R/data.R

In tkonopka/shrt: short R code

# Collection of functions with short names - data manipulation
#


#' compute Jaccard Index of two sets
#'
#' The Jaccard Index of two sets is defined as the ratio of the
#' size of the set intersection (common element) and the
#' size of the set union (all elements).
#'
#' The name is short for: (j)accard (i)ndex
#'
#' @param a vector of set elements
#' @param b vector of set elements
#'
#' @examples
#' ji(c(1,2,3), c(1,4,5)) # one common element out of five, JI=0.2 
#'
#' @export
ji = function(a, b) {
  length(intersect(a,b)) / length(union(a, b))
}


#' compute Overlap Index of two sets
#'
#' The Overlap Index of two sets is defined as the ratio of the
#' size of the set intersection (common elements) and the size
#' of the primary set.
#'
#' @param a vector of set elements
#' @param b vector of set elements
#'
#' @examples
#' oi(c(1,2), c(2,3,4,5)) # one element out of two in "a" is present in "b"
#'
#' @export
oi = function(a, b) {
  length(intersect(a, b)) / length(unique(a))
}


#' Characterize memberships in two sets
#'
#' @param a vector (character, numeric, etc.) with members of set a
#' @param b vector with members of set b
#' @param unique.force logical, set TRUE to enforce that a, b have
#' non-repeating elements (i.e. they are sets), set FALSE to use
#' a,b as-is, i.e. potentially with non-unique elements.
#'
#'
#' @return list holding three vectors and a summary. Vectors contain
#' the set elements that are private and common in a, b, and a vector
#' with membership counts. Warning, membership counts might not be
#' reliable if a, b have duplicate entries
#'
#' @export
setsummary = function(a, b, unique.force=TRUE) {
  
  # perhaps enforce a, b to be sets
  if (unique.force) {
    a = unique(a)
    b = unique(b)
  }
  
  ## compute intersection, a/b unique set members
  ab = intersect(a, b)
  
  a.only = a[!a %in% ab]
  b.only = b[!b %in% ab]
  summary = c("a.private" = length(a.only),
              "b.private" = length(b.only),
              "common" = length(ab))
  
  list(a.private = a.only, b.private = b.only,
       common = ab, summary=summary)
}