R/get_all_distances.R

In UBC-MDS/distrrr: Calculate Distance Metrics Between Vectors

#' Return distance/similarity metric for each row in a dataframe
#'
#' Compares an input reference vector (point) to all rows of an input data frame, calculating the specified distance/similarity metric for each row.
#'
#' @param point vector of length k to compare to the data frame
#' @param data data frame of size n by k to compare ref_vec to
#' @param metric string, default value "euclidean" for euclidean distance, "cosine" will return cosine similiarity distance, and "manhattan" will return manhattan distance
#'
#' @return numeric vector of length n
#'
#' @export
#'
#' @examples
#' df <- data.frame("A" = c(1,2,3), "B" = c(8,2,4))
#' ref_vec <- c(-2,4)
#' get_all_distances(ref_vec, df, metric = "euclidean")
#'

get_all_distances <- function(point, data, metric = "euclidean"){

  # raise error if point isn't class numeric
  if(class(point) != "numeric") stop("first argument point should be a numeric vector")

  # raise error if data isn't a data frame
  if(class(data) != "data.frame") stop("Second Argument data not a dataframe")

  # number of observations and columns in the data frame
  n <- dim(data)[1]
  k <- dim(data)[2]

  # raise error if point isn't length k
  if(length(point) != k) stop("Point should be length k, the number of columns of the input dataframe")

  # raise error if metric isn't correctly specified
  metrics <- c("euclidean","cosine","manhattan")
  if(!metric %in% metrics) stop("dist_type should be a string and one of 'cosine', 'euclidean' or 'manhattan'")

  # raise error if all elements of data aren't numeric
  # first extract type of each column in data frame
  # credit: https://stackoverflow.com/questions/21125222/determine-the-data-types-of-a-data-frames-columns
  data_col_classes <- unique(sapply(data,class))
  if(length(data_col_classes) != 1 & data_col_classes[1] != "numeric") stop("All elements of data should be numeric")

  # empty distances vector to fill
  distances <- numeric(n)

  for(obs in 1:n){
    point2 <- as.numeric(as.vector(data[obs,]))
    distances[obs] <- get_distance(point, point2, metric)
  }

  return(distances)
}