add labels_rsi_count

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: AMR
-Version: 0.3.0.9007
-Date: 2018-09-08
+Version: 0.3.0.9008
+Date: 2018-09-16
 Title: Antimicrobial Resistance Analysis
 Authors@R: c(
     person(

NAMESPACE

@@ -87,6 +87,7 @@ export(is.rsi.eligible)
 export(key_antibiotics)
 export(key_antibiotics_equal)
 export(kurtosis)
+export(labels_rsi_count)
 export(left_join_microorganisms)
 export(like)
 export(mo_aerobic)

NEWS.md

@@ -25,6 +25,7 @@
   * Column names inputs of `EUCAST_rules`, `first_isolate` and `key_antibiotics`
   * Column names of datasets `microorganisms` and `septic_patients`
   * All old syntaxes will still work with this version, but will throw warnings
+* Function `labels_rsi_count` to print datalabels on a RSI `ggplot2` model
 * Functions `as.atc` and `is.atc` to transform/look up antibiotic ATC codes as defined by the WHO. The existing function `guess_atc` is now an alias of `as.atc`.
 * Aliases for existing function `mo_property`: `mo_family`, `mo_genus`, `mo_species`, `mo_subspecies`, `mo_fullname`, `mo_shortname`, `mo_aerobic`, `mo_type` and `mo_gramstain`. They also come with support for German, Dutch, French, Italian, Spanish and Portuguese, and it defaults to the systems locale:
   ```r
@@ -70,6 +71,7 @@
   ```
 * Edited `ggplot_rsi` and `geom_rsi` so they can cope with `count_df`. The new `fun` parameter has value `portion_df` at default, but can be set to `count_df`.
 * Fix for `ggplot_rsi` when the `ggplot2` package was not loaded
+* Added datalabels function `labels_rsi_count` to `ggplot_rsi`
 * Added possibility to set any parameter to `geom_rsi` (and `ggplot_rsi`) so you can set your own preferences
 * Fix for joins, where predefined suffices would not be honoured
 * Added parameter `quote` to the `freq` function

R/ab_property.R

@@ -36,7 +36,7 @@
 #' ab_umcg("amcl")        # AMCL
 ab_property <- function(x, property = 'official') {
   property <- property[1]
-  if (!property %in% colnames(antibiotics)) {
+  if (!property %in% colnames(AMR::antibiotics)) {
     stop("invalid property: ", property, " - use a column name of the `antibiotics` data set")
   }
   if (!is.atc(x)) {

R/count.R

@@ -146,6 +146,10 @@ count_S <- function(...) {
 count_df <- function(data,
                      translate_ab = getOption("get_antibiotic_names", "official")) {
 
+  if (!"data.frame" %in% class(data)) {
+    stop("`count_df` must be called on a data.frame")
+  }
+
   if (data %>% select_if(is.rsi) %>% ncol() == 0) {
     stop("No columns with class 'rsi' found. See ?as.rsi.")
   }
@@ -177,7 +181,7 @@ count_df <- function(data,
 
   res <- bind_rows(resS, resI, resR) %>%
     mutate(Interpretation = factor(Interpretation, levels = c("R", "I", "S"), ordered = TRUE)) %>%
-    tidyr::gather(Antibiotic, Count, -Interpretation, -data.groups)
+    tidyr::gather(Antibiotic, Value, -Interpretation, -data.groups)
 
   if (!translate_ab == FALSE) {
     if (!tolower(translate_ab) %in% tolower(colnames(AMR::antibiotics))) {

