
       main = main,
       group = group,
       replicate = replicate),
     scale = weights)

.state

The internal component state. Normally supplied automatically by the internal methods for evaluating inlabru predictor expressions.

Value

A vector of values for a component

Examples

if (FALSE) {
if (bru_safe_inla()) {
  mesh <- INLA::inla.mesh.2d(
    cbind(0, 0),
    offset = 2, max.edge = 0.25
  )
  spde <- INLA::inla.spde2.pcmatern(mesh,
    prior.range = c(0.1, 0.01),
    prior.sigma = c(2, 0.01)
  )
  data <- sp::SpatialPointsDataFrame(
    matrix(runif(10), 5, 2),
    data = data.frame(z = rnorm(5))
  )
  fit <- bru(z ~ -1 + field(coordinates, model = spde),
    family = "gaussian", data = data
  )
  pred <- predict(
    fit,
    data = data.frame(x = 0.5, y = 0.5),
    formula = ~ field_eval(cbind(x, y))
  )
}
}