cute_little_R_functions.R 501 KB
Newer Older
Gael  MILLOT's avatar
Gael MILLOT committed
4001
4002
4003
4004
4005
4006
4007
4008
4009
4010
4011
4012
4013
4014
4015
4016
4017
4018
4019
4020
4021
4022
4023
4024
4025
4026
4027
}
}
}
# no need loop part
# end no need loop part
suppressWarnings(print(eval(parse(text = paste(paste0(tempo.gg.name, 1:tempo.gg.count), collapse = " + ")))))
if(return == TRUE){
output <- ggplot2::ggplot_build(eval(parse(text = paste(paste0(tempo.gg.name, 1:tempo.gg.count), collapse = " + "))))
if(is.null(unlist(removed.row.nb))){
removed.row.nb <- NULL
removed.rows <- NULL
}else{
for(i3 in 1:length(data1)){
if( ! is.null(removed.row.nb[[i3]])){
removed.row.nb[[i3]] <- sort(removed.row.nb[[i3]])
removed.rows[[i3]] <- data1.ini[[i3]][removed.row.nb[[i3]], ]
}
}
}
output <- list(data = output$data, removed.row.nb = removed.row.nb, removed.rows = removed.rows, warnings = paste0("\n", warning, "\n\n"))
return(output)
}
}


######## fun_gg_bar_mean() #### ggplot2 mean barplot + overlaid dots if required

Gael  MILLOT's avatar
Gael MILLOT committed
4028
4029
# nice breaks

Gael  MILLOT's avatar
Gael MILLOT committed
4030
4031
4032
4033
4034

  


# Check OK: clear to go Apollo
Gael  MILLOT's avatar
Gael MILLOT committed
4035
fun_gg_bar_mean <- function(data1, y, categ, categ.class.order = NULL, categ.legend.name = NULL, categ.color = NULL, bar.width = 0.5, error.disp = NULL, error.whisker.width = 0.5,  dot.color = "same", dot.tidy = FALSE, dot.bin.nb = 30, dot.jitter = 0.25, dot.size = 3, dot.border.size = 0.5, dot.alpha = 0.5, ylim = NULL, ylog = FALSE, y.break.nb = NULL, y.include.zero = FALSE, y.top.extra.margin = 0.05, y.bottom.extra.margin = 0, stat.disp = NULL, stat.size = 4, stat.dist = 2, xlab = NULL, ylab = NULL, vertical = TRUE, title = "", text.size = 12, text.angle = 0, classic = FALSE, grid = FALSE, return = FALSE, path.lib = NULL){
Gael  MILLOT's avatar
Gael MILLOT committed
4036
4037
4038
4039
4040
4041
4042
4043
4044
# AIM
# ggplot2 vertical barplot representing mean values with the possibility to add error bars and to overlay dots
# for ggplot2 specifications, see: https://ggplot2.tidyverse.org/articles/ggplot2-specs.html
# WARNINGS
# rows containing NA in data1[, c(y, categ)] will be removed before processing, with a warning (see below)
# if ever bars disappear, see the end of https://github.com/tidyverse/ggplot2/issues/2887
# to have a single bar, create a factor column with a single class and specify the name of this column in categ argument as unique element (no categ2 in categ argument). For a single set of grouped bars, create a factor column with a single class and specify this column in categ argument as first element (categ1). See categ below
# with several single bars (categ argument with only one element), bar.width argument (i.e., width argument of ggplot2::geom_bar()) defines each bar width. The bar.width argument also defines the space between bars by using (1 - bar.width). In addition, xmin and xmax of the fun_gg_bar_mean() output report the bar boundaries (around x-axis unit 1, 2, 3, etc., for each bar)
# with several sets of grouped bars (categ argument with two elements), bar.width argument defines each set of grouped bar width. The bar.width argument also defines the space between set of grouped bars by using (1 - bar.width). In addition, xmin and xmax of the fun_gg_bar_mean() output report the bar boundaries (around x-axis unit 1, 2, 3, etc., for each set of grouped bar)
Gael  MILLOT's avatar
Gael MILLOT committed
4045
# to manually change the 0 base bar into this code, see https://stackoverflow.com/questions/35324892/ggplot2-setting-geom-bar-baseline-to-1-instead-of-zero
Gael  MILLOT's avatar
Gael MILLOT committed
4046
# ARGUMENTS
Gael  MILLOT's avatar
Gael MILLOT committed
4047
# data1: a dataframe containing one column of values (see y argument below) and one or two columns of categories (see categ argument below). Duplicated column names not allowed
Gael  MILLOT's avatar
Gael MILLOT committed
4048
4049
4050
4051
4052
4053
4054
4055
4056
4057
4058
4059
4060
# y: character string of the data1 column name for y-axis (containing numeric values). Numeric values will be averaged by categ to generate the bars and will also be used to plot the dots
# categ: vector of character strings of the data1 column name for categories (column of characters or factor). Must either be one or two column names. If a single column name (further refered to as categ1), then one bar per class of categ1. If two column names (further refered to as categ1 and categ2), then one bar per class of categ2, which form a group of bars in each class of categ1. Beware, categ1 (and categ2 if it exists) must have a single value of y per class of categ1 (and categ2). To have a single bar, create a factor column with a single class and specify the name of this column in categ argument as unique element (no categ2 in categ argument). For a single set of grouped bars, create a factor column with a single class and specify this column in categ argument as first element (categ1)
# categ.class.order: list indicating the order of the classes of categ1 and categ2 represented on the barplot (the first compartment for categ1 and and the second for categ2). If categ.class.order = NULL, classes are represented according to the alphabetical order. Some compartment can be NULL and other not
# categ.legend.name: character string of the legend title for categ2. If categ.legend.name = NULL, then categ.legend.name <- categ1 if only categ1 is present and categ.legend.name <- categ2 if categ1 and categ2 are present. Write "" if no legend required
# categ.color: vector of character color string for bar filling. If categ.color = NULL, default colors of ggplot2, whatever categ1 and categ2. If categ.color is non null and only categ1 in categ argument, categ.color can be either: (1) a single color string (all the bars will have this color, whatever the classes of categ1), (2) a vector of string colors, one for each class of categ1 (each color will be associated according to categ.class.order of categ1), (3) a vector or factor of string colors, like if it was one of the column of data1 data frame (beware: a single color per class of categ1 and a single class of categ1 per color must be respected). Integers are also accepted instead of character strings, as long as above rules about length are respected. Integers will be processed by fun_gg_palette() using the max integer value among all the integers in categ.color. If categ.color is non null and categ1 and categ2 specified, all the rules described above will apply to categ2 instead of categ1 (colors will be determined for bars inside a group of bars)
# bar.width: numeric value (from 0 to 1) of the bar or set of grouped bar width (see warnings above)
# error.disp: either "SD", "SD.TOP", "SEM" or "SEM.TOP". If NULL, no error bars added
# error.whisker.width: numeric value (from 0 to 1) of the whisker (error bar extremities) width, with 0 meaning no whiskers and 1 meaning a width equal to the corresponding bar width
# dot.color: vector of character string. Idem as categ.color but for dots, except that in the possibility (3), the rule "a single color per class of categ1 and a single class of categ1", cannot be respected (each dot can have a different color). If NULL, no dots plotted
# dot.tidy: logical. Nice dot spreading? If TRUE, use the geom_dotplot() function for a nice representation. If FALSE, dots are randomly spread, using the dot.jitter argument (see below)
# dot.bin.nb: positive integer indicating the number of bins (i.e., nb of separations) of the ylim range. Each dot will then be put in one of the bin, with the size the width of the bin. Not considered if dot.tidy is FALSE
# dot.jitter: numeric value (from 0 to 1) of random dot horizontal dispersion, with 0 meaning no dispersion and 1 meaning a dispersion in the corresponding bar width interval. Not considered if dot.tidy is TRUE
# dot.size: numeric value of dot size. Not considered if dot.tidy is TRUE
Gael  MILLOT's avatar
Gael MILLOT committed
4061
# dot.border.size: numeric value of border dot size. Write zero for no dot border. If dot.tidy is TRUE, value 0 remove the border. Another one leave the border without size control (geom_doplot() feature)
Gael  MILLOT's avatar
Gael MILLOT committed
4062
4063
4064
4065
4066
4067
4068
4069
4070
4071
4072
4073
4074
4075
4076
# dot.alpha: numeric value (from 0 to 1) of dot transparency (full transparent to full opaque, respectively)
# ylim: 2 numeric values for y-axis range. If NULL, range of y in data1
# ylog: logical. Log10 scale for the y-axis? Beware: if TRUE, ylim must not contain null or negative values. In addition, will be automatically set to FALSE if vertical argument is set to FALSE, to prevent a bug in ggplot2 (see https://github.com/tidyverse/ggplot2/issues/881)
# y.break.nb: number of desired values on the y-axis
# y.include.zero: logical. Does ylim range include 0? Beware: if ylog = TRUE, will be automately set to FALSE with a warning message
# y.top.extra.margin: single proportion (between 0 and 1) indicating if extra margins must be added to ylim. If different from 0, add the range of the axis * y.top.extra.margin (e.g., abs(ylim[2] - ylim[1]) * y.top.extra.margin) to the top of y-axis. Beware with ylog = TRUE, the range result must not overlap zero or negative values
# y.bottom.extra.margin: idem as y.top.extra.margin but to the bottom of y-axis
# stat.disp: add the mean number above the corresponding bar. Either NULL (no number shown), "top" (at the top of the figure region) or "above" (above each bar)
# stat.size: numeric value of the stat size (in points). Increase the value to increase text size
# stat.dist: numeric value of the stat distance. Increase the value to increase the distance
# xlab: a character string for x-axis legend. If NULL, character string of categ1
# ylab: a character string y-axis legend. If NULL, character string of the y argument
# vertical: logical. Vertical bars? BEWARE: cannot have horizontal bars with a log axis, i.e., ylog = TRUE & vertical = FALSE (see ylog above)
# title: character string of the graph title
# text.size: numeric value of the text size (in points)
Gael  MILLOT's avatar
Gael MILLOT committed
4077
# text.angle: integer value of the text angle for the x-axis labels. Positive values for counterclockwise rotation: 0 for horizontal, 90 for vertical, 180 for upside down etc. Negative values for clockwise rotation: 0 for horizontal, -90 for vertical, -180 for upside down etc.
Gael  MILLOT's avatar
Gael MILLOT committed
4078
4079
4080
4081
4082
4083
4084
4085
4086
4087
# classic: logical. Use the classic theme (article like)?
# grid: logical. draw horizontal lines in the background to better read the bar values? Not considered if classic = FALSE