R/ggplot_rsi.R

@@ -20,13 +20,16 @@
 #'
 #' Use these functions to create bar plots for antimicrobial resistance analysis. All functions rely on internal \code{\link[ggplot2]{ggplot}} functions.
 #' @param data a \code{data.frame} with column(s) of class \code{"rsi"} (see \code{\link{as.rsi}})
-#' @param position position adjustment of bars, either \code{"fill"}, \code{"stack"} (default when \code{fun} is \code{\link{portion_df}}) or \code{"dodge"} (default when \code{fun} is \code{\link{count_df}})
+#' @param position position adjustment of bars, either \code{"fill"} (default when \code{fun} is \code{\link{count_df}}), \code{"stack"} (default when \code{fun} is \code{\link{portion_df}}) or \code{"dodge"}
 #' @param x variable to show on x axis, either \code{"Antibiotic"} (default) or \code{"Interpretation"} or a grouping variable
 #' @param fill variable to categorise using the plots legend, either \code{"Antibiotic"} (default) or \code{"Interpretation"} or a grouping variable
 #' @param facet variable to split plots by, either \code{"Interpretation"} (default) or \code{"Antibiotic"} or a grouping variable
 #' @param translate_ab a column name of the \code{\link{antibiotics}} data set to translate the antibiotic abbreviations into, using \code{\link{abname}}. Default behaviour is to translate to official names according to the WHO. Use \code{translate_ab = FALSE} to disable translation.
-#' @param fun function to transform \code{data}, either \code{\link{portion_df}} (default) or \code{\link{count_df}}
+#' @param fun function to transform \code{data}, either \code{\link{count_df}} (default) or \code{\link{portion_df}}
 #' @param nrow (when using \code{facet}) number of rows
+#' @param datalabels show datalabels using \code{labels_rsi_count}, will at default only be shown when \code{fun = count_df}
+#' @param datalabels.size size of the datalabels
+#' @param datalabels.colour colour of the datalabels
 #' @param ... other parameters passed on to \code{geom_rsi}
 #' @details At default, the names of antibiotics will be shown on the plots using \code{\link{abname}}. This can be set with the option \code{get_antibiotic_names} (a logical value), so change it e.g. to \code{FALSE} with \code{options(get_antibiotic_names = FALSE)}.
 #'
@@ -35,12 +38,14 @@
 #'
 #' \code{facet_rsi} creates 2d plots (at default based on S/I/R) using \code{\link[ggplot2]{facet_wrap}}.
 #'
-#' \code{scale_y_percent} transforms the y axis to a 0 to 100\% range.
+#' \code{scale_y_percent} transforms the y axis to a 0 to 100\% range using \code{\link[ggplot2]{scale_y_continuous}}.
 #'
-#' \code{scale_rsi_colours} sets colours to the bars: green for S, yellow for I and red for R.
+#' \code{scale_rsi_colours} sets colours to the bars: green for S, yellow for I and red for R, using \code{\link[ggplot2]{scale_fill_brewer}}.
 #'
 #' \code{theme_rsi} is a \code{\link[ggplot2]{theme}} with minimal distraction.
 #'
+#' \code{labels_rsi_count} print datalabels on the bars with percentage and amount of isolates using \code{\link[ggplot2]{geom_text}}
+#'
 #' \code{ggplot_rsi} is a wrapper around all above functions that uses data as first input. This makes it possible to use this function after a pipe (\code{\%>\%}). See Examples.
 #' @rdname ggplot_rsi
 #' @export
@@ -58,6 +63,7 @@
 #'   geom_rsi() +
 #'   scale_y_percent() +
 #'   scale_rsi_colours() +
+#'   labels_rsi_count() +
 #'   theme_rsi()
 #'
 #' # or better yet, simplify this using the wrapper function - a single command:
@@ -131,8 +137,11 @@ ggplot_rsi <- function(data,
                        # params = list(),
                        facet = NULL,
                        translate_ab = "official",
-                       fun = portion_df,
+                       fun = count_df,
                        nrow = NULL,
+                       datalabels = TRUE,
+                       datalabels.size = 3,
+                       datalabels.colour = "grey15",
                        ...) {
 
   if (!"ggplot2" %in% rownames(installed.packages())) {
@@ -174,11 +183,22 @@ ggplot_rsi <- function(data,
     # set RSI colours
     p <- p + scale_rsi_colours()
   }
-  if (fun_name == "portion_df") {
+  if (is.null(position)) {
+    position <- "fill"
+  }
+  if (fun_name == "portion_df"
+      | (fun_name == "count_df" & position == "fill")) {
     # portions, so use y scale with percentage
     p <- p + scale_y_percent()
   }
 
