AMR/R/freq.R

# ==================================================================== #
# TITLE                                                                #
# Antimicrobial Resistance (AMR) Analysis                              #
#                                                                      #
# SOURCE                                                               #
# https://gitlab.com/msberends/AMR                                     #
#                                                                      #
# LICENCE                                                              #
# (c) 2019 Berends MS (m.s.berends@umcg.nl), Luz CF (c.f.luz@umcg.nl)  #
#                                                                      #
# This R package is free software; you can freely use and distribute   #
# it for both personal and commercial purposes under the terms of the  #
# GNU General Public License version 2.0 (GNU GPL-2), as published by  #
# the Free Software Foundation.                                        #
#                                                                      #
# This R package was created for academic research and was publicly    #
# released in the hope that it will be useful, but it comes WITHOUT    #
# ANY WARRANTY OR LIABILITY.                                           #
# Visit our website for more info: https://msberends.gitlab.io/AMR.    #
# ==================================================================== #

#' Frequency table
#'
#' Create a frequency table of a vector with items or a \code{data.frame}. Supports quasiquotation and markdown for reports. Best practice is: \code{data \%>\% freq(var)}.\cr
#' \code{top_freq} can be used to get the top/bottom \emph{n} items of a frequency table, with counts as names.
#' @param x vector of any class or a \code{\link{data.frame}}, \code{\link{tibble}} (may contain a grouping variable) or \code{\link{table}}
#' @param ... up to nine different columns of \code{x} when \code{x} is a \code{data.frame} or \code{tibble}, to calculate frequencies from - see Examples. Also supports quasiquotion.
#' @param sort.count sort on count, i.e. frequencies. This will be \code{TRUE} at default for everything except when using grouping variables.
#' @param nmax number of row to print. The default, \code{15}, uses \code{\link{getOption}("max.print.freq")}. Use \code{nmax = 0}, \code{nmax = Inf}, \code{nmax = NULL} or \code{nmax = NA} to print all rows.
#' @param na.rm a logical value indicating whether \code{NA} values should be removed from the frequency table. The header (if set) will always print the amount of \code{NA}s.
#' @param row.names a logical value indicating whether row indices should be printed as \code{1:nrow(x)}
#' @param markdown a logical value indicating whether the frequency table should be printed in markdown format. This will print all rows (except when \code{nmax} is defined) and is default behaviour in non-interactive R sessions (like when knitting RMarkdown files).
#' @param digits how many significant digits are to be used for numeric values in the header (not for the items themselves, that depends on \code{\link{getOption}("digits")})
#' @param quote a logical value indicating whether or not strings should be printed with surrounding quotes
#' @param header a logical value indicating whether an informative header should be printed
#' @param title text to show above frequency table, at default to tries to coerce from the variables passed to \code{x}
#' @param na a character string that should be used to show empty (\code{NA}) values (only useful when \code{na.rm = FALSE})
#' @param droplevels a logical value indicating whether in factors empty levels should be dropped
#' @param sep a character string to separate the terms when selecting multiple columns
#' @inheritParams base::format
#' @param f a frequency table
#' @param n number of top \emph{n} items to return, use -n for the bottom \emph{n} items. It will include more than \code{n} rows if there are ties.
#' @param property property in header to return this value directly
#' @details Frequency tables (or frequency distributions) are summaries of the distribution of values in a sample. With the `freq` function, you can create univariate frequency tables. Multiple variables will be pasted into one variable, so it forces a univariate distribution. This package also has a vignette available to explain the use of this function further, run \code{browseVignettes("AMR")} to read it.
#'
#' For numeric values of any class, these additional values will all be calculated with \code{na.rm = TRUE} and shown into the header:
#' \itemize{
#'   \item{Mean, using \code{\link[base]{mean}}}
#'   \item{Standard Deviation, using \code{\link[stats]{sd}}}
#'   \item{Coefficient of Variation (CV), the standard deviation divided by the mean}
#'   \item{Mean Absolute Deviation (MAD), using \code{\link[stats]{mad}}}
#'   \item{Tukey Five-Number Summaries (minimum, Q1, median, Q3, maximum), using \code{\link[stats]{fivenum}}}
#'   \item{Interquartile Range (IQR) calculated as \code{Q3 - Q1} using the Tukey Five-Number Summaries, i.e. \strong{not} using the \code{\link[stats]{quantile}} function}
#'   \item{Coefficient of Quartile Variation (CQV, sometimes called coefficient of dispersion), calculated as \code{(Q3 - Q1) / (Q3 + Q1)} using the Tukey Five-Number Summaries}
#'   \item{Outliers (total count and unique count), using \code{\link[grDevices]{boxplot.stats}}}
#' }
#'
#' For dates and times of any class, these additional values will be calculated with \code{na.rm = TRUE} and shown into the header:
#' \itemize{
#'   \item{Oldest, using \code{\link{min}}}
#'   \item{Newest, using \code{\link{max}}, with difference between newest and oldest}
#'   \item{Median, using \code{\link[stats]{median}}, with percentage since oldest}
#' }
#'
#' In factors, all factor levels that are not existing in the input data will be dropped.
#'
#' The function \code{top_freq} uses \code{\link[dplyr]{top_n}} internally and will include more than \code{n} rows if there are ties.
#' @importFrom stats fivenum sd mad
#' @importFrom grDevices boxplot.stats
#' @importFrom dplyr %>% arrange arrange_at bind_cols desc filter_at funs group_by mutate mutate_at n n_distinct pull select summarise tibble ungroup vars all_vars
#' @importFrom utils browseVignettes
#' @importFrom hms is.hms
#' @importFrom crayon red green silver
#' @importFrom rlang enquos eval_tidy as_label
#' @keywords summary summarise frequency freq
#' @rdname freq
#' @name freq
#' @return A \code{data.frame} (with an additional class \code{"frequency_tbl"}) with five columns: \code{item}, \code{count}, \code{percent}, \code{cum_count} and \code{cum_percent}.
#' @export
#' @inheritSection AMR Read more on our website!
#' @examples
#' library(dplyr)
#'
#' # this all gives the same result:
#' freq(septic_patients$hospital_id)
#' freq(septic_patients[, "hospital_id"])
#' septic_patients$hospital_id %>% freq()
#' septic_patients[, "hospital_id"] %>% freq()
#' septic_patients %>% freq("hospital_id")
#' septic_patients %>% freq(hospital_id)  #<- easiest to remember (tidyverse)
#'
#'
#' # you could also use `select` or `pull` to get your variables
#' septic_patients %>%
#'   filter(hospital_id == "A") %>%
#'   select(mo) %>%
#'   freq()
#'
#'
#' # multiple selected variables will be pasted together
#' septic_patients %>%
#'   left_join_microorganisms %>%
#'   freq(genus, species)
#'
#' # functions as quasiquotation are also supported
#' septic_patients %>%
#'   freq(mo_genus(mo), mo_species(mo))
#'
#'
#' # group a variable and analyse another
#' septic_patients %>%
#'   group_by(hospital_id) %>%
#'   freq(gender)
#'
#'
#' # get top 10 bugs of hospital A as a vector
#' septic_patients %>%
#'   filter(hospital_id == "A") %>%
#'   freq(mo) %>%
#'   top_freq(10)
#'
#'
#' # save frequency table to an object
#' years <- septic_patients %>%
#'   mutate(year = format(date, "%Y")) %>%
#'   freq(year)
#'
#'
#' # show only the top 5
#' years %>% print(nmax = 5)
#'
#'
#' # save to an object with formatted percentages
#' years <- format(years)
#'
#'
#' # print a histogram of numeric values
#' septic_patients %>%
#'   freq(age) %>%
#'   hist()
#'
#'
#' # or print all points to a regular plot
#' septic_patients %>%
#'   freq(age) %>%
#'   plot()
#'
#'
#' # transform to a data.frame or tibble
#' septic_patients %>%
#'   freq(age) %>%
#'   as.data.frame()
#'
#'
#' # or transform (back) to a vector
#' septic_patients %>%
#'   freq(age) %>%
#'   as.vector()
#'
#' identical(septic_patients %>%
#'             freq(age) %>%
#'             as.vector() %>%
#'             sort(),
#'           sort(septic_patients$age)) # TRUE
#'
#'
#' # it also supports `table` objects
#' table(septic_patients$gender,
#'       septic_patients$age) %>%
#'   freq(sep = " **sep** ")
#'
#'
#' # only get selected columns
#' septic_patients %>%
#'   freq(hospital_id) %>%
#'   select(item, percent)
#'
#' septic_patients %>%
#'   freq(hospital_id) %>%
#'   select(-count, -cum_count)
#'
#'
#' # check differences between frequency tables
#' diff(freq(septic_patients$TMP),
#'      freq(septic_patients$SXT))
frequency_tbl <- function(x,
                          ...,
                          sort.count = TRUE,
                          nmax = getOption("max.print.freq"),
                          na.rm = TRUE,
                          row.names = TRUE,
                          markdown = !interactive(),
                          digits = 2,
                          quote = FALSE,
                          header = TRUE,
                          title = NULL,
                          na = "<NA>",
                          droplevels = TRUE,
                          sep = " ",
                          decimal.mark = getOption("OutDec"),
                          big.mark = ifelse(decimal.mark != ",", ",", ".")) {