# return: logical. Return the graph parameters?
# path.lib: absolute path of the required packages, if not in the default folders
# REQUIRED PACKAGES
# ggplot2
# REQUIRED FUNCTIONS FROM CUTE_LITTLE_R_FUNCTION
# fun_param_check()
# fun_pack_import()
# fun_gg_palette()
Gael  MILLOT's avatar
Gael MILLOT committed
4088
# fun_gg_just()
Gael  MILLOT's avatar
Gael MILLOT committed
4089
4090
# fun_round()
# fun_2D_comp()
Gael  MILLOT's avatar
Gael MILLOT committed
4091
# fun_name_change()
Gael  MILLOT's avatar
Gael MILLOT committed
4092
4093
4094
# RETURN
# a barplot
# a list of the graph info if return argument is TRUE:
Gael  MILLOT's avatar
Gael MILLOT committed
4095
4096
4097
4098
4099
# $stat: the graphic statistics
# $removed.row.nb: which rows have been removed due to NA detection in y and categ columns (NULL if no row removed)
# $removed.rows: removed rows containing NA (NULL if no row removed)
# $data: the graphic info coordinates
# $warnings: the warning messages. Use cat() for proper display. NULL if no warning
Gael  MILLOT's avatar
Gael MILLOT committed
4100
4101
# EXAMPLES
# nice representation (1)
Gael  MILLOT's avatar
Gael MILLOT committed
4102
# obs1 <- data.frame(Time = 1:20, Group1 = rep(c("G", "H"), times = 10), Group2 = rep(c("A", "B"), each = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), categ.class.order = list(NULL, c("B", "A")), categ.legend.name = "LEGEND", categ.color = NULL, bar.width = 0.3, error.disp = "SD.TOP", error.whisker.width = 0.8, dot.color = "same", dot.jitter = 0.5, dot.size = 3.5, dot.border.size = 0.2, dot.alpha = 0.5, ylim = c(10, 25), y.include.zero = TRUE, stat.disp = "above", stat.size = 4, xlab = "GROUP", ylab = "MEAN", title = "GRAPH1", text.size = 20, text.angle = 0, classic = TRUE, grid = TRUE, return = TRUE)
Gael  MILLOT's avatar
Gael MILLOT committed
4103
# nice representation (2)
Gael  MILLOT's avatar
Gael MILLOT committed
4104
# set.seed(1) ; obs1 <- data.frame(Time = c(rnorm(24, 0), rnorm(24, -10), rnorm(24, 10), rnorm(24, 20)), Group1 = rep(c("CAT", "DOG"), times = 48), Group2 = rep(c("A", "B", "C", "D"), each = 24)) ; set.seed(NULL) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), categ.class.order = list(NULL, c("B", "A", "D", "C")), categ.legend.name = "LEGEND", categ.color = NULL, bar.width = 0.8, dot.color = "same", dot.tidy = TRUE, dot.bin.nb = 60, dot.size = 3.5, dot.border.size = 0.2, dot.alpha = 1, ylim= c(-20, 25), stat.disp = "above", stat.size = 4, stat.dist = 1, xlab = "GROUP", ylab = "MEAN", vertical = FALSE, title = "GRAPH1", text.size = 20, text.angle = 45, classic = FALSE, return = TRUE)
Gael  MILLOT's avatar
Gael MILLOT committed
4105
4106
4107
4108
4109
4110
4111
4112
4113
4114
4115
4116
4117
4118
4119
4120
4121
4122
4123
4124
4125
4126
4127
4128
4129
4130
4131
4132
4133
4134
4135
4136
4137
4138
4139
4140
4141
4142
4143
4144
4145
4146
4147
4148
4149
4150
4151
4152
4153
4154
4155
4156
4157
4158
4159
4160
4161
4162
4163
4164
4165
4166
4167
4168
4169
4170
4171
4172
4173
4174
4175
4176
# simple example
# obs1 <- data.frame(Time = 1:20, Group1 = rep(c("G", "H"), times = 10), Group2 = rep(c("A", "B"), each = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = "Group1")
# separate bars: example (1) of modification of bar color using a single value
# obs1 <- data.frame(Time = 1:20, Group1 = rep(c("G", "H"), times = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = "Group1", categ.color = "white")
# separate bars: example (2) of modification of bar color using one value par class of categ2
# obs1 <- data.frame(Time = 1:20, Group1 = rep(c("G", "H"), times = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = "Group1", categ.color = c("coral", "lightblue"))
# separate bars: example (3) of modification of bar color using the bar.color data frame column, with respect of the correspondence between categ2 and bar.color columns
# obs1 <- data.frame(Time = 1:20, Group1 = rep(c("G", "H"), times = 10), bar.color = rep(c("coral", "lightblue"), time = 10)) ; obs1 ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = "Group1", categ.color = obs1$bar.color)
# separate bars: example (1) of modification of dot color, using the same dot color as the corresponding bar
# obs1 <- data.frame(Time = 1:20, Group1 = rep(c("G", "H"), times = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = "Group1", dot.color = "same")
# separate bars: example (2) of modification of dot color, using a single color for all the dots
# obs1 <- data.frame(Time = 1:20, Group1 = rep(c("G", "H"), times = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = "Group1", dot.color = "green")
# separate bars: example (3) of modification of dot color, using one value par class of categ2
# obs1 <- data.frame(Time = 1:20, Group1 = rep(c("G", "H"), times = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = "Group1", dot.color = c("green", "brown"))
# separate bars: example (4) of modification of dot color, using different colors for each dot
# obs1 <- data.frame(Time = 1:10, Group1 = rep(c("G", "H"), times = 5)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = "Group1", dot.color = hsv(h = (1:nrow(obs1)) / nrow(obs1)))
# grouped bars: simple example
# obs1 <- data.frame(Time = 1:20, Group1 = rep(c("G", "H"), times = 10), Group2 = rep(c("A", "B"), each = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"))
# grouped bars: more grouped bars
# obs1 <- data.frame(Time = 1:24, Group1 = rep(c("G", "H"), times = 12), Group2 = rep(c("A", "B", "C", "D"), each = 6)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"))
# grouped bars: example (1) of modification of bar color (1), using a single value
# obs1 <- data.frame(Time = 1:20, Group1 = rep(c("G", "H"), times = 10), Group2 = rep(c("A", "B"), each = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), categ.color = "white")
# grouped bars: example (2) of modification of bar color (2), using one value par class of categ2
# obs1 <- data.frame(Time = 1:20, Group1 = rep(c("G", "H"), times = 10), Group2 = rep(c("A", "B"), each = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), categ.color = c("coral", "lightblue"))
# grouped bars: example (3) of modification of bar color (3), using one value per line of obs1, with respect of the correspondence between categ2 and bar.color columns
# obs1 <- data.frame(Time = 1:20, Group1 = rep(c("G", "H"), times = 10), Group2 = rep(c("A", "B"), each = 10), bar.color = rep(c("coral", "lightblue"), each = 10)) ; obs1 ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), categ.color = obs1$bar.color)
# grouped bars: example (1) of modification of dot color, using the same dot color as the corresponding bar
# obs1 <- data.frame(Time = 1:20, Group1 = rep(c("G", "H"), times = 10), Group2 = rep(c("A", "B"), each = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), dot.color = "same")
# grouped bars: example (2) of modification of dot color, using a single color for all the dots
# obs1 <- data.frame(Time = 1:20, Group1 = rep(c("G", "H"), times = 10), Group2 = rep(c("A", "B"), each = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), dot.color = "green")
# grouped bars: example (3) of modification of dot color, using one value par class of categ2
# obs1 <- data.frame(Time = 1:20, Group1 = rep(c("G", "H"), times = 10), Group2 = rep(c("A", "B"), each = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), dot.color = c("green", "brown"))
# grouped bars: example (4) of modification of dot color, using different colors for each dot
# obs1 <- data.frame(Time = 1:10, Group1 = rep(c("G", "H"), times = 5), Group2 = rep(c("A", "B"), each = 5)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), dot.color = hsv(h = (1:nrow(obs1)) / nrow(obs1)))
# no dots (y.include.zero set to TRUE to see the lowest bar):
# obs1 <- data.frame(Time = 1:20, Group1 = rep(c("G", "H"), times = 10), Group2 = rep(c("A", "B"), each = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), dot.color = NULL, y.include.zero = TRUE)
# bar width: example (1) with bar.width = 0.25 -> three times more space between single bars than the bar width (y.include.zero set to TRUE to see the lowest bar)
# obs1 <- data.frame(Time = 1:1000, Group1 = rep(c("G", "H"), each = 500)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = "Group1", dot.color = NULL, y.include.zero = TRUE, bar.width = 0.25)
# bar width: example (2) with bar.width = 1, no space between single bars
# obs1 <- data.frame(Time = 1:1000, Group1 = rep(c("G", "H"), each = 500)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = "Group1", dot.color = NULL, y.include.zero = TRUE, bar.width = 1)
# bar width: example (3) with bar.width = 0.25 -> three times more space between sets of grouped bars than the set width
# obs1 <- data.frame(Time = 1:1000, Group1 = rep(c("G", "H"), times = 500), Group2 = rep(LETTERS[1:5], each = 200)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), dot.color = NULL, y.include.zero = TRUE, bar.width = 0.25)
# bar width: example (4) with bar.width = 0 -> no space between sets of grouped bars
# obs1 <- data.frame(Time = 1:1000, Group1 = rep(c("G", "H"), times = 500), Group2 = rep(LETTERS[1:5], each = 200)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), dot.color = NULL, y.include.zero = TRUE, bar.width = 1)
# whisker width: example (1) with error.whisker.width = 1 -> whiskers have the width of the corresponding bar
# obs1 <- data.frame(Time = 1:1000, Group1 = rep(c("G", "H"), times = 500), Group2 = rep(LETTERS[1:5], each = 200)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), dot.color = NULL, error.disp = "SD", error.whisker.width = 1)
# whisker width: example (2) error bars with no whiskers
# obs1 <- data.frame(Time = 1:1000, Group1 = rep(c("G", "H"), times = 500), Group2 = rep(LETTERS[1:5], each = 200)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), dot.color = NULL, error.disp = "SD", error.whisker.width = 0)
# dot jitter: example (1) with dot.jitter = 1 -> dispersion around the corresponding bar width
# obs1 <- data.frame(Time = 1:1000, Group1 = rep(c("G", "H"), times = 500), Group2 = rep(LETTERS[1:5], each = 200)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), dot.color = "grey", dot.size = 3, dot.alpha = 1,  dot.jitter = 1)
# dot jitter: example (2) with no dispersion
# obs1 <- data.frame(Time = 1:100, Group1 = rep(c("G", "H"), times = 50), Group2 = rep(LETTERS[1:5], each = 20)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), dot.color = "grey", dot.size = 3, dot.alpha = 1,  dot.jitter = 0)
# dot size, dot border size and dot transparency:
# obs1 <- data.frame(Time = 1:100, Group1 = rep(c("G", "H"), times = 50), Group2 = rep(LETTERS[1:5], each = 20)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), dot.color = "grey", dot.size = 4, dot.border.size = 0, dot.alpha = 0.6)
# tidy dot distribution: example (1)
# obs1 <- data.frame(Time = 1:1000, Group1 = rep(c("G", "H"), times = 500), Group2 = rep(LETTERS[1:5], each = 200)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), dot.color = "same", dot.tidy = TRUE, dot.bin.nb = 100)
# tidy dot distribution: example (2) reducing the dot size with dot.bin.nb
# obs1 <- data.frame(Time = 1:1000, Group1 = rep(c("G", "H"), times = 500), Group2 = rep(LETTERS[1:5], each = 200)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), dot.color = "same", dot.tidy = TRUE, dot.bin.nb = 150)
# tidy dot distribution: comparison with random spreading
# obs1 <- data.frame(Time = 1:1000, Group1 = rep(c("G", "H"), times = 500), Group2 = rep(LETTERS[1:5], each = 200)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), dot.color = "same", dot.tidy = FALSE, dot.jitter = 1, dot.size = 2)
# log scale: beware, y column must be log, otherwise incoherent scale
# obs1 <- data.frame(Time = log10((1:20) * 100), Group1 = rep(c("G", "H"), times = 10), Group2 = rep(c("A", "B"), each = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), ylog = TRUE)
# break number: (make nice)
# obs1 <- data.frame(Time = log10((1:20) * 100), Group1 = rep(c("G", "H"), times = 10), Group2 = rep(c("A", "B"), each = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), y.break.nb = 10)
# extra margins for the plot region: to avoid dot cuts
# obs1 <- data.frame(Time = log10((1:20) * 100), Group1 = rep(c("G", "H"), times = 10), Group2 = rep(c("A", "B"), each = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), y.top.extra.margin = 0.25, y.bottom.extra.margin = 0.25)
# mean diplay: example (1) at the top of the plot region
# obs1 <- data.frame(Time = log10((1:20) * 100), Group1 = rep(c("G", "H"), times = 10), Group2 = rep(c("A", "B"), each = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), stat.disp = "top", stat.size = 4, stat.dist = 2)
# mean diplay: example (2) above bars
# obs1 <- data.frame(Time = log10((1:20) * 100), Group1 = rep(c("G", "H"), times = 10), Group2 = rep(c("A", "B"), each = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), stat.disp = "above", stat.size = 4, stat.dist = 2)
# label orientation: beware, log scale automatically set to FALSE for horizontal display, because of a bug in ggplot2 (https://github.com/tidyverse/ggplot2/issues/881)
# obs1 <- data.frame(Time = log10((1:20) * 100), Group1 = rep(c("G", "H"), times = 10), Group2 = rep(c("A", "B"), each = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), vertical = FALSE)
Gael  MILLOT's avatar