+  if (fun_name == "count_df" & datalabels == TRUE) {
+    p <- p + labels_rsi_count(position = position,
+                              x = x,
+                              datalabels.size = datalabels.size,
+                              datalabels.colour = datalabels.colour)
+  }
+
   if (!is.null(facet)) {
     p <- p + facet_rsi(facet = facet, nrow = nrow)
   }
@@ -192,20 +212,19 @@ geom_rsi <- function(position = NULL,
                      x = c("Antibiotic", "Interpretation"),
                      fill = "Interpretation",
                      translate_ab = "official",
-                     fun = portion_df,
+                     fun = count_df,
                      ...)  {
 
   fun_name <- deparse(substitute(fun))
   if (!fun_name %in% c("portion_df", "count_df", "fun")) {
     stop("`fun` must be portion_df or count_df")
   }
+  y <- "Value"
   if (identical(fun, count_df)) {
-    y <- "Count"
     if (missing(position) | is.null(position)) {
-      position <- "dodge"
+      position <- "fill"
     }
   } else {
-    y <- "Percentage"
     if (missing(position) | is.null(position)) {
       position <- "stack"
     }
@@ -264,7 +283,6 @@ facet_rsi <- function(facet = c("Interpretation", "Antibiotic"), nrow = NULL) {
 #' @export
 scale_y_percent <- function() {
   ggplot2::scale_y_continuous(breaks = seq(0, 1, 0.1),
-                              limits = c(0, 1),
                               labels = percent(seq(0, 1, 0.1)))
 }
 
@@ -282,3 +300,35 @@ theme_rsi <- function() {
                    panel.grid.minor = ggplot2::element_blank(),
                    panel.grid.major.y = ggplot2::element_line(colour = "grey75"))
 }
+
+#' @rdname ggplot_rsi
+#' @export
+labels_rsi_count <- function(position = NULL,
+                             x = "Antibiotic",
+                             datalabels.size = 3,
+                             datalabels.colour = "grey15") {
+  if (is.null(position)) {
+    position <- "fill"
+  }
+  if (position == "fill") {
+    position <- ggplot2::position_fill(vjust = 0.5)
+  }
+  ggplot2::geom_text(mapping = ggplot2::aes_string(label = "lbl",
+                                                   x = x,
+                                                   y = "Value"),
+                     position = position,
+                     data = getlbls,
+                     inherit.aes = FALSE,
+                     size = datalabels.size,
+                     colour = datalabels.colour)
+}
+
+#' @importFrom dplyr %>% group_by mutate
+getlbls <- function(data) {
+  data %>%
+    count_df() %>%
+    group_by(Antibiotic) %>%
+    mutate(lbl = paste0(percent(Value / sum(Value, na.rm = TRUE), force_zero = TRUE),
+                        " (n=", Value, ")")) %>%
+    mutate(lbl = ifelse(lbl == "0.0% (n=0)", "", lbl))
+}

R/globals.R

@@ -22,7 +22,6 @@ globalVariables(c(".",
                   "antibiotics",
                   "cnt",
                   "count",
-                  "Count",
                   "cum_count",
                   "cum_percent",
                   "date_lab",
@@ -37,6 +36,7 @@ globalVariables(c(".",
                   "key_ab",
                   "key_ab_lag",
                   "key_ab_other",
+                  "lbl",
                   "median",
                   "mic",
                   "microorganisms",
@@ -46,7 +46,6 @@ globalVariables(c(".",
                   "other_pat_or_mo",
                   "Pasted",
                   "patient_id",
-                  "Percentage",
                   "prevalence",
                   "R",
                   "real_first_isolate",
@@ -54,4 +53,5 @@ globalVariables(c(".",
                   "septic_patients",
                   "species",
                   "value",
+                  "Value",
                   "y"))