  mult.columns <- 0
  x.group = character(0)
  df <- NULL
  x.name <- NULL
  cols <- NULL
  cols.names <- NULL
  if (any(class(x) == "list")) {
    cols <- names(x)
    x <- as.data.frame(x, stringsAsFactors = FALSE)
    x.name <- "a list"
  } else if (any(class(x) == "matrix")) {
    x <- as.data.frame(x, stringsAsFactors = FALSE)
    x.name <- "a matrix"
    cols <- colnames(x)
    if (all(cols %like% "V[0-9]")) {
      cols <- NULL
    }
  }

  if (any(class(x) == "data.frame")) {

    if (is.null(x.name)) {
      x.name <- deparse(substitute(x))
    }
    if (x.name %like% "(%>%)") {
      x.name <- x.name %>% strsplit("%>%", fixed = TRUE) %>% unlist() %>% .[1] %>% trimws()
    }
    if (x.name == ".") {
      x.name <- "a data.frame"
    } else {
      x.name <- paste0("`", x.name, "`")
    }
    x.name.dims <- x %>%
      dim() %>%
      format(decimal.mark = decimal.mark, big.mark = big.mark) %>%
      trimws() %>%
      paste(collapse = " x ")
    x.name <- paste0(x.name, " (", x.name.dims, ")")

    x.group <- group_vars(x)
    if (length(x.group) > 1) {
      x.group <- x.group[1L]
      warning("freq supports one grouping variable, only `", x.group, "` will be kept.", call. = FALSE)
    }

    user_exprs <- enquos(...)