Gael MILLOT committed
4177
4178
# classic representation (use grid = TRUE to display the background lines of the y axis ticks)
# obs1 <- data.frame(Time = log10((1:20) * 100), Group1 = rep(c("G", "H"), times = 10), Group2 = rep(c("A", "B"), each = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), classic = TRUE, grid = FALSE)
Gael  MILLOT's avatar
Gael MILLOT committed
4179
4180
4181
4182
# graphic info: 
# obs1 <- data.frame(Time = log10((1:20) * 100), Group1 = rep(c("G", "H"), times = 10), Group2 = rep(c("A", "B"), each = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "Time", categ = c("Group1", "Group2"), return = TRUE)
# all the arguments:
# obs1 <- data.frame(x = 1:20, Group1 = rep(c("G", "H"), times = 10), Group2 = rep(c("A", "B"), each = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "x", categ = c("Group1", "Group2"), categ.class.order = list(NULL, c("B", "A")), categ.legend.name = "", categ.color = c("red", "blue"), bar.width = 0.25, error.disp = "SD", error.whisker.width = 0.8, dot.color = "grey", dot.tidy = FALSE, dot.bin.nb = 30, dot.jitter = 1, dot.size = 4, dot.border.size = 0, dot.alpha = 1, ylim = NULL, ylog = FALSE, y.break.nb = NULL, y.include.zero = FALSE, y.top.extra.margin = 0.05, y.bottom.extra.margin = 0.05, stat.disp = "above", stat.size = 4, xlab = "GROUP", ylab = "MEAN", vertical = FALSE, title = "GRAPH1", text.size = 14, text.angle = 45, classic = TRUE, grid = TRUE, return = TRUE, path.lib = NULL)
Gael  MILLOT's avatar
Gael MILLOT committed
4183
# DEBUGGING
Gael  MILLOT's avatar
Gael MILLOT committed
4184
4185
4186
# data1 <- data.frame(a = 1:20, group1 = rep(c("G", "H"), times = 10), group2 = rep(c("A", "B"), each = 10), bar.color = rep(c("brown", "orange"), each = 10)) ; data1[2:3, 1] <- NA ; data1[7:8, 2] <- NA ; y = names(data1)[1] ; categ = c(names(data1)[2], names(data1)[3]) ; categ.class.order = list(L1 = NULL, L2 = c("B", "A")) ; categ.legend.name = NULL ; categ.color = na.omit(data1)$bar.color ; bar.width = 0.5 ; error.disp = "SD" ; error.whisker.width = 0.5 ; dot.color = "same" ; dot.tidy = FALSE ; dot.bin.nb = 30 ; dot.jitter = 0.25 ; dot.size = 3 ; dot.border.size = 0.5 ; dot.alpha = 1 ; ylim = NULL ; ylog = FALSE ; y.break.nb = NULL ; y.include.zero = FALSE ; y.top.extra.margin = 0.05 ; y.bottom.extra.margin = 0 ; stat.disp = NULL ; stat.size = 4 ; stat.dist = 2 ; xlab = NULL ; ylab = NULL ; vertical = TRUE ; title = "" ; text.size = 12 ; text.angle = 0 ; classic = FALSE ; grid = FALSE ; return = FALSE ; path.lib = NULL
# data1 <-data.frame(a = rep(1:20, 5), group1 = rep(c("G", "H"), times = 50), group2 = rep(LETTERS[1:5], each = 20)) ; y = names(data1)[1] ; categ = c(names(data1)[2], names(data1)[3]) ; categ.class.order = list(L1 = NULL, L2 = c("B", "A", "E", "D", "C")) ; categ.legend.name = NULL ; categ.color = NULL ; bar.width = 0.5 ; error.disp = "SD" ; error.whisker.width = 0.5 ; dot.color = "same" ; dot.tidy = TRUE ; dot.bin.nb = 30 ; dot.jitter = 0.25 ; dot.size = 3 ; dot.border.size = 0.5 ; dot.alpha = 1 ; ylim = NULL ; ylog = FALSE ; y.break.nb = NULL ; y.include.zero = FALSE ; y.top.extra.margin = 0.05 ; y.bottom.extra.margin = 0 ; stat.disp = NULL ; stat.size = 4 ; stat.dist = 2 ; xlab = NULL ; ylab = NULL ; vertical = TRUE ; title = "" ; text.size = 12 ; text.angle = 0 ; classic = FALSE ; grid = FALSE ; return = FALSE ; path.lib = NULL
# data1 <- data.frame(a = 1:20, group1 = rep(c("G", "H"), times = 10), group2 = rep(c("A", "B"), each = 10), bar.color = rep(c("brown", "orange"), each = 10)) ; data1[2:3, 1] <- NA ; data1[7:8, 2] <- NA ; y = names(data1)[1] ; categ = c(names(data1)[2], names(data1)[3]) ; categ.class.order = list(L1 = NULL, L2 = c("B", "A")) ; categ.legend.name = NULL ; categ.color = na.omit(data1)$bar.color ; bar.width = 0.5 ; error.disp = "SD" ; error.whisker.width = 0.5 ; dot.color = "same" ; dot.tidy = TRUE ; dot.bin.nb = 30 ; dot.jitter = 0.25 ; dot.size = 3 ; dot.border.size = 0.5 ; dot.alpha = 1 ; ylim = NULL ; ylog = FALSE ; y.break.nb = NULL ; y.include.zero = FALSE ; y.top.extra.margin = 0.05 ; y.bottom.extra.margin = 0 ; stat.disp = "above" ; stat.size = 4 ; stat.dist = 2 ; xlab = NULL ; ylab = NULL ; vertical = TRUE ; title = "" ; text.size = 12 ; text.angle = 0 ; classic = FALSE ; grid = FALSE ; return = FALSE ; path.lib = NULL
Gael  MILLOT's avatar
Gael MILLOT committed
4187
# set.seed(1) ; data1 <- data.frame(a = c(rnorm(25, 0), rnorm(25, -10), rnorm(25, 10), rnorm(25, 20)), group1 = rep(c("G", "H"), times = 50), group2 = rep(c("A", "B", "C", "D"), each = 25)) ; set.seed(NULL) ; y = "Time" ; categ = c("group1", "group2") ; categ.class.order = list(NULL, c("B", "A", "D", "C")) ; categ.legend.name = "LEGEND" ; categ.color = NULL ; bar.width = 0.8 ; error.disp = "SD" ; error.whisker.width = 0.5 ; dot.color = "same" ; dot.tidy = TRUE ; dot.bin.nb = 60 ; dot.jitter = 0.25 ; dot.size = 3.5 ; dot.border.size = 0 ; dot.alpha = 1 ; ylim= c(-15, 25) ; ylog = FALSE ; y.break.nb = NULL ; y.include.zero = FALSE ; y.top.extra.margin = 0.05 ; y.bottom.extra.margin = 0 ; stat.disp = "above" ; stat.size = 4 ; stat.dist = 2 ; xlab = "GROUP" ; ylab = "MEAN" ; vertical = FALSE ; title = "GRAPH1" ; text.size = 20 ; text.angle = -200 ; classic = FALSE ; grid = FALSE ; return = FALSE ; path.lib = NULL
Gael  MILLOT's avatar
Gael MILLOT committed
4188
# set.seed(1) ; data1 <- data.frame(x = 1:1000, group1 = rep(c("G", "H"), times = 500), group2 = rep(LETTERS[1:5], each = 200)) ; set.seed(NULL) ; y = "x" ; categ <- c("group1", "group2") ; categ.class.order = list(NULL, c("B", "A", "D", "C", "E")) ; categ.legend.name = "LEGEND" ; categ.color = NULL ; bar.width = 0.8 ; error.disp = "SD" ; error.whisker.width = 1 ; dot.color = NULL ; dot.tidy = FALSE ; dot.bin.nb = 60 ; dot.jitter = 0.25 ; dot.size = 3.5 ; dot.border.size = 0.2 ; dot.alpha = 1 ; ylim= c(-15, 25) ; ylog = FALSE ; y.break.nb = NULL ; y.include.zero = FALSE ; y.top.extra.margin = 0.05 ; y.bottom.extra.margin = 0 ; stat.disp = "above" ; stat.size = 4 ; stat.dist = 1 ; xlab = "GROUP" ; ylab = "MEAN" ; vertical = TRUE ; title = "GRAPH1" ; text.size = 20 ; text.angle = -200 ; classic = FALSE ; grid = FALSE ; return = FALSE ; path.lib = NULL
Gael  MILLOT's avatar
Gael MILLOT committed
4189
4190
4191
4192
4193
4194
4195
4196
4197
4198
4199
4200
4201
4202
4203
4204
4205
4206
4207
4208
4209
4210
4211
4212
# function name
function.name <- paste0(as.list(match.call(expand.dots=FALSE))[[1]], "()")
# end function name
# required function checking
if(length(find("fun_param_check", mode = "function")) == 0){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": REQUIRED fun_param_check() FUNCTION IS MISSING IN THE R ENVIRONMENT\n\n================\n\n")
stop(tempo.cat)
}
if(length(find("fun_pack_import", mode = "function")) == 0){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": REQUIRED fun_pack_import() FUNCTION IS MISSING IN THE R ENVIRONMENT\n\n================\n\n")
stop(tempo.cat)
}
if(length(find("fun_gg_palette", mode = "function")) == 0){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": REQUIRED fun_gg_palette() FUNCTION IS MISSING IN THE R ENVIRONMENT\n\n================\n\n")
stop(tempo.cat)
}
if(length(find("fun_round", mode = "function")) == 0){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": REQUIRED fun_round() FUNCTION IS MISSING IN THE R ENVIRONMENT\n\n================\n\n")
stop(tempo.cat)
}
if(length(find("fun_2D_comp", mode = "function")) == 0){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": REQUIRED fun_2D_comp() FUNCTION IS MISSING IN THE R ENVIRONMENT\n\n================\n\n")
stop(tempo.cat)
}
Gael  MILLOT's avatar
Gael MILLOT committed
4213
4214
4215
4216
if(length(find("fun_2D_comp", mode = "function")) == 0){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": REQUIRED fun_name_change() FUNCTION IS MISSING IN THE R ENVIRONMENT\n\n================\n\n")
stop(tempo.cat)
}
Gael  MILLOT's avatar
Gael MILLOT committed
4217
# end required function checking
Gael  MILLOT's avatar
Gael MILLOT committed
4218
4219
4220
# reserved words to avoid bugs (used in this function)
reserved.words <- c("categ.check", "categ.color", "dot.color", "dot.max", "dot.min", "ERROR.INF", "ERROR.SUP", "group", "group.check", "max.dot.error", "MEAN", "min.dot.error", "SD", "SEM", "tempo.categ1", "tempo.categ2", "text.max.pos", "text.min.pos", "x", "x.y", "y", "y.check", "y_from.dot.max", "ymax")
# end reserved words to avoid bugs (used in this function)
Gael  MILLOT's avatar
Gael MILLOT committed
4221
4222
4223
4224
4225
4226
# argument checking (and modification for proper color management)
warning <- NULL
arg.check <- NULL # for function debbuging
checked.arg.names <- NULL # for function debbuging
ee <- expression(arg.check <- c(arg.check, tempo$problem) , checked.arg.names <- c(checked.arg.names, tempo$param.name))
tempo <- fun_param_check(data = data1, class = "data.frame", na.contain = TRUE, fun.name = function.name) ; eval(ee)
Gael  MILLOT's avatar
Gael MILLOT committed
4227
4228
4229
4230
if(tempo$problem == FALSE & any(duplicated(names(data1)))){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": DUPLICATED COLUMN NAMES OF data1 ARGUMENT NOT ALLOWED:\n", paste(names(data1)[duplicated(names(data1))], collapse = " "), "\n\n================\n\n")
stop(tempo.cat)
}
Gael  MILLOT's avatar
Gael MILLOT committed
4231
4232
4233
4234
4235
4236
4237
4238
4239
4240
4241
4242
4243
4244
tempo <- fun_param_check(data = y, class = "vector", mode = "character", length = 1, fun.name = function.name) ; eval(ee)
if(tempo$problem == FALSE & ! (y %in% names(data1))){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": y ARGUMENT MUST BE A COLUMN NAME OF data1\n\n================\n\n")
stop(tempo.cat)
}else if(tempo$problem == FALSE){
tempo <- fun_param_check(data = data1[, y], data.name = "y COLUMN OF data1", class = "vector", mode = "numeric", na.contain = TRUE, fun.name = function.name) ; eval(ee)
}
tempo <- fun_param_check(data = categ, class = "vector", mode = "character", fun.name = function.name) ; eval(ee)
if(tempo$problem == FALSE & length(categ) > 2){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": categ ARGUMENT CANNOT HAVE MORE THAN 2 COLUMN NAMES OF data1\n\n================\n\n")
stop(tempo.cat)
}else if(tempo$problem == FALSE & ! all(categ %in% names(data1))){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": categ ARGUMENT MUST BE COLUMN NAMES OF data1. HERE IT IS:\n", paste(categ, collapse = " "), "\n\n================\n\n")
stop(tempo.cat)
Gael  MILLOT's avatar
Gael MILLOT committed
4245
4246
4247
}
# reserved word checking
if(any(names(data1) %in% reserved.words)){
Gael  MILLOT's avatar
Gael MILLOT committed
4248
4249
4250
4251
if(any(duplicated(names(data1)))){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": DUPLICATED COLUMN NAMES OF data1 ARGUMENT NOT ALLOWED:\n", paste(names(data1)[duplicated(names(data1))], collapse = " "), "\n\n================\n\n")
stop(tempo.cat)
}
Gael  MILLOT's avatar
Gael MILLOT committed
4252
4253
tempo.output <- fun_name_change(names(data1), reserved.words)
for(i3 in 1:length(tempo.output$ini)){ # a loop to be sure to take the good ones
Gael  MILLOT's avatar
Gael MILLOT committed
4254
names(data1)[names(data1) == tempo.output$ini[i3]] <- tempo.output$post[i3]
Gael  MILLOT's avatar
Gael MILLOT committed
4255
4256
if(any(y == tempo.output$ini[i3])){
y[y == tempo.output$ini[i3]] <- tempo.output$post[i3]
Gael  MILLOT's avatar
Gael MILLOT committed
4257
tempo.warning <- paste0("FROM FUNCTION ", function.name, ": IN y ARGUMENT (COLUMN NAMES OF data1 ARGUMENT),\n", tempo.output$ini[i3], " HAS BEEN REPLACED BY ", tempo.output$post[i3], "\nBECAUSE RISK OF BUG AS SOME NAMES IN y ARGUMENT ARE RESERVED WORD USED BY THE ", function.name, " FUNCTION")
Gael  MILLOT's avatar
Gael MILLOT committed
4258
4259
4260
4261
warning <- paste0(ifelse(is.null(warning), tempo.warning, paste0(warning, "\n\n", tempo.warning)))
}
if(any(categ == tempo.output$ini[i3])){
categ[categ == tempo.output$ini[i3]] <- tempo.output$post[i3]
Gael  MILLOT's avatar
Gael MILLOT committed
4262
tempo.warning <- paste0("FROM FUNCTION ", function.name, ": IN categ ARGUMENT (COLUMN NAMES OF data1 ARGUMENT),\n", tempo.output$ini[i3], " HAS BEEN REPLACED BY ", tempo.output$post[i3], "\nBECAUSE RISK OF BUG AS SOME NAMES IN categ ARGUMENT ARE RESERVED WORD USED BY THE ", function.name, " FUNCTION")
