R/cdb_id_stages.R

In jonesor/Rcompadre: Utilities for using the 'COM(P)ADRE' Matrix Model Database

#' Create integer identifiers for a COM(P)ADRE database corresponding to unique
#' combinations of species and matrix stage class definitions
#'
#' Creates a vector of integer identifiers corresponding to the rows of a
#' CompadreDB object, based on unique combinations of the column
#' 'SpeciesAccepted' and a list of matrix stage class definitions (either
#' 'MatrixClassAuthor' or 'MatrixClassOrganized').
#'
#' @param cdb A CompadreDB object
#' @param stage_def Whether to define matrix stage class based on
#'   "MatrixClassAuthor" or "MatrixClassOrganized" (see \emph{Details}).
#'   Defaults to "MatrixClassAuthor".
#'
#' @return Vector of integer identifiers corresponding to the rows of
#'   \code{cdb}.
#'
#' @author Patrick Barks <patrick.barks@@gmail.com>
#'
#' @family data management
#'
#' @details
#' The vector 'MatrixClassOrganized' reflects standardized stage classes
#' ('prop', 'active', or 'dorm'), whereas 'MatrixClassAuthor' reflects a
#' description of the stage classes as defined by the original author (e.g.
#' \code{c('Seedling', 'Medium rosette', 'Large (2 rosettes)', 'Flowering')}).
#'
#' Because the 'MatrixClassAuthor' definitions are less standardized, they are
#' more prone to typos that could lead to slight differences between stage
#' descriptions of matrices that really do have the same stage classes (e.g. a
#' set of matrices from a single study/species/population). Therefore, using
#' 'MatrixClassAuthor' to define stage classes is potentially prone to
#' mistakenly 'splitting' identifiers that should really be the same.
#'
#' 'MatrixClassOrganized' has the opposite problem. It's possible for two
#' matrices from a given study to have the same stage definitions based on
#' 'MatrixClassOrganized', but legitimately differ in stage definitions as
#' defined by the author. Therefore, using 'MatrixClassAuthor' to define stage
#' classes is potentially prone to mistakenly 'lumping' identifiers that should
#' actually differ.
#'
#' Because the majority of studies in COM(P)ADRE use a single set of stage
#' definitions for all matrices, and typos are rare, results for the different
#' stage definitions will usually be similar. Note, however, that the actual
#' integers returned for the different stage definitions are likely to be very
#' different (because they are based on alphabetical order).
#'
#' @seealso \link{cdb_id}
#'
#' @examples
#' cdb_id_stages(Compadre, stage_def = "MatrixClassOrganized")
#'
#' @export cdb_id_stages
cdb_id_stages <- function(cdb, stage_def = "MatrixClassAuthor") {
  if (!inherits(cdb, "CompadreDB")) {
    stop("db must be of class CompadreDB. See function as_cdb")
  }

  stage_def <- match.arg(
    stage_def,
    c("MatrixClassOrganized", "MatrixClassAuthor")
  )

  stage_fun <- match.fun(stage_def)
  cdb@data$stage_def <- stage_fun(cdb)
  cdb@data$stage_def <- vapply(cdb@data$stage_def, paste, "", collapse = ", ")

  cdb_id(cdb, columns = c("SpeciesAccepted", "stage_def"))
}