    if (length(user_exprs) > 0) {
      new_list <- list(0)
      for (i in 1:length(user_exprs)) {
        new_list[[i]] <- tryCatch(eval_tidy(user_exprs[[i]], data = x),
                                  error = function(e) stop(e$message, call. = FALSE))
        if (length(new_list[[i]]) == 1) {
          if (is.character(new_list[[i]]) & new_list[[i]] %in% colnames(x)) {
            # support septic_patients %>% freq("hospital_id")
            new_list[[i]] <- x %>% pull(new_list[[i]])
          }
        }
        cols <- c(cols, as_label(user_exprs[[i]]))
      }

      if (length(new_list) == 1 & length(x.group) == 0) {
        # is now character
        x <- new_list[[1]]
        df <- NULL
      } else {
        # create data frame
        df <- as.data.frame(new_list, col.names = cols, stringsAsFactors = FALSE)
        cols.names <- colnames(df)
      }
    } else {
      # complete data frame
      df <- x
    }

    # support grouping variables
    if (length(x.group) > 0) {
      x.group_cols <- c(x.group, cols.names)
      x <- bind_cols(x, df)
      # if (droplevels == TRUE) {
      #   x <- x %>% mutate_at(vars(x.group_cols), droplevels)
      # }
      suppressWarnings(
        df <- x %>%
          group_by_at(vars(x.group_cols)) %>%
          summarise(count = n())
      )
      if (na.rm == TRUE) {
        df <- df %>% filter_at(vars(x.group_cols), all_vars(!is.na(.)))
      }
      if (!missing(sort.count)) {
        if (sort.count == TRUE) {
          df <- df %>% arrange_at(c(x.group_cols, "count"), desc)
        }
      }
      df <- df %>%
        mutate(cum_count = cumsum(count))

      df.topleft <- df[1, 1]
      df <- df %>%
        ungroup() %>%
        # do not repeat group labels
        mutate_at(vars(x.group), funs(ifelse(lag(.) == ., "", .)))
      df[1, 1] <- df.topleft
      colnames(df)[1:2] <- c("group", "item")

      if (!is.null(levels(df$item)) & droplevels == TRUE) {
        # is factor
        df <- df %>% filter(count != 0)
      }
    } else {
      if (!is.null(df)) {
        # no groups, multiple values like: septic_patients %>% freq(mo, mo_genus(mo))
        x <- df
        df <- NULL
      }
    }
    if (length(cols) > 0 & is.data.frame(x)) {
      x <- x[, cols.names]
    }

  } else if (any(class(x) == "table")) {
    x <- as.data.frame(x, stringsAsFactors = FALSE)
    # now this DF contains 3 columns: the 2 vars and a Freq column
    # paste the first 2 cols and repeat them Freq times:
    x <- rep(x = do.call(paste, c(x[colnames(x)[1:2]], sep = sep)),
             times = x$Freq)
    x.name <- "a `table` object"
    cols <- NULL
    # mult.columns <- 2
  } else {
    x.name <- deparse(substitute(x))
    if (x.name %like% "[$]") {
      cols <- unlist(strsplit(x.name, "$", fixed = TRUE))[2]
      x.name <- unlist(strsplit(x.name, "$", fixed = TRUE))[1]
      # try to find the object to determine dimensions

      x.obj <- tryCatch(get(x.name), error = function(e) NULL)
      x.name <- paste0("`", x.name , "`")
      if (!is.null(dim(x.obj))) {
        x.name <- paste0(x.name,
                         " (",
                         x.obj %>%
                           dim() %>%
                           format(decimal.mark = decimal.mark, big.mark = big.mark) %>%
                           trimws() %>%
                           paste(collapse = " x "),
                         ")")
      }
    } else {
      x.name <- NULL
      cols <- NULL
    }
  }

  if (!is.null(ncol(x))) {
    if (ncol(x) == 1 & any(class(x) == "data.frame")) {
      x <- x %>% pull(1)
    } else if (ncol(x) < 10) {
      mult.columns <- ncol(x)
      # paste old columns together
      x <- do.call(paste, c(x[colnames(x)], sep = sep))
    } else {
      stop("A maximum of 9 columns can be analysed at the same time.", call. = FALSE)
    }
  }

  if (mult.columns > 1) {
    NAs <- x[is.na(x) | x == trimws(strrep("NA ", mult.columns))]
  } else {
    NAs <- x[is.na(x)]
  }

  if (mult.columns > 0) {
    header_list <- list(columns = mult.columns)
  } else {
    header_list <- list(class = class(x),
                        mode = mode(x))
  }
  header_list$length <- length(x)

  if (na.rm == TRUE) {
    x_class <- class(x)
    x <- x[!x %in% NAs]
    class(x) <- x_class
  }

  markdown_line <- ""
  if (markdown == TRUE) {
    markdown_line <- "  "
  }
  x_align <- "l"

  if (!is.null(levels(x))) {
    header_list$levels <- levels(x)
    header_list$ordered <- is.ordered(x)
    # drop levels of non-existing factor values,
    # since dplyr >= 0.8.0 does not do this anymore in group_by
    if (droplevels == TRUE) {
      x <- droplevels(x)
    }
  }