Gael  MILLOT's avatar
Gael MILLOT committed
4263
4264
warning <- paste0(ifelse(is.null(warning), tempo.warning, paste0(warning, "\n\n", tempo.warning)))
}
Gael  MILLOT's avatar
Gael MILLOT committed
4265
}
Gael  MILLOT's avatar
Gael MILLOT committed
4266
tempo.warning <- paste0("FROM FUNCTION ", function.name, ": IN COLUMN NAMES OF data1 ARGUMENT,\n", paste(tempo.output$ini, collapse = " "), "\nNAMES HAVE BEEN REPLACED BY\n", paste(tempo.output$post, collapse = " "), "\nBECAUSE RISK OF BUG AS THESE NAMES ARE RESERVED WORD USED BY THE ", function.name, " FUNCTION")
Gael  MILLOT's avatar
Gael MILLOT committed
4267
4268
warning <- paste0(ifelse(is.null(warning), tempo.warning, paste0(warning, "\n\n", tempo.warning)))
}
Gael  MILLOT's avatar
Gael MILLOT committed
4269
# end reserved word checking
Gael  MILLOT's avatar
Gael MILLOT committed
4270
4271
4272
4273
4274
# na detection and removal (done now to be sure of the correct length of categ)
if(any(is.na(data1[, c(y, categ)]))){
removed.row.nb <- unlist(lapply(lapply(c(data1[c(y, categ)]), FUN = is.na), FUN = which))
removed.rows <- data1[removed.row.nb, ]
data1 <- data1[-removed.row.nb, ]
Gael  MILLOT's avatar
Gael MILLOT committed
4275
tempo.warning <- paste0("FROM FUNCTION ", function.name, ": NA DETECTED IN COLUMN ", paste(c(y, categ), collapse = " "), " OF data1 AND CORRESPONDING ROWS REMOVED (SEE $removed.row.nb AND $removed.rows)")
Gael  MILLOT's avatar
Gael MILLOT committed
4276
4277
4278
4279
4280
4281
4282
4283
warning <- paste0(ifelse(is.null(warning), tempo.warning, paste0(warning, "\n\n", tempo.warning)))
}else{
removed.row.nb <- NULL
removed.rows <- NULL
}
# end na detection and removal (done now to be sure of the correct length of categ)
for(i1 in 1:length(categ)){
if(any(is.na(data1[, categ[i1]]))){
Gael  MILLOT's avatar
Gael MILLOT committed
4284
tempo.warning <- paste0("FROM FUNCTION ", function.name, ": IN categ NUMBER ", i1, " IN data1, THE CATEGORY COLUMN ", categ[i1], " CONTAINS NA")
Gael  MILLOT's avatar
Gael MILLOT committed
4285
4286
4287
4288
4289
4290
4291
4292
warning <- paste0(ifelse(is.null(warning), tempo.warning, paste0(warning, "\n\n", tempo.warning)))
}
tempo1 <- fun_param_check(data = data1[, categ[i1]], data.name = paste0("categ NUMBER ", i1, " OF data1"), class = "vector", mode = "character", na.contain = TRUE, fun.name = function.name, print = FALSE)
tempo2 <- fun_param_check(data = data1[, categ[i1]], data.name = paste0("categ NUMBER ", i1, " OF data1"), class = "factor", na.contain = TRUE, fun.name = function.name, print = FALSE)
if(tempo1$problem == TRUE & tempo2$problem == TRUE){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": ", paste0("categ NUMBER ", i1, " OF data1"), " MUST BE A FACTOR OR CHARACTER VECTOR\n\n================\n\n")
stop(tempo.cat)
}else if(tempo1$problem == FALSE){
Gael  MILLOT's avatar
Gael MILLOT committed
4293
tempo.warning <- paste0("FROM FUNCTION ", function.name, ": IN categ NUMBER ", i1, " IN data1, THE CHARACTER COLUMN HAS BEEN CONVERTED TO FACTOR")
Gael  MILLOT's avatar
Gael MILLOT committed
4294
4295
4296
4297
4298
4299
4300
4301
4302
4303
4304
4305
warning <- paste0(ifelse(is.null(warning), tempo.warning, paste0(warning, "\n\n", tempo.warning)))
}
data1[, categ[i1]] <- factor(data1[, categ[i1]]) # if already a factor, change nothing, if characters, levels according to alphabetical order
}
if( ! is.null(categ.class.order)){
tempo <- fun_param_check(data = categ.class.order, class = "list", fun.name = function.name) ; eval(ee)
if(tempo$problem == FALSE & length(categ.class.order) > 2){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": categ.class.order ARGUMENT MUST BE A LIST OF MAX LENGTH 2\n\n================\n\n")
stop(tempo.cat)
}else if(tempo$problem == FALSE){
for(i3 in 1:length(categ.class.order)){
if(is.null(categ.class.order[[i3]])){
Gael  MILLOT's avatar
Gael MILLOT committed
4306
tempo.warning <- paste0("FROM FUNCTION ", function.name, ": THE categ.class.order COMPARTMENT ", i3, " IS NULL. ALPHABETICAL ORDER WILL BE APPLIED")
Gael  MILLOT's avatar
Gael MILLOT committed
4307
4308
4309
4310
4311
4312
4313
4314
4315
4316
4317
4318
4319
4320
4321
4322
4323
4324
4325
4326
4327
4328
4329
4330
4331
4332
4333
4334
4335
4336
4337
4338
4339
4340
4341
4342
4343
4344
4345
4346
warning <- paste0(ifelse(is.null(warning), tempo.warning, paste0(warning, "\n\n", tempo.warning)))
data1[, categ[i3]] <- factor(as.character(data1[, categ[i3]])) # if already a factor, change nothing, if characters, levels according to alphabetical order
}else if(any(duplicated(categ.class.order[[i3]]))){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": COMPARTMENT ", i3, " OF categ.class.order ARGUMENT CANNOT HAVE DUPLICATED CLASSES: ", paste(categ.class.order[[i3]], collapse = " "), "\n\n================\n\n")
stop(tempo.cat)
}else if( ! (all(categ.class.order[[i3]] %in% unique(data1[, categ[i3]])) & all(unique(data1[, categ[i3]]) %in% categ.class.order[[i3]]))){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": COMPARTMENT ", i3, " OF categ.class.order ARGUMENT MUST BE CLASSES OF ELEMENT ", i3, " OF categ\nHERE IT IS:\nCOMPARTMENT ", i3, " OF categ.class.order:", paste(categ.class.order[[i3]], collapse = " "), "\nCOLUMN ", categ[i3], " OF data1: ", paste( unique(data1[, categ[i3]]), collapse = " "), "\n\n================\n\n")
stop(tempo.cat)
}else{
data1[, categ[i3]] <- factor(data1[, categ[i3]], levels = categ.class.order[[i3]]) # reorder the factor

}
}
}
}
if( ! is.null(categ.legend.name)){
tempo <- fun_param_check(data = categ.legend.name, class = "vector", mode = "character", fun.name = function.name) ; eval(ee)
}else{
categ.legend.name <- categ[length(categ)] # if only categ1, then legend name of categ1, if length(categ) == 2, then legend name of categ2
}
if( ! is.null(categ.color)){
# check the nature of color
tempo1 <- fun_param_check(data = categ.color, class = "vector", mode = "character", na.contain = TRUE, fun.name = function.name, print = FALSE)
tempo2 <- fun_param_check(data = categ.color, class = "factor", na.contain = TRUE, fun.name = function.name, print = FALSE)
if(tempo1$problem == TRUE & tempo2$problem == TRUE){
# integer colors into gg_palette
tempo.check.color <- fun_param_check(data = categ.color, class = "integer", double.as.integer.allowed = TRUE, na.contain = TRUE, fun.name = function.name, print = FALSE)$problem
if(tempo.check.color == TRUE){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": categ.color MUST BE A FACTOR OR CHARACTER VECTOR OR INTEGER VECTOR\n\n================\n\n") # integer possible because dealt above
stop(tempo.cat)
}else{ # convert integers into colors
categ.color <- fun_gg_palette(max(categ.color, na.rm = TRUE))
}
# end integer colors into gg_palette
}
if( ! (all(categ.color %in% colors() | grepl(pattern = "^#", categ.color)))){ # check that all strings of low.color start by #
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": categ.color ARGUMENT MUST BE A HEXADECIMAL COLOR VECTOR STARTING BY # AND/OR COLOR NAMES GIVEN BY colors(): ", paste(unique(categ.color), collapse = " "), "\n\n================\n\n")
stop(tempo.cat)
}
if(any(is.na(categ.color))){
Gael  MILLOT's avatar
Gael MILLOT committed
4347
tempo.warning <- paste0("FROM FUNCTION ", function.name, ": categ.color ARGUMENT CONTAINS NA")
Gael  MILLOT's avatar
Gael MILLOT committed
4348
4349
4350
4351
4352
4353
4354
4355
4356
4357
warning <- paste0(ifelse(is.null(warning), tempo.warning, paste0(warning, "\n\n", tempo.warning)))
}
# end check the nature of color
# check the length of color
# No problem of NA management by ggplot2 because already removed
i0 <- length(categ) # if only categ1, then colors for classes of categ1, if length(categ) == 2, then colors for classes of categ2
if(length(categ.color) == length(unique(data1[, categ[i0]]))){ # here length(categ.color) is equal to the different number of categ
data1[, categ[i0]] <- factor(data1[, categ[i0]]) # if already a factor, change nothing, if characters, levels according to alphabetical order
data1 <- data.frame(data1, categ.color = data1[, categ[i0]])
levels(data1$categ.color) <- categ.color
Gael  MILLOT's avatar
Gael MILLOT committed
4358
tempo.warning <- paste0("FROM FUNCTION ", function.name, ": IN ", categ[i0], " OF categ ARGUMENT, THE FOLLOWING COLORS:\n", paste(categ.color, collapse = " "), "\nHAVE BEEN ATTRIBUTED TO THESE CLASSES:\n", paste(levels(factor(data1[, categ[i0]])), collapse = " "))
Gael  MILLOT's avatar
Gael MILLOT committed
4359
4360
4361
4362
4363
4364
4365
4366
4367
4368
warning <- paste0(ifelse(is.null(warning), tempo.warning, paste0(warning, "\n\n", tempo.warning)))
}else if(length(categ.color) == length(data1[, categ[i0]])){# here length(categ.color) is equal to nrow(data1) -> Modif to have length(categ.color) equal to the different number of categ (length(categ.color) == length(levels(data1[, categ[i0]])))
data1 <- data.frame(data1, categ.color = categ.color)
tempo.check <- unique(data1[ , c(categ[i0], "categ.color")])
if( ! (nrow(tempo.check) == length(unique(categ.color)) & nrow(tempo.check) == length(unique(data1[ , categ[i0]])))){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": categ.color ARGUMENT HAS THE LENGTH OF data1 ROW NUMBER\nBUT IS INCORRECTLY ASSOCIATED TO EACH CLASS OF categ ", categ[i0], ":\n", paste(unique(mapply(FUN = "paste", data1[ ,categ[i0]], data1[ ,"categ.color"])), collapse = "\n"), "\n\n================\n\n")
stop(tempo.cat)
}else{
data1[, categ[i0]] <- factor(data1[, categ[i0]]) # if already a factor, change nothing, if characters, levels according to alphabetical order
categ.color <- unique(categ.color[order(data1[, categ[i0]])]) # Modif to have length(categ.color) equal to the different number of categ (length(categ.color) == length(levels(data1[, categ[i0]])))
Gael  MILLOT's avatar
Gael MILLOT committed
4369
tempo.warning <- paste0("FROM FUNCTION ", function.name, ": categ.color ARGUMENT HAS THE LENGTH OF data1 ROW NUMBER\nCOLORS HAVE BEEN RESPECTIVELY ASSOCIATED TO EACH CLASS OF categ ", categ[i0], " AS:\n", paste(levels(factor(data1[, categ[i0]])), collapse = " "), "\n", paste(categ.color, collapse = " "))
Gael  MILLOT's avatar
Gael MILLOT committed
4370
4371
4372
4373
4374
4375
warning <- paste0(ifelse(is.null(warning), tempo.warning, paste0(warning, "\n\n", tempo.warning)))
}
}else if(length(categ.color) == 1){
data1[, categ[i0]] <- factor(data1[, categ[i0]]) # if already a factor, change nothing, if characters, levels according to alphabetical order
data1 <- data.frame(data1, categ.color = categ.color)
categ.color <- rep(categ.color, length(levels(data1[, categ[i0]])))
Gael  MILLOT's avatar
Gael MILLOT committed
4376
tempo.warning <- paste0("FROM FUNCTION ", function.name, ": categ.color ARGUMENT HAS LENGTH 1, MEANING THAT ALL THE DIFFERENT CLASSES OF ", categ[i0], "\n", paste(levels(factor(data1[, categ[i0]])), collapse = " "), "\nWILL HAVE THE SAME COLOR\n", paste(categ.color, collapse = " "))
Gael  MILLOT's avatar
Gael MILLOT committed
4377
4378
4379
4380
4381
4382
4383
4384
4385
4386
4387
warning <- paste0(ifelse(is.null(warning), tempo.warning, paste0(warning, "\n\n", tempo.warning)))
}else{
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": categ.color ARGUMENT MUST BE (1) LENGTH 1, OR (2) THE LENGTH OF data1 NROWS, OR (3) THE LENGTH OF THE CLASSES IN THE categ ", categ[i0], " COLUMN. HERE IT IS COLOR LENGTH ", length(categ.color), " VERSUS CATEG LENGTH ", length(data1[, categ[i0]]), " AND CATEG CLASS LENGTH ", length(unique(data1[, categ[i0]])), "\nPRESENCE OF NA COULD BE THE PROBLEM\n\n================\n\n")
stop(tempo.cat)
}
}else{
i0 <- length(categ) # if only categ1, then colors for classes of categ1, if length(categ) == 2, then colors for classes of categ2
data1[, categ[i0]] <- factor(data1[, categ[i0]]) # if already a factor, change nothing, if characters, levels according to alphabetical order
categ.color <- fun_gg_palette(length(levels(data1[, categ[i0]])))
data1 <- data.frame(data1, categ.color = data1[, categ[i0]])
levels(data1$categ.color) <- categ.color
Gael  MILLOT's avatar
Gael MILLOT committed
4388
tempo.warning <- paste0("FROM FUNCTION ", function.name, ": NULL categ.color ARGUMENT -> COLORS RESPECTIVELY ATTRIBUTED TO EACH CLASS OF ", categ[i0], " IN data1:\n", paste(categ.color, collapse = " "), "\n", paste(levels(data1[, categ[i0]]), collapse = " "))
Gael  MILLOT's avatar
Gael MILLOT committed
4389
4390
4391
4392
4393
4394
4395
4396
4397
4398
4399
4400
4401
4402
4403
4404
4405
4406
4407
4408
4409
4410
4411
4412
warning <- paste0(ifelse(is.null(warning), tempo.warning, paste0(warning, "\n\n", tempo.warning)))
}
tempo <- fun_param_check(data = bar.width, prop = TRUE, length = 1, fun.name = function.name) ; eval(ee)
if( ! is.null(error.disp)){