R/mo_property.R

@@ -187,7 +187,7 @@ mo_aerobic <- function(x) {
 #' @export
 mo_property <- function(x, property = 'fullname', Becker = FALSE, Lancefield = FALSE, language = NULL) {
   property <- tolower(property[1])
-  if (!property %in% colnames(microorganisms)) {
+  if (!property %in% colnames(AMR::microorganisms)) {
     stop("invalid property: ", property, " - use a column name of the `microorganisms` data set")
   }
   result1 <- as.mo(x = x, Becker = Becker, Lancefield = Lancefield) # this will give a warning if x cannot be coerced

R/portion.R

@@ -203,6 +203,10 @@ portion_df <- function(data,
                        minimum = 30,
                        as_percent = FALSE) {
 
+  if (!"data.frame" %in% class(data)) {
+    stop("`portion_df` must be called on a data.frame")
+  }
+
   if (data %>% select_if(is.rsi) %>% ncol() == 0) {
     stop("No columns with class 'rsi' found. See ?as.rsi.")
   }
@@ -240,7 +244,7 @@ portion_df <- function(data,
 
   res <- bind_rows(resS, resI, resR) %>%
     mutate(Interpretation = factor(Interpretation, levels = c("R", "I", "S"), ordered = TRUE)) %>%
-    tidyr::gather(Antibiotic, Percentage, -Interpretation, -data.groups)
+    tidyr::gather(Antibiotic, Value, -Interpretation, -data.groups)
 
   if (!translate_ab == FALSE) {
     if (!tolower(translate_ab) %in% tolower(colnames(AMR::antibiotics))) {

README.md

@@ -161,7 +161,8 @@ Adjust it with any parameter you know from the `ggplot2` package:
 ```r
 septic_patients %>%
   select(amox, nitr, fosf, trim, cipr) %>%
-  ggplot_rsi(width = 0.5, colour = "black", size = 1, linetype = 2, alpha = 0.25)
+  ggplot_rsi(datalabels = FALSE, 
+             width = 0.5, colour = "black", size = 1, linetype = 2, alpha = 0.25)
 ```
 