  header_list$na_length <- length(NAs)
  header_list$unique <- n_distinct(x)

  if (NROW(x) > 0 & any(class(x) == "character")) {
    header_list$shortest <- x %>% base::nchar() %>% base::min(na.rm = TRUE)
    header_list$longest <- x %>% base::nchar() %>% base::max(na.rm = TRUE)
  }

  if (NROW(x) > 0 & any(class(x) == "mo")) {
    x_mo <- as.mo(x) # do it once for all three
    header_list$families <- x_mo %>% mo_family() %>% n_distinct()
    header_list$genera <- x_mo %>% mo_genus() %>% n_distinct()
    header_list$species <- x_mo %>% mo_species() %>% n_distinct()
  }

  if (NROW(x) > 0 & any(class(x) == "difftime") & !is.hms(x)) {
    header_list$units <- attributes(x)$units
    x <- as.double(x)
    # after this, the numeric header_txt continues
  }

  if (NROW(x) > 0 & any(class(x) %in% c("double", "integer", "numeric", "raw", "single"))) {
    # right align number
    x_align <- "r"
    header_list$mean <- base::mean(x, na.rm = TRUE)
    header_list$sd <- stats::sd(x, na.rm = TRUE)
    header_list$cv <- cv(x, na.rm = TRUE)
    header_list$mad <- stats::mad(x, na.rm = TRUE)
    Tukey_five <- stats::fivenum(x, na.rm = TRUE)
    header_list$fivenum <- Tukey_five
    header_list$IQR <- Tukey_five[4] - Tukey_five[2]
    header_list$cqv <- cqv(x, na.rm = TRUE)
    header_list$outliers_total <- length(boxplot.stats(x)$out)
    header_list$outliers_unique <- n_distinct(boxplot.stats(x)$out)
  }

  if (any(class(x) == "rsi")) {
    header_list$count_SI <- max(0, sum(x %in% c("S", "I"), na.rm = TRUE), na.rm = TRUE)
    header_list$count_R <- max(0, sum(x == "R", na.rm = TRUE), na.rm = TRUE)
  }

  formatdates <- "%e %B %Y" # = d mmmm yyyy
  if (is.hms(x)) {
    x <- x %>% as.POSIXlt()
    formatdates <- "%H:%M:%S"
  }
  if (NROW(x) > 0 & any(class(x) %in% c("Date", "POSIXct", "POSIXlt"))) {
    if (formatdates == "%H:%M:%S") {
      # hms
      header_list$earliest <- min(x, na.rm = TRUE)
      header_list$latest <- max(x, na.rm = TRUE)

    } else {
      # other date formats
      header_list$oldest <- min(x, na.rm = TRUE)
      header_list$newest <- max(x, na.rm = TRUE)
    }
    header_list$median <- median(x, na.rm = TRUE)
    header_list$date_format <- formatdates
  }
  if (any(class(x) == "POSIXlt")) {
    x <- x %>% format(formatdates)
  }

  nmax.set <- !missing(nmax)
  if (!nmax.set & is.null(nmax) & is.null(base::getOption("max.print.freq", default = NULL))) {
    # default for max print setting
    nmax <- 15
  } else if (is.null(nmax)) {
    nmax <- length(x)
  }

  if (nmax %in% c(0, Inf, NA, NULL)) {
    nmax <- length(x)
  }

  column_names <- c("Item", "Count", "Percent", "Cum. Count", "Cum. Percent")
  column_names_df <- c("item", "count", "percent", "cum_count", "cum_percent")
  column_align <- c(x_align, "r", "r", "r", "r")

  if (is.null(df)) {

    suppressWarnings( # suppress since dplyr 0.8.0, which idiotly warns about included NAs :(
      # create table with counts and percentages
      df <- tibble(item = x) %>%
        group_by(item) %>%
        summarise(count = n())
    )

    # sort according to setting
    if (sort.count == TRUE) {
      df <- df %>% arrange(desc(count), item)
    } else {
      df <- df %>% arrange(item)
    }
  } else {
    column_names <- c("Group", column_names)
    column_names_df <-c("group", column_names_df)
    column_align <- c("l", column_align)
  }

  if (df$item %>% paste(collapse = ",") %like% "\033") {
    # remove escape char
    # see https://en.wikipedia.org/wiki/Escape_character#ASCII_escape_character
    df <- df %>% mutate(item = item %>% gsub("\033", " ", ., fixed = TRUE))
  }

  if (quote == TRUE) {
    df$item <- paste0('"', df$item, '"')
    if (length(x.group) != 0) {
      df$group <- paste0('"', df$group, '"')
    }
  }

  df <- as.data.frame(df, stringsAsFactors = FALSE)

  df$percent <- df$count / base::sum(df$count, na.rm = TRUE)
  if (length(x.group) == 0) {
    df$cum_count <- base::cumsum(df$count)
  }
  df$cum_percent <- df$cum_count / base::sum(df$count, na.rm = TRUE)
  if (length(x.group) != 0) {
    # sort columns
    df <- df[, column_names_df]
  }

  if (markdown == TRUE) {
    tbl_format <- "markdown"
  } else {
    tbl_format <- "pandoc"
  }

  if (!is.null(title)) {
    title <- trimws(gsub("^Frequency table of", "", title[1L], ignore.case = TRUE))
  }