tempo <- fun_param_check(data = error.disp, options = c("SD", "SD.TOP", "SEM", "SEM.TOP"), length = 1, fun.name = function.name) ; eval(ee)
}
tempo <- fun_param_check(data = error.whisker.width, prop = TRUE, length = 1, fun.name = function.name) ; eval(ee)
if( ! is.null(dot.color)){
# check the nature of color
tempo1 <- fun_param_check(data = dot.color, class = "vector", mode = "character", na.contain = TRUE, fun.name = function.name, print = FALSE)
tempo2 <- fun_param_check(data = dot.color, class = "factor", na.contain = TRUE, fun.name = function.name, print = FALSE)
if(tempo1$problem == TRUE & tempo2$problem == TRUE){
# integer colors into gg_palette
tempo.check.color <- fun_param_check(data = dot.color, class = "integer", double.as.integer.allowed = TRUE, na.contain = TRUE, fun.name = function.name, print = FALSE)$problem
if(tempo.check.color == TRUE){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": dot.color MUST BE A FACTOR OR CHARACTER VECTOR OR INTEGER VECTOR\n\n================\n\n") # integer possible because dealt above
stop(tempo.cat)
}else{ # convert integers into colors
dot.color <- fun_gg_palette(max(dot.color, na.rm = TRUE))
}
# end integer colors into gg_palette
}
if(all(dot.color == "same") & length(dot.color) == 1){
dot.color <- categ.color # same color of the dots as the corresponding bar color
Gael  MILLOT's avatar
Gael MILLOT committed
4413
tempo.warning <- paste0("FROM FUNCTION ", function.name, ": dot.color ARGUMENT HAS BEEN SET TO \"SAME\"\nTHUS, DOT COLORS HAVE BEEN RESPECTIVELY ASSOCIATED TO EACH CLASS OF categ ", categ[i0], " AS:\n", paste(levels(factor(data1[, categ[i0]])), collapse = " "), "\n", paste(levels(factor(dot.color)), collapse = " "))
Gael  MILLOT's avatar
Gael MILLOT committed
4414
4415
4416
4417
4418
4419
warning <- paste0(ifelse(is.null(warning), tempo.warning, paste0(warning, "\n\n", tempo.warning)))
}else if( ! (all(dot.color %in% colors() | grepl(pattern = "^#", dot.color)))){ # check that all strings of low.color start by #
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": dot.color ARGUMENT MUST BE (1) A HEXADECIMAL COLOR VECTOR STARTING BY #, OR (2) COLOR NAMES GIVEN BY colors(), OR (3) INTEGERS, OR THE STRING\"same\"\nHERE IT IS: ", paste(unique(dot.color), collapse = " "), "\n\n================\n\n")
stop(tempo.cat)
}
if(any(is.na(dot.color))){
Gael  MILLOT's avatar
Gael MILLOT committed
4420
tempo.warning <- paste0("FROM FUNCTION ", function.name, ": dot.color ARGUMENT CONTAINS NA")
Gael  MILLOT's avatar
Gael MILLOT committed
4421
4422
4423
4424
4425
4426
4427
4428
4429
4430
warning <- paste0(ifelse(is.null(warning), tempo.warning, paste0(warning, "\n\n", tempo.warning)))
}
# end check the nature of color
# check the length of color
# No problem of NA management by ggplot2 because already removed
i0 <- length(categ) # if only categ1, then colors for classes of categ1, if length(categ) == 2, then colors for classes of categ2
if(length(dot.color) == length(unique(data1[, categ[i0]]))){ # here length(dot.color) is equal to the different number of categ
data1[, categ[i0]] <- factor(data1[, categ[i0]]) # if already a factor, change nothing, if characters, levels according to alphabetical order
data1 <- data.frame(data1, dot.color = data1[, categ[i0]])
levels(data1$dot.color) <- dot.color
Gael  MILLOT's avatar
Gael MILLOT committed
4431
tempo.warning <- paste0("FROM FUNCTION ", function.name, ": IN ", categ[i0], " OF categ ARGUMENT, THE FOLLOWING COLORS:\n", paste(dot.color, collapse = " "), "\nHAVE BEEN ATTRIBUTED TO THESE CLASSES:\n", paste(levels(factor(data1[, categ[i0]])), collapse = " "))
Gael  MILLOT's avatar
Gael MILLOT committed
4432
4433
4434
4435
4436
4437
4438
warning <- paste0(ifelse(is.null(warning), tempo.warning, paste0(warning, "\n\n", tempo.warning)))
}else if(length(dot.color) == length(data1[, categ[i0]])){# here length(dot.color) is equal to nrow(data1) -> Modif to have length(dot.color) equal to the different number of categ (length(dot.color) == length(levels(data1[, categ[i0]])))
data1 <- data.frame(data1, dot.color = dot.color)
}else if(length(dot.color) == 1 & ! all(dot.color == "same")){
data1[, categ[i0]] <- factor(data1[, categ[i0]]) # if already a factor, change nothing, if characters, levels according to alphabetical order
data1 <- data.frame(data1, dot.color = dot.color)
dot.color <- rep(dot.color, length(levels(data1[, categ[i0]])))
Gael  MILLOT's avatar
Gael MILLOT committed
4439
tempo.warning <- paste0("FROM FUNCTION ", function.name, ": dot.color ARGUMENT HAS LENGTH 1, MEANING THAT ALL THE DIFFERENT CLASSES OF ", categ[i0], "\n", paste(levels(factor(data1[, categ[i0]])), collapse = " "), "\nWILL HAVE THE SAME COLOR\n", paste(dot.color, collapse = " "))
Gael  MILLOT's avatar
Gael MILLOT committed
4440
4441
4442
4443
4444
4445
4446
warning <- paste0(ifelse(is.null(warning), tempo.warning, paste0(warning, "\n\n", tempo.warning)))
}else{
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": dot.color ARGUMENT MUST BE (1) LENGTH 1, OR (2) THE LENGTH OF data1 NROWS, OR (3) THE LENGTH OF THE CLASSES IN THE categ ", categ[i0], " COLUMN. HERE IT IS COLOR LENGTH ", length(dot.color), " VERSUS CATEG LENGTH ", length(data1[, categ[i0]]), " AND CATEG CLASS LENGTH ", length(unique(data1[, categ[i0]])), "\nPRESENCE OF NA COULD BE THE PROBLEM\n\n================\n\n")
stop(tempo.cat)
}
}
tempo <- fun_param_check(data = dot.tidy, class = "vector", mode = "logical", length = 1, fun.name = function.name) ; eval(ee)
Gael  MILLOT's avatar
Gael MILLOT committed
4447
tempo <- fun_param_check(data = dot.bin.nb, class = "vector", typeof = "integer", length = 1, double.as.integer.allowed = TRUE, neg.values = FALSE, fun.name = function.name) ; eval(ee)
Gael  MILLOT's avatar
Gael MILLOT committed
4448
4449
4450
4451
4452
4453
4454
4455
4456
4457
4458
4459
4460
tempo <- fun_param_check(data = dot.jitter, prop = TRUE, length = 1, fun.name = function.name) ; eval(ee)
tempo <- fun_param_check(data = dot.size, class = "vector", mode = "numeric", length = 1, neg.values = FALSE, fun.name = function.name) ; eval(ee)
tempo <- fun_param_check(data = dot.border.size, class = "vector", mode = "numeric", length = 1, neg.values = FALSE, fun.name = function.name) ; eval(ee)
tempo <- fun_param_check(data = dot.alpha, prop = TRUE, length = 1, fun.name = function.name) ; eval(ee)
if( ! is.null(ylim)){
tempo <- fun_param_check(data = ylim, class = "vector", mode = "numeric", length = 2, fun.name = function.name) ; eval(ee)
}
tempo <- fun_param_check(data = ylog, class = "vector", mode = "logical", length = 1, fun.name = function.name) ; eval(ee)
if( ! is.null(y.break.nb)){
tempo <- fun_param_check(data = y.break.nb, class = "vector", typeof = "integer", length = 1, double.as.integer.allowed = TRUE, fun.name = function.name) ; eval(ee)
}
tempo <- fun_param_check(data = y.include.zero, class = "vector", mode = "logical", length = 1, fun.name = function.name) ; eval(ee)
if(tempo$problem == FALSE & ylog == TRUE & y.include.zero == TRUE){
Gael  MILLOT's avatar
Gael MILLOT committed
4461
tempo.warning <- paste0("FROM FUNCTION ", function.name, ": BOTH ylog AND y.include.zero ARGUMENTS SET TO TRUE -> y.include.zero ARGUMENT RESET TO FALSE")
Gael  MILLOT's avatar
Gael MILLOT committed
4462
4463
4464
4465
4466
4467
4468
4469
4470
4471
4472
4473
4474
4475
4476
4477
4478
4479
warning <- paste0(ifelse(is.null(warning), tempo.warning, paste0(warning, "\n\n", tempo.warning)))
}
tempo <- fun_param_check(data = y.top.extra.margin, prop = TRUE, length = 1, fun.name = function.name) ; eval(ee)
tempo <- fun_param_check(data = y.bottom.extra.margin, prop = TRUE, length = 1, fun.name = function.name) ; eval(ee)
if( ! is.null(stat.disp)){
tempo <- fun_param_check(data = stat.disp, options = c("top", "above"), length = 1, fun.name = function.name) ; eval(ee)
}
tempo <- fun_param_check(data = stat.size, class = "vector", mode = "numeric", length = 1, neg.values = FALSE, fun.name = function.name) ; eval(ee)
tempo <- fun_param_check(data = stat.dist, class = "vector", mode = "numeric", length = 1, fun.name = function.name) ; eval(ee)
if( ! is.null(xlab)){
tempo <- fun_param_check(data = xlab, class = "vector", mode = "character", length = 1, fun.name = function.name) ; eval(ee)
}
if( ! is.null(ylab)){
tempo <- fun_param_check(data = ylab, class = "vector", mode = "character", length = 1, fun.name = function.name) ; eval(ee)
}
tempo <- fun_param_check(data = vertical, class = "vector", mode = "logical", length = 1, fun.name = function.name) ; eval(ee)
if(tempo$problem == FALSE & ylog == TRUE & vertical == FALSE){
ylog <- FALSE
Gael  MILLOT's avatar
Gael MILLOT committed
4480
tempo.warning <- paste0("FROM FUNCTION ", function.name, ": BECAUSE OF A BUG IN ggplot2, CANNOT FLIP BARS HORIZONTALLY WITH A YLOG SCALE -> ylog ARGUMENT RESET TO FALSE")
Gael  MILLOT's avatar
Gael MILLOT committed
4481
4482
4483
4484
warning <- paste0(ifelse(is.null(warning), tempo.warning, paste0(warning, "\n\n", tempo.warning)))
}
tempo <- fun_param_check(data = title, class = "vector", mode = "character", length = 1, fun.name = function.name) ; eval(ee)
tempo <- fun_param_check(data = text.size, class = "vector", mode = "numeric", length = 1, neg.values = FALSE, fun.name = function.name) ; eval(ee)
Gael  MILLOT's avatar
Gael MILLOT committed
4485
tempo <- fun_param_check(data = text.angle, class = "vector", typeof = "integer", double.as.integer.allowed = TRUE, length = 1, neg.values = TRUE, fun.name = function.name) ; eval(ee)
Gael  MILLOT's avatar
Gael MILLOT committed
4486
4487
4488
4489
4490
4491
4492
4493
4494
4495
4496
4497
4498
4499
4500
4501
4502
4503
4504
4505
4506
4507
4508
4509
4510
4511
4512
4513
4514
4515
4516
4517
4518
4519
4520
4521
4522
4523
4524
4525
4526
4527
4528
4529
4530
4531
4532
4533
4534
4535
4536
4537
4538
4539
4540
4541
4542
4543
4544
4545
4546
4547
4548
4549
4550
4551
4552
4553
4554
4555
4556
4557
4558
4559
4560
4561
4562
4563
4564
4565
4566
4567
4568
4569
4570
4571
4572
4573
4574
4575
tempo <- fun_param_check(data = return, class = "logical", length = 1, fun.name = function.name) ; eval(ee)
tempo <- fun_param_check(data = classic, class = "logical", length = 1, fun.name = function.name) ; eval(ee)
tempo <- fun_param_check(data = grid, class = "logical", length = 1, fun.name = function.name) ; eval(ee)
if( ! is.null(path.lib)){
tempo <- fun_param_check(data = path.lib, class = "vector", mode = "character", fun.name = function.name) ; eval(ee)
if(tempo$problem == FALSE & ! all(dir.exists(path.lib))){
cat(paste0("\n\n============\n\nERROR IN ", function.name, ": \nDIRECTORY PATH INDICATED IN THE path.lib PARAMETER DOES NOT EXISTS: ", path.lib, "\n\n============\n\n"))
arg.check <- c(arg.check, TRUE)
}
}
if(any(arg.check) == TRUE){
stop() # nothing else because print = TRUE by default in fun_param_check()
}
# source("C:/Users/Gael/Documents/Git_versions_to_use/debugging_tools_for_r_dev-v1.2/r_debugging_tools-v1.2.R") ; eval(parse(text = str_basic_arg_check_dev)) ; eval(parse(text = str_arg_check_with_fun_param_check_dev)) # activate this line and use the function (with no arguments left as NULL) to check arguments status and if they have been checked using fun_param_check()
# end argument checking (and modification for proper color management)
# package checking
fun_pack_import(req.package = c("ggplot2"), path.lib = path.lib)
# end package checking
# main code
if(length(categ) == 1){
# new data frames for bar and error bars
mean.dataframe <- aggregate(x = data1[y], by = {x.env <- list(data1[, categ[1]]) ; names(x.env) <-categ[1] ; x.env}, FUN = mean, na.rm = TRUE)
sd.dataframe <- aggregate(x = data1[y], by = {x.env <- list(data1[, categ[1]]) ; names(x.env) <-categ[1] ; x.env}, FUN = sd, na.rm = TRUE)
nb.dataframe <- aggregate(x = data1[y], by = {x.env <- list(data1[, categ[1]]) ; names(x.env) <- categ[1] ; x.env}, FUN = function(x.env2){length(x.env2[ ! is.na(x.env2)])})
if( ! all(identical(mean.dataframe[, categ[1]], sd.dataframe[, categ[1]]) & identical(mean.dataframe[, categ[1]], nb.dataframe[, categ[1]]))){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": aggregate OUTPUT IS DIFFERENT IN TERM OF CLASS ORDER FOR mean.dataframe, sd.dataframe AND nb.dataframe. CODE HAS TO BE MODIFIED\n\n================\n\n")
stop(tempo.cat)
}else{
sem.dataframe <- sd.dataframe
sem.dataframe[, y] <- sd.dataframe[, y] / (nb.dataframe[, y])^0.5
}
# end new data frames for bar and error bars
# data1 check categ order for dots coordinates recovery
data1 <- data.frame(data1, categ.check = data1[, categ[1]])
data1$categ.check <- as.integer(data1$categ.check) # to check that data1[, categ[1]] and dot.coord$group are similar, during merging
# end data1 check categ order for dots coordinates recovery
# per bar dots coordinates recovery
tempo.gg.name <- "gg.indiv.plot."
