R/interpolant.R

In chebpol: Multivariate Interpolation

#' Evaluate an interpolant in a point
#'
#' An interpolant is a function returned by \code{\link{ipol}} which has prespecified values in some
#' points, and which fills in between with some reasonable values.
#'
#' @name interpolant
#' @param x The argument of the function. A function of more then one variable takes a
#' vector. \code{x} can also be a matrix of column vectors.
#' @param threads The number of threads to use for evaluation. All  interpolants created by
#' \pkg{chebpol} are parallelized. If given a matrix argument \code{x}, the vectors can
#' be evaluated in parallel.
#' @param ... Other parameters. Currently used for simplex linear interpolants with the logical argument
#' \code{epol} which makes the interpolant extrapolate to points outside the domain.
#' The stalker spline has the argument \code{blend=c("linear","cubic","sigmoid")} where a
#' blending function can be chosen as described in a vignette. The \code{"multilinear"} interpolant also
#' has such a blending function.
#' @return A numeric. If more than one point was evaluated, a vector.
#' @examples
#' grid <- list(x=seq(0,1,length.out=10), y=seq(0,1,length.out=10))
#' val <- runif(100)
#' dim(val) <- c(10,10)
#' ip <- ipol(val, grid=grid, method='fh')
#' ip(c(0.3, 0.8))
#' ip(matrix(runif(12),2), threads=2)
NULL