 ![example_3_rsi](man/figures/rsi_example3.png)
@@ -174,7 +175,8 @@ septic_patients %>%
   group_by(hospital_id) %>%
   ggplot_rsi(x = "hospital_id",
              facet = "Antibiotic",
-             nrow = 1) +
+             nrow = 1,
+             datalabels = FALSE) +
   labs(title = "AMR of Anti-UTI Drugs Per Hospital",
        x = "Hospital")
 ```

man/ggplot_rsi.Rd

@@ -7,15 +7,17 @@
 \alias{scale_y_percent}
 \alias{scale_rsi_colours}
 \alias{theme_rsi}
+\alias{labels_rsi_count}
 \title{AMR bar plots with \code{ggplot}}
 \usage{
 ggplot_rsi(data, position = NULL, x = "Antibiotic",
   fill = "Interpretation", facet = NULL, translate_ab = "official",
-  fun = portion_df, nrow = NULL, ...)
+  fun = count_df, nrow = NULL, datalabels = TRUE,
+  datalabels.size = 3, datalabels.colour = "grey15", ...)
 
 geom_rsi(position = NULL, x = c("Antibiotic", "Interpretation"),
-  fill = "Interpretation", translate_ab = "official",
-  fun = portion_df, ...)
+  fill = "Interpretation", translate_ab = "official", fun = count_df,
+  ...)
 
 facet_rsi(facet = c("Interpretation", "Antibiotic"), nrow = NULL)
 
@@ -24,11 +26,14 @@ scale_y_percent()
 scale_rsi_colours()
 
 theme_rsi()
+
+labels_rsi_count(position = NULL, x = "Antibiotic",
+  datalabels.size = 3, datalabels.colour = "grey15")
 }
 \arguments{
 \item{data}{a \code{data.frame} with column(s) of class \code{"rsi"} (see \code{\link{as.rsi}})}
 
-\item{position}{position adjustment of bars, either \code{"fill"}, \code{"stack"} (default when \code{fun} is \code{\link{portion_df}}) or \code{"dodge"} (default when \code{fun} is \code{\link{count_df}})}
+\item{position}{position adjustment of bars, either \code{"fill"} (default when \code{fun} is \code{\link{count_df}}), \code{"stack"} (default when \code{fun} is \code{\link{portion_df}}) or \code{"dodge"}}
 
 \item{x}{variable to show on x axis, either \code{"Antibiotic"} (default) or \code{"Interpretation"} or a grouping variable}
 
@@ -38,10 +43,16 @@ theme_rsi()
 
 \item{translate_ab}{a column name of the \code{\link{antibiotics}} data set to translate the antibiotic abbreviations into, using \code{\link{abname}}. Default behaviour is to translate to official names according to the WHO. Use \code{translate_ab = FALSE} to disable translation.}
 
-\item{fun}{function to transform \code{data}, either \code{\link{portion_df}} (default) or \code{\link{count_df}}}
+\item{fun}{function to transform \code{data}, either \code{\link{count_df}} (default) or \code{\link{portion_df}}}
 
 \item{nrow}{(when using \code{facet}) number of rows}
 
+\item{datalabels}{show datalabels using \code{labels_rsi_count}, will at default only be shown when \code{fun = count_df}}
+
+\item{datalabels.size}{size of the datalabels}
+
+\item{datalabels.colour}{colour of the datalabels}
+
 \item{...}{other parameters passed on to \code{geom_rsi}}
 }
 \description{
@@ -55,12 +66,14 @@ At default, the names of antibiotics will be shown on the plots using \code{\lin
 
 \code{facet_rsi} creates 2d plots (at default based on S/I/R) using \code{\link[ggplot2]{facet_wrap}}.
 
-\code{scale_y_percent} transforms the y axis to a 0 to 100\% range.
+\code{scale_y_percent} transforms the y axis to a 0 to 100\% range using \code{\link[ggplot2]{scale_y_continuous}}.
 
-\code{scale_rsi_colours} sets colours to the bars: green for S, yellow for I and red for R.
+\code{scale_rsi_colours} sets colours to the bars: green for S, yellow for I and red for R, using \code{\link[ggplot2]{scale_fill_brewer}}.
 
 \code{theme_rsi} is a \code{\link[ggplot2]{theme}} with minimal distraction.
 
+\code{labels_rsi_count} print datalabels on the bars with percentage and amount of isolates using \code{\link[ggplot2]{geom_text}}
+
 \code{ggplot_rsi} is a wrapper around all above functions that uses data as first input. This makes it possible to use this function after a pipe (\code{\%>\%}). See Examples.
 }
 \examples{
@@ -77,6 +90,7 @@ ggplot(df) +
   geom_rsi() +
   scale_y_percent() +
   scale_rsi_colours() +
+  labels_rsi_count() +
   theme_rsi()
 
 # or better yet, simplify this using the wrapper function - a single command:

tests/testthat/test-count.R

@@ -41,4 +41,6 @@ test_that("counts work", {
   expect_error(count_S("test", minimum = "test"))
   expect_error(count_S("test", as_percent = "test"))
 
+  expect_error(count_df(c("A", "B", "C")))
+
 })

tests/testthat/test-portion.R

@@ -112,7 +112,7 @@ test_that("old rsi works", {
 
   # portion_df
   expect_equal(
-    septic_patients %>% select(amox) %>% portion_df() %>% pull(Percentage),
+    septic_patients %>% select(amox) %>% portion_df() %>% pull(Value),
     c(septic_patients$amox %>% portion_S(),
       septic_patients$amox %>% portion_I(),
       septic_patients$amox %>% portion_R())
@@ -165,4 +165,6 @@ test_that("prediction of rsi works", {
                                   col_ab = "mero",
                                   col_date = "date",
                                   info = TRUE))
+
+  expect_error(portion_df(c("A", "B", "C")))
 })