  # if (nmax.set == FALSE) {
  #   nmax <- nrow(df)
  # }

  structure(.Data = df,
            class = c("frequency_tbl", class(df)),
            header = header_list,
            opt = list(title = title,
                       data = x.name,
                       vars = cols,
                       group_var = x.group,
                       header = header,
                       row_names = row.names,
                       column_names = column_names,
                       column_align = column_align,
                       decimal.mark = decimal.mark,
                       big.mark = big.mark,
                       tbl_format = tbl_format,
                       na = na,
                       digits = digits,
                       nmax = nmax,
                       nmax.set = nmax.set))
}

#' @rdname freq
#' @export
freq <- frequency_tbl

#' @importFrom crayon silver green red
#' @importFrom dplyr %>%
format_header <- function(x, markdown = FALSE, decimal.mark = ".", big.mark = ",", digits = 2) {
  newline <-"\n"
  if (markdown == TRUE) {
    newline <- "  \n"
    # no colours in markdown
    silver <- function(x) x
    green <- function(x) x
    red <- function(x) x
  }

  header <- header(x)
  x_class <- header$class
  has_length <- header$length > 0

  # FORMATTING
  # rsi
  if (has_length == TRUE & any(x_class == "rsi")) {
    if (!is.null(attributes(x)$opt$vars)) {
      ab <- tryCatch(as.ab(attributes(x)$opt$vars), error = function(e) NA)
      if (!is.na(ab) & isTRUE(length(ab) > 0)) {
        header$drug <- paste0(ab_name(ab[1L]), " (", ab[1L], ", ", ab_atc(ab[1L]), ")")
        header$group <- ab_group(ab[1L])
      }
    }
    header$`%SI` <- percent(header$count_SI / (header$count_SI + header$count_R),
                            force_zero = TRUE, round = digits, decimal.mark = decimal.mark)
  }
  header <- header[!names(header) %in% c("count_SI", "count_R")]
  # dates
  if (!is.null(header$date_format)) {
    if (header$date_format == "%H:%M:%S") {
      header$median <- paste0(format(header$median, header$date_format),
                              " (",
                              (as.double(difftime(header$median, header$earliest, units = "auto")) /
                                 as.double(difftime(header$latest, header$earliest, units = "auto"))) %>%
                                percent(round = digits, decimal.mark = decimal.mark), ")")
      header$latest <- paste0(format(header$latest, header$date_format),
                              " (+",
                              difftime(header$latest, header$earliest, units = "mins") %>%
                                as.double() %>%
                                format(digits = digits, decimal.mark = decimal.mark, big.mark = big.mark),
                              " min.)")
      header$earliest <- format(header$earliest, header$date_format)

      header$median <- trimws(header$median)
      header$latest <- trimws(header$latest)
      header$earliest <- trimws(header$earliest)
    } else {
      header$median <- paste0(format(header$median, header$date_format),
                              " (",
                              (as.double(difftime(header$median, header$oldest, units = "auto")) /
                                 as.double(difftime(header$newest, header$oldest, units = "auto"))) %>%
                                percent(round = digits, decimal.mark = decimal.mark), ")")
      header$newest <- paste0(format(header$newest, header$date_format),
                              " (+",
                              difftime(header$newest, header$oldest, units = "auto") %>%
                                as.double() %>%
                                format(digits = digits, decimal.mark = decimal.mark, big.mark = big.mark),
                              ")")
      header$oldest <- format(header$oldest, header$date_format)

      header$median <- trimws(header$median)
      header$newest <- trimws(header$newest)
      header$oldest <- trimws(header$oldest)
    }
    header <- header[names(header) != "date_format"]
  }

  # class and mode
  if (is.null(header$columns)) {
    # if (markdown == TRUE) {
    #   header$class <- paste0("`", header$class, "`")
    # }
    if (!header$mode %in% header$class) {
    #   if (markdown == TRUE) {
    #     header$mode <- paste0("`", header$mode, "`")
    #   }
      header$class <- header$class %>% rev() %>% paste(collapse = " > ") %>% paste0(silver(paste0(" (", header$mode, ")")))
    } else {
      header$class <- header$class %>% rev() %>% paste(collapse = " > ")
    }
    header <- header[names(header) != "mode"]
  }
  # levels
  if (!is.null(header$levels)) {
    # if (markdown == TRUE) {
    #   header$levels <- paste0("`", header$levels, "`")
    # }
    if (header$ordered == TRUE) {
      levels_text <- paste0(header$levels, collapse = " < ")
    } else {
      levels_text <- paste0(header$levels, collapse = ", ")
    }
    if (nchar(levels_text) > 70) {
      # levels text wider than half the console
      levels_text <- paste0(substr(levels_text, 1, 70 - 3), "...")
      if (nchar(gsub("[^`]", "", levels_text)) %% 2 == 1) {
        # odd number of backticks, should be even
        levels_text <- paste0(levels_text, "`")
      }
    }
    header$levels <- paste0(length(header$levels), ": ", levels_text)
    header <- header[names(header) != "ordered"]
  }
  # length and NAs
  if (has_length == TRUE) {
    na_txt <- paste0(header$na_length %>% format(decimal.mark = decimal.mark, big.mark = big.mark), " = ",
                     (header$na_length / header$length) %>% percent(force_zero = TRUE, round = digits, decimal.mark = decimal.mark) %>%
                       sub("NaN", "0", ., fixed = TRUE))
    if (!na_txt %like% "^0 =") {
      na_txt <- red(na_txt)
    } else {
      na_txt <- green(na_txt)
    }
    na_txt <- paste0("(of which NA: ", na_txt, ")")
  } else {
    na_txt <- ""
  }
  header$length <- paste(format(header$length, decimal.mark = decimal.mark, big.mark = big.mark),
                         na_txt)
  header <- header[names(header) != "na_length"]