tempo.gg.count <- 0
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ggplot(data = data1, mapping = ggplot2::aes_string(x = categ[1], y = y, fill = categ[1]))) # fill because this is what is used with geom_bar
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_point(stroke = dot.border.size, size = dot.size, alpha = dot.alpha, pch = 21))
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_boxplot()) # to easily have the equivalent of the grouped bars
dot.coord <- ggplot2::ggplot_build(eval(parse(text = paste(paste0(tempo.gg.name, 1:tempo.gg.count), collapse = " + "))))$data[[1]]
if( ! is.null(dot.color)){
dot.coord <- data.frame(dot.coord[order(dot.coord$group, dot.coord$y), ], y.check = as.double(data1[order(data1$categ.check, data1[, y]), y]), categ.check = data1[order(data1$categ.check, data1[, y]), "categ.check"], dot.color = data1[order(data1$categ.check, data1[, y]), "dot.color"], tempo.categ1 = data1[order(data1$categ.check, data1[, y]), categ[1]]) # y.check to be sure that the order is the same between the y of data1 and the y of dot.coord
names(dot.coord)[names(dot.coord) == "tempo.categ1"] <- categ[1]
if( ! identical(dot.coord$y, dot.coord$y.check)){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": (dot.coord$y AND dot.coord$y.check) AS WELL AS (dot.coord$group AND dot.coord$categ.check) MUST BE IDENTICAL. CODE HAS TO BE MODIFIED\n\n================\n\n")
stop(tempo.cat)
}
}
# end per bar dots coordinates recovery
}else if(length(categ) == 2){
# new data frames for bar and error bars
mean.dataframe <- aggregate(x = data1[y], by = {x.env <- list(data1[, categ[1]], data1[, categ[2]]) ; names(x.env) <- c(categ[1], categ[2]) ; x.env}, FUN = mean, na.rm = TRUE)
sd.dataframe <- aggregate(x = data1[y], by = {x.env <- list(data1[, categ[1]], data1[, categ[2]]) ; names(x.env) <- c(categ[1], categ[2]) ; x.env}, FUN = sd, na.rm = TRUE)
nb.dataframe <- aggregate(x = data1[y], by = {x.env <- list(data1[, categ[1]], data1[, categ[2]]) ; names(x.env) <- c(categ[1], categ[2]) ; x.env}, FUN = function(x.env2){length(x.env2[ ! is.na(x.env2)])})
tempo.check.mean <- mapply(FUN = "paste", mean.dataframe[, categ[1]], mean.dataframe[, categ[2]], sep = "_")
tempo.check.sd <- mapply(FUN = "paste", sd.dataframe[, categ[1]], sd.dataframe[, categ[2]], sep = "_")
tempo.check.nb <- mapply(FUN = "paste", nb.dataframe[, categ[1]], nb.dataframe[, categ[2]], sep = "_")
if( ! all(identical(tempo.check.mean, tempo.check.sd) & identical(tempo.check.mean, tempo.check.nb))){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": aggregate OUTPUT IS DIFFERENT IN TERM OF CLASS ORDER FOR mean.dataframe, sd.dataframe AND nb.dataframe. CODE HAS TO BE MODIFIED\n\n================\n\n")
stop(tempo.cat)
}else{
sem.dataframe <- sd.dataframe
sem.dataframe[, y] <- sd.dataframe[, y] / (nb.dataframe[, y])^0.5
}
# end new data frames for bar and error bars
# data1 check categ order for dots coordinates recovery
tempo.factor <- paste0(data1[order(data1[, categ[2]], data1[, categ[1]]), categ[2]], "_", data1[order(data1[, categ[2]], data1[, categ[1]]), categ[1]])
data1 <- data.frame(data1[order(data1[, categ[2]], data1[, categ[1]]), ], categ.check = factor(tempo.factor, levels = unique(tempo.factor)))
data1$categ.check <- as.integer(data1$categ.check)
# end data1 check categ order for dots coordinates recovery
# per bar dots coordinates recovery
tempo.gg.name <- "gg.indiv.plot."
tempo.gg.count <- 0
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ggplot(data = data1, mapping = ggplot2::aes_string(x = categ[1], y = y, fill = categ[2]))) # fill because this is what is used with geom_bar
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_point(stroke = dot.border.size, size = dot.size, alpha = dot.alpha, pch = 21))
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_boxplot()) # to easily have the equivalent of the grouped bars
dot.coord <- ggplot2::ggplot_build(eval(parse(text = paste(paste0(tempo.gg.name, 1:tempo.gg.count), collapse = " + "))))$data[[1]]
if( ! is.null(dot.color)){
dot.coord <- data.frame(dot.coord[order(dot.coord$group, dot.coord$y), ], y.check = as.double(data1[order(data1$categ.check, data1[, y]), y]), categ.check = data1[order(data1$categ.check, data1[, y]), "categ.check"], dot.color = data1[order(data1$categ.check, data1[, y]), "dot.color"], tempo.categ1 = data1[order(data1$categ.check, data1[, y]), categ[1]], tempo.categ2 = data1[order(data1$categ.check, data1[, y]), categ[2]]) # y.check to be sure that the order is the same between the y of data1 and the y of dot.coord
names(dot.coord)[names(dot.coord) == "tempo.categ1"] <- categ[1]
names(dot.coord)[names(dot.coord) == "tempo.categ2"] <- categ[2]
if( ! (identical(dot.coord$y, dot.coord$y.check) & identical(dot.coord$group, dot.coord$categ.check))){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": (dot.coord$y AND dot.coord$y.check) AS WELL AS (dot.coord$group AND dot.coord$categ.check) MUST BE IDENTICAL. CODE HAS TO BE MODIFIED\n\n================\n\n")
stop(tempo.cat)
}
}
}else{
Gael  MILLOT's avatar
Gael MILLOT committed
4576
tempo.cat <- (paste0("\n\n============\n\nERROR IN ", function.name, ": CODE INCONSISTENCY 2\n\n============\n\n"))
Gael  MILLOT's avatar
Gael MILLOT committed
4577
4578
4579
4580
4581
4582
4583
4584
4585
4586
4587
4588
4589
stop(tempo.cat)
}
data2 <- mean.dataframe
if( ! is.null(error.disp)){
if(error.disp == "SD"){
data2 <- data.frame(data2, SD = sd.dataframe[, y], ERROR.INF = mean.dataframe[, y] - sd.dataframe[, y], ERROR.SUP = mean.dataframe[, y] + sd.dataframe[, y])
}else if(error.disp == "SD.TOP"){
data2 <- data.frame(data2, SD = sd.dataframe[, y], ERROR.INF = mean.dataframe[, y], ERROR.SUP = mean.dataframe[, y] + sd.dataframe[, y])
}else if(error.disp == "SEM"){
data2 <- data.frame(data2, SEM = sem.dataframe[, y], ERROR.INF = mean.dataframe[, y] - sem.dataframe[, y], ERROR.SUP = mean.dataframe[, y] + sem.dataframe[, y])
}else if(error.disp == "SEM.TOP"){
data2 <- data.frame(data2, SEM = sem.dataframe[, y], ERROR.INF = mean.dataframe[, y], ERROR.SUP = mean.dataframe[, y] + sem.dataframe[, y])
}else{
Gael  MILLOT's avatar
Gael MILLOT committed
4590
tempo.cat <- (paste0("\n\n============\n\nERROR IN ", function.name, ": CODE INCONSISTENCY 3\n\n============\n\n"))
Gael  MILLOT's avatar
Gael MILLOT committed
4591
4592
4593
4594
4595
4596
4597
4598
4599
4600
4601
4602
4603
4604
4605
4606
4607
4608
4609
4610
4611
4612
4613
4614
4615
4616
4617
4618
4619
4620
4621
4622
4623
4624
4625
4626
stop(tempo.cat)
}
}
# stat output
stat <- data2
names(stat)[names(stat) == y] <- "MEAN"
# end stat output
# range depending on means and error bars
if(is.null(ylim)){
if(is.null(dot.color)){ # no dots plotted
if( ! is.null(error.disp)){
ylim <- range(c(data2[, "ERROR.INF"], data2[, "ERROR.SUP"]), na.rm = TRUE)
}else{
ylim <- range(data2[, y], na.rm = TRUE)
}
}else{
ylim <- range(data1[, y], na.rm = TRUE)
}
}
# end range depending on means and error bars
ylim <- sort(ylim)
ylim[1] <- ylim[1] - abs(ylim[2] - ylim[1]) * y.bottom.extra.margin
ylim[2] <- ylim[2] + abs(ylim[2] - ylim[1]) * y.top.extra.margin
if(y.include.zero == TRUE){ # no need to check ylog == TRUE because done before
ylim <- range(c(ylim, 0), na.rm = TRUE)
}
if(ylog == TRUE & any(ylim < 0)){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": FINAL ylim RANGE SPAN NULL OR NEGATIVE VALUES:", paste(ylim, collapse = " "), "\nWHICH IS IMCOMPATIBLE WITH ylog PARAMETER SET TO TRUE\n\n================\n\n")
stop(tempo.cat)
}
# width commputations
if(length(categ) == 2){
bar.width2 <- bar.width / length(unique(data1[, categ[length(categ)]])) # real width of each bar in x-axis unit, among the set of grouped bar. Not relevant if no grouped bars length(categ) == 1
}else if(length(categ) == 1){
bar.width2 <- bar.width
}else{
Gael  MILLOT's avatar
Gael MILLOT committed
4627
tempo.cat <- (paste0("\n\n============\n\nERROR IN ", function.name, ": CODE INCONSISTENCY 4\n\n============\n\n"))
Gael  MILLOT's avatar
Gael MILLOT committed
4628
4629
4630
4631
4632
4633
4634
4635
4636
4637
4638
4639
4640
stop(tempo.cat)
}
error.whisker.width <- bar.width * error.whisker.width # real error bar width
dot.jitter <- bar.width2 * dot.jitter # real dot.jitter
# end width commputations
# barplot
# constant part
tempo.gg.name <- "gg.indiv.plot."
