tour-little.r

#' A little tour path.
#'
#' The little tour is a planned tour that travels between all axis parallel
#' projections. (John McDonald named this type of tour.)
#'
#' Usually, you will not call this function directly, but will pass it to
#' a method that works with tour paths like \code{\link{animate}},
#' \code{\link{save_history}} or \code{\link{render}}.
#'
#' @param d target dimensionality
#' @export
#' @examples
#' animate_xy(flea[, 1:6], little_tour())
#' animate_pcp(flea[, 1:6], little_tour(3))
#' animate_scatmat(flea[, 1:6], little_tour(3))
#' animate_pcp(flea[, 1:6], little_tour(4))
little_tour <- function(d = 2) {
  little <- NULL
  step <- 0

  generator <- function(current, data, ...) {
    if (is.null(little)) {
      # Initialise bases
      little <<- bases_little(ncol(data), d)
    }
    step <<- (step %% length(little)) + 1
    if (is.null(current)) {
      # need an intercept to start the little tour
      return(little[[step]])
    }
    target <- little[[step]]
    list(target = target)
  }

  new_geodesic_path("little", generator)
}

#' Generate bases for the little tour
#'
#' @keywords internal
#' @param p dimensionality of data
#' @param d dimensionality of target projection
bases_little <- function(p, d = 2) {
  b <- diag(rep(1, p))
  vars <- utils::combn(p, d)
  # Let's make the list twice as long, so it visits pairs multiple
  # times, and in different order
  vars <- cbind(vars[, sample(1:ncol(vars))], vars[, sample(1:ncol(vars))])
  lapply(seq_len(ncol(vars)), function(i) b[, vars[, i]])
}