  # format all numeric values
  header <- lapply(header, function(x) {
    if (is.numeric(x)) {
      if (any(x < 1000, na.rm = TRUE)) {
        format(round2(x, digits = digits), decimal.mark = decimal.mark, big.mark = big.mark)
      } else {
        format(x, digits = digits, decimal.mark = decimal.mark, big.mark = big.mark)
      }
    } else {
      x
    }
  })

  # numeric values
  if (has_length == TRUE & any(x_class %in% c("double", "integer", "numeric", "raw", "single"))) {
    header$sd <- paste0(header$sd, " (CV: ", header$cv, ", MAD: ", header$mad, ")")
    header$fivenum <- paste0(paste(trimws(header$fivenum), collapse = " | "), " (IQR: ", header$IQR, ", CQV: ", header$cqv, ")")
    header$outliers_total <- paste0(header$outliers_total, " (unique count: ", header$outliers_unique, ")")
    header <- header[!names(header) %in% c("cv", "mad", "IQR", "cqv", "outliers_unique")]
  }

  # header names
  header_names <- paste0(names(header), ":  ")
  header_names <- gsub("sd", "SD", header_names)
  header_names <- gsub("fivenum", "Five-Num", header_names)
  header_names <- gsub("outliers_total", "Outliers", header_names)
  # capitalise first character
  header_names <- gsub("^(.)", "\\U\\1", header_names, perl = TRUE)
  # make all header captions equal size
  header_names <- gsub("\\s", " ", format(header_names,
                                          width = max(nchar(header_names),
                                                      na.rm = TRUE)))
  header <- paste0(header_names, header)
  header <- paste(header, collapse = newline)
  # add newline after 'Unique'
  gsub("(.*Unique.*\\n)(.*?)", paste0("\\1", newline, "\\2"), header)
}

#' @rdname freq
#' @export
#' @importFrom dplyr top_n pull
top_freq <- function(f, n) {
  if (!"frequency_tbl" %in% class(f)) {
    stop("`top_freq` can only be applied to frequency tables", call. = FALSE)
  }
  if (!is.numeric(n) | length(n) != 1L) {
    stop("For `top_freq`, 'n' must be a number of length 1", call. = FALSE)
  }
  top <- f %>% top_n(n, count)
  vect <- top %>% pull(item)
  names(vect) <- top %>% pull(count)
  if (length(vect) > abs(n)) {
    message("top_freq: selecting ", length(vect), " items instead of ", abs(n), ", because of ties")
  }
  vect
}

#' @rdname freq
#' @export
header <- function(f, property = NULL) {
  if (!"frequency_tbl" %in% class(f)) {
    stop("`header` can only be applied to frequency tables", call. = FALSE)
  }
  if (is.null(property)) {
    attributes(f)$header
  } else {
    a <- attributes(f)$header
    if (any(property %in% names(f))) {
      a[names(a) %in% property]
    }
  }
}

#' @noRd
#' @exportMethod diff.frequency_tbl
#' @importFrom dplyr %>% full_join mutate
#' @export
diff.frequency_tbl <- function(x, y, ...) {
  # check classes
  if (!"frequency_tbl" %in% class(x)
      | !"frequency_tbl" %in% class(y)) {
    stop("Both x and y must be a frequency table.")
  }

  cat("Differences between frequency tables")
  if (identical(x, y)) {
    cat("\n\nNo differences found.\n")
    return(invisible())
  }

  x.attr <- attributes(x)$opt

  # only keep item and count
  x <- x[, 1:2]
  y <- y[, 1:2]

  x <- x %>%
    full_join(y,
              by = colnames(x)[1],
              suffix = c(".x", ".y")) %>%
    mutate(
      diff = case_when(
        is.na(count.y) ~ -count.x,
        is.na(count.x) ~ count.y,
        TRUE ~ count.y - count.x)) %>%
    mutate(
      diff.percent = percent(
        diff / count.x,
        force_zero = TRUE)) %>%
    mutate(diff = ifelse(diff %like% "^-",
                         diff,
                         paste0("+", diff)),
           diff.percent = ifelse(diff.percent %like% "^-",
                                 diff.percent,
                                 paste0("+", diff.percent)))

  print(
    knitr::kable(x,
                 format = x.attr$tbl_format,
                 col.names = c("Item", "Count #1", "Count #2", "Difference", "Diff. percent"),
                 align = paste0(x.attr$column_align[1], "rrrr"),
                 padding = 1)
  )
}