tempo.gg.count <- 0
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ggplot())
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::xlab(if(is.null(xlab)){categ[1]}else{xlab}))
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ylab(if(is.null(ylab)){y}else{ylab}))
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ggtitle(title))
Gael  MILLOT's avatar
Gael MILLOT committed
4641
4642
4643
# text angle management
tempo.just <- fun_gg_just(angle = text.angle, axis = ifelse(vertical == TRUE, "x", "y"))
# end text angle management
Gael  MILLOT's avatar
Gael MILLOT committed
4644
if(classic == TRUE){
Gael  MILLOT's avatar
Gael MILLOT committed
4645
# BEWARE: not possible to add several times theme(). NO message but the last one overwrites the others
Gael  MILLOT's avatar
Gael MILLOT committed
4646
4647
4648
4649
4650
4651
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::theme_classic(base_size = text.size))
if(grid == TRUE){
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), m.gg <- ggplot2::theme(
line = ggplot2::element_line(size = 0.5), 
axis.line.y.left = ggplot2::element_line(colour = "black"), # draw lines for the y axis
axis.line.x.bottom = ggplot2::element_line(colour = "black"), # draw lines for the x axis
Gael  MILLOT's avatar
Gael MILLOT committed
4652
4653
panel.grid.major.x = if(vertical == TRUE){NULL}else{ggplot2::element_line(colour = "grey75")},
panel.grid.major.y = if(vertical == TRUE){ggplot2::element_line(colour = "grey75")}else{NULL},
Gael  MILLOT's avatar
Gael MILLOT committed
4654
4655
axis.text.x = if(vertical == TRUE){ggplot2::element_text(angle = tempo.just$angle, hjust = tempo.just$hjust, vjust = tempo.just$vjust)}else{NULL},
axis.text.y = if(vertical == TRUE){NULL}else{ggplot2::element_text(angle = tempo.just$angle, hjust = tempo.just$hjust, vjust = tempo.just$vjust)}
Gael  MILLOT's avatar
Gael MILLOT committed
4656
4657
4658
4659
4660
))
}else{
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), m.gg <- ggplot2::theme(
line = ggplot2::element_line(size = 0.5), 
axis.line.y.left = ggplot2::element_line(colour = "black"), 
Gael  MILLOT's avatar
Gael MILLOT committed
4661
4662
4663
axis.line.x.bottom = ggplot2::element_line(colour = "black"),
axis.text.x = if(vertical == TRUE){ggplot2::element_text(angle = tempo.just$angle, hjust = tempo.just$hjust, vjust = tempo.just$vjust)}else{NULL},
axis.text.y = if(vertical == TRUE){NULL}else{ggplot2::element_text(angle = tempo.just$angle, hjust = tempo.just$hjust, vjust = tempo.just$vjust)}
Gael  MILLOT's avatar
Gael MILLOT committed
4664
4665
4666
4667
4668
4669
4670
4671
4672
4673
4674
4675
4676
))
}
}else{
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), m.gg <- ggplot2::theme(
text = ggplot2::element_text(size = text.size), 
line = ggplot2::element_line(size = 0.5), 
panel.background = ggplot2::element_rect(fill = "grey95"), 
axis.line.y.left = ggplot2::element_line(colour = "black"), 
axis.line.x.bottom = ggplot2::element_line(colour = "black"), 
panel.grid.major.x = ggplot2::element_line(colour = "grey75"), 
panel.grid.major.y = ggplot2::element_line(colour = "grey75"), 
panel.grid.minor.x = ggplot2::element_blank(), 
panel.grid.minor.y = ggplot2::element_blank(), 
Gael  MILLOT's avatar
Gael MILLOT committed
4677
4678
4679
strip.background = ggplot2::element_rect(fill = "white", colour = "black"),
axis.text.x = if(vertical == TRUE){ggplot2::element_text(angle = tempo.just$angle, hjust = tempo.just$hjust, vjust = tempo.just$vjust)}else{NULL},
axis.text.y = if(vertical == TRUE){NULL}else{ggplot2::element_text(angle = tempo.just$angle, hjust = tempo.just$hjust, vjust = tempo.just$vjust)}
Gael  MILLOT's avatar
Gael MILLOT committed
4680
4681
4682
4683
4684
4685
4686
4687
4688
4689
4690
4691
4692
4693
4694
4695
4696
4697
4698
4699
4700
4701
4702
4703
4704
4705
4706
4707
4708
4709
4710
4711
4712
4713
4714
4715
))
}
# end constant part
# barplot and error bars
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_bar(data = data2, mapping = ggplot2::aes_string(x = categ[1], y = y, fill = categ[length(categ)]), stat = "identity", position = ggplot2::position_dodge(width = NULL), color = "black", width = bar.width)) # stat = "identity" because already counted, position = position_dodge(width = NULL) for grouped bars (width = NULL means no overlap between grouped bars). Please, see explanation in https://stackoverflow.com/questions/34889766/what-is-the-width-argument-in-position-dodge/35102486#35102486
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::scale_discrete_manual(aesthetics = "fill", name = categ.legend.name, values = as.character(categ.color), guide = ggplot2::guide_legend(override.aes = list(fill = categ.color)))) # values are the values of color (which is the border color in geom_bar. Beware: values = categ.color takes the numbers to make the colors if categ.color is a factor
if( ! is.null(error.disp)){
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_errorbar(data = data2, mapping = ggplot2::aes_string(x = categ[1], group = categ[length(categ)], ymin = "ERROR.INF", ymax = "ERROR.SUP"), position = ggplot2::position_dodge(width = bar.width), color = "black", width = error.whisker.width)) # cannot use fill = categ[length(categ)] because not an aesthetic of geom_errorbar, but if only x = categ[1], wrong x coordinates with grouped bars
}
# end barplot and error bars
# coordinates management (for random plotting and for stat display)
# bars
bar.coord <- ggplot2::ggplot_build(eval(parse(text = paste(paste0(tempo.gg.name, 1:tempo.gg.count), collapse = " + "))))$data[[1]] # to have the summary statistics of the plot. Here because can be required for stat.disp when just bar are plotted
# end bars
if( ! is.null(dot.color)){
# random dots
if(dot.tidy == FALSE){
dot.coord.rd1 <- merge(dot.coord, bar.coord[c("fill", "group", "x")], by = intersect("group", "group"), sort = FALSE) # rd for random. Send the coord of the bars into the coord data.frame of the dots (in the column x.y). Beware: by = intersect("group", "group") because group is enough as only one value of x per group number in bar.coord. Thus, no need to consider fill
if(nrow(dot.coord.rd1) != nrow(dot.coord)){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": THE merge() FUNCTION DID NOT RETURN A CORRECT dot.coord.rd1 DATA FRAME. CODE HAS TO BE MODIFIED\n\n================\n\n")
stop(tempo.cat)
}
set.seed(1)
sampled.dot.jitter <- if(nrow(dot.coord.rd1) == 1){runif(n = nrow(dot.coord.rd1), min = - dot.jitter / 2, max = dot.jitter / 2)}else{sample(x = runif(n = nrow(dot.coord.rd1), min = - dot.jitter / 2, max = dot.jitter / 2), size = nrow(dot.coord.rd1), replace = FALSE)}
dot.coord.rd2 <- data.frame(dot.coord.rd1, dot.x = dot.coord.rd1$x.y + sampled.dot.jitter) # set the dot.jitter thanks to runif and dot.jitter range. Then, send the coord of the bars into the coord data.frame of the dots (in the column x.y)
set.seed(NULL)
if(length(categ) == 1){
tempo.data1 <- unique(data.frame(data1[categ[1]], group = as.integer(factor(as.numeric(data1[, categ[1]]))))) # categ[2] first if categ[2] is used to make the categories in ggplot and categ[1] is used to make the x-axis
names(tempo.data1)[names(tempo.data1) == categ[1]] <- paste0(categ[1], ".check")
verif <- paste0(categ[1], ".check")
}else if(length(categ) == 2){
tempo.data1 <- unique(data.frame(data1[c(categ[1], categ[2])], group = as.integer(factor(paste0(as.numeric(data1[, categ[2]]), ".", as.numeric(data1[, categ[1]])))))) # categ[2] first if categ[2] is used to make the categories in ggplot and categ[1] is used to make the x-axis
names(tempo.data1)[names(tempo.data1) == categ[1]] <- paste0(categ[1], ".check")
names(tempo.data1)[names(tempo.data1) == categ[2]] <- paste0(categ[2], ".check")
verif <- c(paste0(categ[1], ".check"), paste0(categ[2], ".check"))
}else{
Gael  MILLOT's avatar
Gael MILLOT committed
4716
tempo.cat <- (paste0("\n\n============\n\nERROR IN ", function.name, ": CODE INCONSISTENCY 5\n\n============\n\n"))
Gael  MILLOT's avatar
Gael MILLOT committed
4717
4718
4719
4720
4721
4722
4723
4724
4725
4726
4727
4728
4729
4730
4731
4732
4733
4734
4735
4736
4737
4738
4739
stop(tempo.cat)
}
dot.coord.rd3 <- merge(dot.coord.rd2, tempo.data1, by = "group", sort = FALSE) # send the factors of data1 into coord
if(nrow(dot.coord.rd3) != nrow(dot.coord) | ( ! fun_2D_comp(dot.coord.rd3[categ], dot.coord.rd3[verif])$identical.content)){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": THE merge() FUNCTION DID NOT RETURN A CORRECT dot.coord.rd3 DATA FRAME. CODE HAS TO BE MODIFIED\n\n================\n\n")
stop(tempo.cat)
}
# end random dots
}
# tidy dots
# coordinates are recover during plotting (see dot.coord.tidy1 below)
# end tidy dots
}
# end coordinates management (for random plotting and for stat display)
# dot display
if( ! is.null(dot.color)){
if(dot.tidy == FALSE){
if(dot.border.size == 0){
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_point(data = dot.coord.rd3, mapping = ggplot2::aes_string(x = "dot.x", y = "y", group = categ[length(categ)]), size = dot.size, color = dot.coord.rd3$dot.color, alpha = dot.alpha, pch = 16)) # group used in aesthetic to do not have it in the legend. Here ggplot2::scale_discrete_manual() cannot be used because of the group easthetic
}else{
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_point(data = dot.coord.rd3, mapping = ggplot2::aes_string(x = "dot.x", y = "y", group = categ[length(categ)]), stroke = dot.border.size, size = dot.size, fill = dot.coord.rd3$dot.color, alpha = dot.alpha, pch = 21)) # group used in aesthetic to do not have it in the legend. Here ggplot2::scale_discrete_manual() cannot be used because of the group easthetic
}
}else if(dot.tidy == TRUE){
Gael  MILLOT's avatar
Gael MILLOT committed
4740
4741
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_dotplot(data = dot.coord, mapping = ggplot2::aes_string(x = categ[1], y = "y", color = categ[length(categ)]), position = ggplot2::position_dodge(width = bar.width), binaxis = "y", stackdir = "center", alpha = dot.alpha, fill = dot.coord[rev(order(dot.coord[, categ[1]], decreasing = TRUE)), "dot.color"], show.legend = FALSE, binwidth = (ylim[2] - ylim[1]) / dot.bin.nb)) # very weird behavior of geom_dotplot, because data1 seems reorderer according to x = categ[1] before plotting. Thus, I have  to use fill = dot.coord[rev(order(dot.coord[, categ[1]], decreasing = TRUE)), "dot.color"] to have the good corresponding colors
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::scale_discrete_manual(aesthetics = "color", name = categ.legend.name, values = if(dot.border.size == 0){as.character(levels(dot.coord[rev(order(dot.coord[, categ[1]], decreasing = TRUE)), "dot.color"]))}else{rep("black", length(categ.color))})) # values = rep("black", length(categ.color)) are the values of color (which is the border color of dots), and this modify the border color on the plot. Beware: values = categ.color takes the numbers to make the colors if categ.color is a factor. BEWARE: , guide = ggplot2::guide_legend(override.aes = list(fill = levels(dot.color))) here
Gael  MILLOT's avatar
Gael MILLOT committed
4742
4743
4744
4745
4746
4747
4748
4749
4750
4751
4752
4753
4754
4755
4756
4757
4758
4759
4760
4761
4762
4763
4764
4765
4766
4767
4768
4769
# coordinates of tidy dots
tempo.coord <- ggplot2::ggplot_build(eval(parse(text = paste(paste0(tempo.gg.name, 1:tempo.gg.count), collapse = " + "))))$data # to have the tidy dot coordinates
if(length(which(sapply(tempo.coord, FUN = nrow) == nrow(data1))) > 1){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": MORE THAN 2 COMPARTMENT WITH NROW EQUAL TO nrow(data1) IN THE tempo.coord LIST (FOR TIDY DOT COORDINATES). CODE HAS TO BE MODIFIED\n\n================\n\n")
stop(tempo.cat)
}else{
dot.coord.tidy1 <- tempo.coord[[which(sapply(tempo.coord, FUN = nrow) == nrow(data1))]]
}
tempo.bar.coord <- merge(bar.coord, unique(dot.coord[, c("group", categ)]), by = intersect("group", "group"), sort = FALSE) # add the categ in bar.coord. Beware: by = intersect("group", "group") because group is enough as only one value of x per group number in bar.coord. Thus, no need to consider fill
if(nrow(tempo.bar.coord) != nrow(bar.coord)){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": THE merge() FUNCTION DID NOT RETURN A CORRECT tempo.bar.coord DATA FRAME. CODE HAS TO BE MODIFIED\n\n================\n\n")