#' @rdname freq
#' @exportMethod print.frequency_tbl
#' @importFrom knitr kable
#' @importFrom dplyr n_distinct
#' @importFrom crayon bold silver
#' @export
print.frequency_tbl <- function(x,
                                nmax = getOption("max.print.freq", default = 15),
                                markdown = !interactive(),
                                header = TRUE,
                                decimal.mark = getOption("OutDec"),
                                big.mark = ifelse(decimal.mark != ",", ",", "."),
                                ...) {

  opt <- attr(x, "opt")
  opt$header_txt <- header(x)

  dots <- list(...)
  if ("markdown" %in% names(dots)) {
    if (dots$markdown == TRUE) {
      opt$tbl_format <- "markdown"
    } else {
      opt$tbl_format <- "pandoc"
    }
  }
  if (!missing(markdown)) {
    if (markdown == TRUE) {
      opt$tbl_format <- "markdown"
    } else {
      opt$tbl_format <- "pandoc"
    }
  }

  if (length(opt$vars) == 0) {
    opt$vars <- NULL
  }

  if (is.null(opt$title)) {
    if (isTRUE(opt$data %like% "^a data.frame") & opt$tbl_format == "markdown") {
      opt$data <- gsub("data.frame", "`data.frame`", opt$data, fixed = TRUE)
    }
    if (!is.null(opt$data) & !is.null(opt$vars)) {
      title <- paste0("`", paste0(opt$vars, collapse = "` and `"), "` from ", opt$data)
    } else if (!is.null(opt$data) & is.null(opt$vars)) {
      title <- opt$data
    } else if (is.null(opt$data) & !is.null(opt$vars)) {
      title <- paste0("`", paste0(opt$vars, collapse = "` and `"), "`")
    } else {
      title <- ""
    }
    if (title != "" & length(opt$group_var) != 0) {
      group_var <- paste0("(grouped by `", opt$group_var, "`)")
      if (opt$tbl_format == "pandoc") {
        group_var <- silver(group_var)
      }
      title <- paste(title, group_var)
    }
    title <- trimws(title)
    if (title == "") {
      title <- "Frequency table"
    } else {
      title <- paste("Frequency table of", trimws(title))
    }
  } else {
    title <- opt$title
  }

  if (!missing(nmax)) {
    opt$nmax <- nmax
    opt$nmax.set <- TRUE
  }
  if (opt$nmax %in% c(0, Inf, NA, NULL)) {
    opt$nmax <- NROW(x)
    opt$nmax.set <- FALSE
  } else if (opt$nmax >= NROW(x)) {
    opt$nmax.set <- FALSE
  }

  if (!missing(decimal.mark)) {
    opt$decimal.mark <- decimal.mark
  }
  if (!missing(big.mark)) {
    opt$big.mark <- big.mark
  }
  if (!missing(header)) {
    opt$header <- header
  }

  # bold title
  if (opt$tbl_format == "pandoc") {
    title <- bold(title)
  } else if (opt$tbl_format == "markdown") {
    title <- paste0("\n\n**", title, "**  ") # two space for newline
  }

  cat(title, "\n\n")

  if (NROW(x) == 0 | isTRUE(all(is.na(x$item)))) {
    cat("No observations")
    if (isTRUE(all(is.na(x$item)))) {
      cat(" - all values are missing (<NA>)")
    }
    cat(".\n")
    if (opt$tbl_format == "markdown") {
      cat("\n")
    }
    return(invisible())
  }

  if (opt$header == TRUE) {
    if (!is.null(opt$header_txt)) {
      if (is.null(opt$digits)) {
        opt$digits <- 2
      }
      cat(format_header(x, digits = opt$digits, markdown = (opt$tbl_format == "markdown"),
                        decimal.mark = decimal.mark, big.mark = big.mark))
    }
  }

  # save old NA setting for kable
  opt.old <- options()$knitr.kable.NA
  if (is.null(opt$na)) {
    opt$na <- "<NA>"
  }
  if (opt$tbl_format == "markdown") {
    # no HTML tags
    opt$na <- gsub("<", "(", opt$na, fixed = TRUE)
    opt$na <- gsub(">", ")", opt$na, fixed = TRUE)
  }
  options(knitr.kable.NA = opt$na)

  x.rows <- nrow(x)
  x.unprinted <- base::sum(x[(opt$nmax + 1):nrow(x), "count"], na.rm = TRUE)
  x.printed <- base::sum(x$count) - x.unprinted

  if (nrow(x) > opt$nmax & opt$tbl_format != "markdown") {

    if (opt$nmax.set == TRUE) {
      nmax <- opt$nmax
    } else {
      nmax <- getOption("max.print.freq", default = 15)
    }

    x <- x[1:nmax,]

    if (opt$nmax.set == TRUE) {
      footer <- paste("[ reached `nmax = ", opt$nmax, "`", sep = "")
    } else {
      footer <- '[ reached getOption("max.print.freq")'
    }
    footer <- paste(footer,
                    " -- omitted ",
                    format(x.rows - opt$nmax, big.mark = opt$big.mark, decimal.mark = opt$decimal.mark),
                    " entries, n = ",
                    format(x.unprinted, big.mark = opt$big.mark, decimal.mark = opt$decimal.mark),
                    " (",
                    (x.unprinted / (x.unprinted + x.printed)) %>% percent(force_zero = TRUE, decimal.mark = opt$decimal.mark),
                    ") ]\n", sep = "")
    if (opt$tbl_format == "pandoc") {
      footer <- silver(footer) # only silver in regular printing
    }
  } else if (opt$tbl_format == "markdown") {
    if (opt$nmax.set == TRUE) {
      x <- x[1:opt$nmax,]
      footer <- paste("\n(omitted ",
                      format(x.rows - opt$nmax, big.mark = opt$big.mark, decimal.mark = opt$decimal.mark),
                      " entries, n = ",
                      format(x.unprinted, big.mark = opt$big.mark, decimal.mark = opt$decimal.mark),
                      " [",
                      (x.unprinted / (x.unprinted + x.printed)) %>% percent(force_zero = TRUE, decimal.mark = opt$decimal.mark),
                      "])\n", sep = "")
    } else {
      footer <- NULL
    }
  } else {
    footer <- NULL
  }

  if ("item" %in% colnames(x)) {
    if (any(class(x$item) %in% c("double", "integer", "numeric", "raw", "single"))) {
      x$item <- format(x$item, decimal.mark = opt$decimal.mark, big.mark = opt$big.mark)
    }
  } else {
    opt$column_names <- opt$column_names[!opt$column_names == "Item"]
  }
  if ("count" %in% colnames(x)) {
    if (all(x$count == 1)) {
      warning("All observations are unique.", call. = FALSE)
    }
    x$count <- format(x$count, decimal.mark = opt$decimal.mark, big.mark = opt$big.mark)
  } else {
    opt$column_names <- opt$column_names[!opt$column_names == "Count"]
  }
  if ("percent" %in% colnames(x)) {
    x$percent <- percent(x$percent, force_zero = TRUE, decimal.mark = opt$decimal.mark)
  } else {
    opt$column_names <- opt$column_names[!opt$column_names == "Percent"]
  }
  if ("cum_count" %in% colnames(x)) {
    x$cum_count <- format(x$cum_count, decimal.mark = opt$decimal.mark, big.mark = opt$big.mark)
  } else {
    opt$column_names <- opt$column_names[!opt$column_names == "Cum. Count"]
  }
  if ("cum_percent" %in% colnames(x)) {
    x$cum_percent <- percent(x$cum_percent, force_zero = TRUE, decimal.mark = opt$decimal.mark)
  } else {
    opt$column_names <- opt$column_names[!opt$column_names == "Cum. Percent"]
  }

  if (opt$tbl_format == "markdown") {
    cat("\n")
  }

  print(
    knitr::kable(x,
                 format = opt$tbl_format,
                 row.names = opt$row_names,
                 col.names = opt$column_names,
                 align = opt$column_align,
                 padding = 1)
  )

  if (!is.null(footer)) {
    cat(footer)
  }

  if (opt$tbl_format == "markdown") {
    cat("\n\n")
  } else {
    cat("\n")
  }

  # reset old kable setting
  options(knitr.kable.NA = opt.old)
  return(invisible())

}

#' @noRd
#' @exportMethod as.data.frame.frequency_tbl
#' @export
as.data.frame.frequency_tbl <- function(x, ...) {
  attr(x, "package") <- NULL
  attr(x, "opt") <- NULL
  as.data.frame.data.frame(x, ...)
}

#' @exportMethod select.frequency_tbl
#' @export
#' @importFrom dplyr select
#' @noRd
select.frequency_tbl <- function(.data, ...) {
  select(as.data.frame(.data), ...)
}

#' @noRd
#' @exportMethod as_tibble.frequency_tbl
#' @export
#' @importFrom dplyr as_tibble
as_tibble.frequency_tbl <- function(x, validate = TRUE, ..., rownames = NA) {
  attr(x, "package") <- NULL
  attr(x, "opt") <- NULL
  as_tibble(x = as.data.frame(x), validate = validate, ..., rownames = rownames)
}

#' @noRd
#' @exportMethod hist.frequency_tbl
#' @export
#' @importFrom graphics hist
hist.frequency_tbl <- function(x, breaks = "Sturges", main = NULL, xlab = NULL, ...) {
  opt <- attr(x, "opt")
  if (!class(x$item) %in% c("numeric", "double", "integer", "Date")) {
    stop("`x` must be numeric or Date.", call. = FALSE)
  }
  if (!is.null(opt$vars)) {
    title <- opt$vars
  } else if (!is.null(opt$data)) {
    title <- opt$data
  } else {
    title <- "frequency table"
  }
  if (class(x$item) == "Date") {
    x <- as.Date(as.vector(x), origin = "1970-01-01")
  } else {
    x <- as.vector(x)
  }
  if (is.null(main)) {
    main <- paste("Histogram of", title)
  }
  if (is.null(xlab)) {
    xlab <- title
  }
  hist(x, main = main, xlab = xlab, breaks = breaks, ...)
}

#' @noRd
#' @exportMethod plot.frequency_tbl
#' @export
plot.frequency_tbl <- function(x, y, ...) {
  opt <- attr(x, "opt")
  if (!is.null(opt$vars)) {
    title <- opt$vars
  } else {
    title <- ""
  }
  plot(x = x$item, y = x$count, ylab = "Count", xlab = title, ...)
}

#' @noRd
#' @exportMethod as.vector.frequency_tbl
#' @export
as.vector.frequency_tbl <- function(x, mode = "any") {
  as.vector(rep(x$item, x$count), mode = mode)
}

#' @noRd
#' @exportMethod format.frequency_tbl
#' @export
format.frequency_tbl <- function(x, digits = 1, ...) {
  opt <- attr(x, "opt")
  if (opt$nmax.set == TRUE) {
    nmax <- opt$nmax
  } else {
    nmax <- getOption("max.print.freq", default = 15)
  }

  x <- x[1:nmax,]
  x$percent <- percent(x$percent, round = digits, force_zero = TRUE)
  x$cum_percent <- percent(x$cum_percent, round = digits, force_zero = TRUE)
  base::format.data.frame(x, ...)
}