stop(tempo.cat)
}
dot.coord.tidy2 <- merge(dot.coord.tidy1, tempo.bar.coord[c("fill", "group", "x", categ)], by = intersect("group", "group"), sort = FALSE) # send the coord of the bars into the coord data.frame of the dots (in the column x.y). Beware: by = intersect("group", "group") because group is enough as only one value of x per group number in bar.coord. Thus, no need to consider fill
if(nrow(dot.coord.tidy2) != nrow(dot.coord)){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": THE merge() FUNCTION DID NOT RETURN A CORRECT dot.coord.tidy2 DATA FRAME. CODE HAS TO BE MODIFIED\n\n================\n\n")
stop(tempo.cat)
}
if(length(categ) == 1){
tempo.data1 <- unique(data.frame(data1[categ[1]], group = as.integer(factor(as.numeric(data1[, categ[1]]))))) # categ[2] first if categ[2] is used to make the categories in ggplot and categ[1] is used to make the x-axis
names(tempo.data1)[names(tempo.data1) == categ[1]] <- paste0(categ[1], ".check")
verif <- paste0(categ[1], ".check")
}else if(length(categ) == 2){
tempo.data1 <- unique(data.frame(data1[c(categ[1], categ[2])], group = as.integer(factor(paste0(as.numeric(data1[, categ[2]]), ".", as.numeric(data1[, categ[1]])))))) # categ[2] first if categ[2] is used to make the categories in ggplot and categ[1] is used to make the x-axis
names(tempo.data1)[names(tempo.data1) == categ[1]] <- paste0(categ[1], ".check")
names(tempo.data1)[names(tempo.data1) == categ[2]] <- paste0(categ[2], ".check")
verif <- c(paste0(categ[1], ".check"), paste0(categ[2], ".check"))
}else{
Gael  MILLOT's avatar
Gael MILLOT committed
4770
tempo.cat <- (paste0("\n\n============\n\nERROR IN ", function.name, ": CODE INCONSISTENCY 6\n\n============\n\n"))
Gael  MILLOT's avatar
Gael MILLOT committed
4771
4772
4773
4774
4775
4776
4777
4778
4779
4780
4781
4782
4783
4784
4785
4786
4787
4788
4789
4790
4791
4792
4793
4794
4795
4796
4797
4798
4799
4800
4801
4802
4803
4804
4805
4806
4807
4808
4809
4810
4811
4812
stop(tempo.cat)
}
dot.coord.tidy3 <- merge(dot.coord.tidy2, tempo.data1, by = "group", sort = FALSE) # send the factors of data1 into coord
if(nrow(dot.coord.tidy3) != nrow(dot.coord) | ( ! fun_2D_comp(dot.coord.tidy3[categ], dot.coord.tidy3[verif])$identical.content)){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": THE merge() FUNCTION DID NOT RETURN A CORRECT dot.coord.tidy3 DATA FRAME. CODE HAS TO BE MODIFIED\n\n================\n\n")
stop(tempo.cat)
}
# end coordinates of tidy dots
}
}
# end dot display
# stat display
# layer after dots but ok, behind dots on the plot
if( ! is.null(stat.disp)){
if(stat.disp == "top"){
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1),  ggplot2::annotate(geom = "text", x = bar.coord$x, y = ylim[2], label = fun_round(bar.coord$y, 2), size = stat.size, color = "black", hjust = ifelse(vertical == TRUE, 0.5, 1.1), vjust = ifelse(vertical == TRUE, 1.1, 0.5))) # beware: no need of order() for labels because bar.coord$x set the order. For justification, see https://stackoverflow.com/questions/7263849/what-do-hjust-and-vjust-do-when-making-a-plot-using-ggplot
}else if(stat.disp == "above"){
# stat coordinates
if( ! is.null(dot.color)){ # for text just above max dot
if(dot.tidy == FALSE){
tempo.stat.ini <- dot.coord.rd3
}else if(dot.tidy == TRUE){
tempo.stat.ini <- dot.coord.tidy3
}
stat.coord1 <- aggregate(x = tempo.stat.ini["y"], by = {x.env <- if(length(categ) == 1){list(tempo.stat.ini$group, tempo.stat.ini$x.y, tempo.stat.ini[, categ[1]])}else if(length(categ) == 2){list(tempo.stat.ini$group, tempo.stat.ini$x.y, tempo.stat.ini[, categ[1]], tempo.stat.ini[, categ[2]])} ; names(x.env) <- if(length(categ) == 1){c("group", "x.y", categ[1])}else if(length(categ) == 2){c("group", "x.y", categ[1], categ[2])} ; x.env}, FUN = min, na.rm = TRUE)
names(stat.coord1)[names(stat.coord1) == "y"] <- "dot.min"
stat.coord2 <- aggregate(x = tempo.stat.ini["y"], by = {x.env <- if(length(categ) == 1){list(tempo.stat.ini$group, tempo.stat.ini$x.y, tempo.stat.ini[, categ[1]])}else if(length(categ) == 2){list(tempo.stat.ini$group, tempo.stat.ini$x.y, tempo.stat.ini[, categ[1]], tempo.stat.ini[, categ[2]])} ; names(x.env) <- if(length(categ) == 1){c("group", "x.y", categ[1])}else if(length(categ) == 2){c("group", "x.y", categ[1], categ[2])} ; x.env}, FUN = max, na.rm = TRUE)
names(stat.coord2) <- paste0(names(stat.coord2), "_from.dot.max")
names(stat.coord2)[names(stat.coord2) == "y_from.dot.max"] <- "dot.max"
stat.coord3 <- cbind(bar.coord[order(bar.coord$x), ], stat.coord1[order(stat.coord1$x.y), ], stat.coord2[order(stat.coord2$x.y), ]) # should be ok to use bar.coord$x and stat.coord$x.y to assemble the two data frames because x coordinates of the bars. Thus, we cannot have identical values
if( ! all(identical(round(stat.coord3$x, 9), round(stat.coord3$x.y, 9)))){
tempo.cat <- paste0("\n\n================\n\nERROR IN ", function.name, ": FUSION OF bar.coord, stat.coord1 AND stat.coord2 ACCORDING TO bar.coord$x, stat.coord1$x.y AND stat.coord2$x.y IS NOT CORRECT. CODE HAS TO BE MODIFIED\n\n================\n\n")
stop(tempo.cat)
}
dot.text.coord <- stat.coord3[, c("x", "group", "dot.min", "dot.max")]
names(dot.text.coord)[names(dot.text.coord) == "dot.min"] <- "text.min.pos"
names(dot.text.coord)[names(dot.text.coord) == "dot.max"] <- "text.max.pos"
}
if( ! is.null(error.disp)){ # for text just above error bars
if(length(categ) == 1){
tempo.data1 <- unique(data.frame(data1[categ[1]], group = as.integer(factor(as.numeric(data1[, categ[1]]))))) # categ[2] first if categ[2] is used to make the categories in ggplot and categ[1] is used to make the x-axis
if( ! identical(stat[order(stat[, categ[1]]), categ[1]], tempo.data1[order(tempo.data1[, categ[1]]), categ[1]])){
Gael  MILLOT's avatar
Gael MILLOT committed
4813
tempo.cat <- (paste0("\n\n============\n\nERROR IN ", function.name, ": CODE PROBLEM IN TRYING TO ASSEMBLE stat AND tempo.data1\n\n============\n\n"))
Gael  MILLOT's avatar
Gael MILLOT committed
4814
4815
4816
4817
4818
4819
4820
4821
4822
stop(tempo.cat)
}else{
names(tempo.data1)[names(tempo.data1) == categ[1]] <- paste0(categ[1], ".check")
names(tempo.data1)[names(tempo.data1) == "group"] <- "group.check"
stat.coord4 <- cbind(stat[order(stat[, categ[1]]), ], tempo.data1[order(tempo.data1[, paste0(categ[1], ".check")]), ])
}
}else if(length(categ) == 2){
tempo.data1 <- unique(data.frame(data1[c(categ[1], categ[2])], group = as.integer(factor(paste0(as.numeric(data1[, categ[2]]), ".", as.numeric(data1[, categ[1]])))))) # categ[2] first if categ[2] is used to make the categories in ggplot and categ[1] is used to make the x-axis
if( ! fun_2D_comp(stat[order(stat[, categ[1]], stat[, categ[2]]), c(categ[1], categ[2])], tempo.data1[order(tempo.data1[, categ[1]], tempo.data1[, categ[2]]), c(categ[1], categ[2])])$identical.content){
Gael  MILLOT's avatar
Gael MILLOT committed
4823
tempo.cat <- (paste0("\n\n============\n\nERROR IN ", function.name, ": CODE PROBLEM IN TRYING TO ASSEMBLE stat AND tempo.data1\n\n============\n\n"))
Gael  MILLOT's avatar
Gael MILLOT committed
4824
4825
4826
4827
4828
4829
4830
4831
stop(tempo.cat)
}else{
names(tempo.data1)[names(tempo.data1) == categ[1]] <- paste0(categ[1], ".check")
names(tempo.data1)[names(tempo.data1) == categ[2]] <- paste0(categ[2], ".check")
names(tempo.data1)[names(tempo.data1) == "group"] <- "group.check"
stat.coord4 <- cbind(stat[order(stat[, categ[1]], stat[, categ[2]]), ], tempo.data1[order(tempo.data1[, paste0(categ[1], ".check")], tempo.data1[,paste0(categ[2], ".check")]), ])
}
}else{
Gael  MILLOT's avatar
Gael MILLOT committed
4832
tempo.cat <- (paste0("\n\n============\n\nERROR IN ", function.name, ": CODE INCONSISTENCY 7\n\n============\n\n"))
Gael  MILLOT's avatar
Gael MILLOT committed
4833
4834
4835
stop(tempo.cat)
}
if( ! identical(bar.coord$group[order(bar.coord$group)], stat.coord4$group.check[order(stat.coord4$group.check)])){
Gael  MILLOT's avatar
Gael MILLOT committed
4836
tempo.cat <- (paste0("\n\n============\n\nERROR IN ", function.name, ": CODE PROBLEM IN TRYING TO ASSEMBLE bar.coord AND stat.coord4\n\n============\n\n"))
Gael  MILLOT's avatar
Gael MILLOT committed
4837
4838
4839
4840
4841
4842
4843
4844
4845
4846
4847
4848
stop(tempo.cat)
}else{
stat.coord5 <- cbind(bar.coord[order(bar.coord$group), ], stat.coord4[order(stat.coord4$group.check), ])
error.text.coord <- stat.coord5[, c("x", "group", "ERROR.INF", "ERROR.SUP")] # 
names(error.text.coord)[names(error.text.coord) == "ERROR.INF"] <- "text.min.pos"
names(error.text.coord)[names(error.text.coord) == "ERROR.SUP"] <- "text.max.pos"
}
}
if(( ! is.null(dot.color)) & ! is.null(error.disp)){ # for text above max dot or error bar
stat.coord3 <- stat.coord3[order(stat.coord3$x), ]
stat.coord5 <- stat.coord5[order(stat.coord5$x), ]
if( ! identical(stat.coord3$group, stat.coord5$group)){
Gael  MILLOT's avatar
Gael MILLOT committed
4849
tempo.cat <- (paste0("\n\n============\n\nERROR IN ", function.name, ": CODE PROBLEM IN TRYING TO ASSEMBLE stat.coord3 AND stat.coord5\n\n============\n\n"))
Gael  MILLOT's avatar
Gael MILLOT committed
4850
4851
4852
4853
4854
4855
4856
4857
4858
4859
4860
4861
4862
4863
4864
4865
4866
4867
4868
4869
4870
4871
4872
4873
4874
4875
4876
4877
4878
4879
4880
4881
4882
4883
4884
4885
4886
stop(tempo.cat)
}else{
stat.coord6 <- data.frame(stat.coord3, min.dot.error =  mapply(FUN = min, stat.coord3$dot.min, stat.coord5$ERROR.INF, na.rm = TRUE))
stat.coord7 <- data.frame(stat.coord6, max.dot.error =  mapply(FUN = max, stat.coord3$dot.max, stat.coord5$ERROR.SUP, na.rm = TRUE))
both.text.coord <- stat.coord7[, c("x", "group", "min.dot.error", "max.dot.error")] # 
names(both.text.coord)[names(both.text.coord) == "min.dot.error"] <- "text.min.pos"
names(both.text.coord)[names(both.text.coord) == "max.dot.error"] <- "text.max.pos"
}
}
if(( ! is.null(dot.color)) & is.null(error.disp)){
text.coord <- dot.text.coord
}else if(is.null(dot.color) & ! is.null(error.disp)){
text.coord <- error.text.coord
}else if(( ! is.null(dot.color)) & ! is.null(error.disp)){
text.coord <- both.text.coord
}
if( ! (is.null(dot.color) & is.null(error.disp))){
bar.coord <- bar.coord[order(bar.coord$x), ]
text.coord <- text.coord[order(text.coord$x), ] # to be sure to have the two objects in the same order for x. BEWARE: cannot add identical(as.integer(text.coord$group), as.integer(bar.coord$group)) because with error, the correspondence between x and group is not the same
if( ! identical(text.coord$x, bar.coord$x)){
tempo.cat <- (paste0("\n\n============\n\nERROR: text.coord AND bar.coord DO NOT HAVE THE SAME x COLUMN CONTENT\n\n============\n\n"))
stop(tempo.cat)
}
}
# end stat coordinates
# stat display
if(is.null(dot.color) & is.null(error.disp)){ # text just above bars
# performed twice: first for y values >=0, then y values < 0, because only a single value allowed for hjust anf vjust
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::annotate(geom = "text", x = bar.coord$x[bar.coord$y >= 0], y = bar.coord$y[bar.coord$y >= 0], label = fun_round(bar.coord$y, 2)[bar.coord$y >= 0], size = stat.size, color = "black", hjust = ifelse(vertical == TRUE, 0.5, 0.5 - stat.dist), vjust = ifelse(vertical == TRUE, 0.5 - stat.dist, 0.5))) # beware: no need of order() for labels because bar.coord$x set the order
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::annotate(geom = "text", x = bar.coord$x[bar.coord$y < 0], y = bar.coord$y[bar.coord$y < 0], label = fun_round(bar.coord$y, 2)[bar.coord$y < 0], size = stat.size, color = "black", hjust = ifelse(vertical == TRUE, 0.5, 0.5 + stat.dist), vjust = ifelse(vertical == TRUE, 0.5 + stat.dist, 0.5))) # beware: no need of order() for labels because bar.coord$x set the order
}else{ # text just above error bars or dots
# I checked that text.coord and bar.coord have the same x and group column content. Thus, ok to use them together
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::annotate(geom = "text", x = text.coord$x[bar.coord$y >= 0], y = text.coord$text.max.pos[bar.coord$y >= 0], label = fun_round(bar.coord$y, 2)[bar.coord$y >= 0], size = stat.size, color = "black", hjust = ifelse(vertical == TRUE, 0.5, 0.5 - stat.dist), vjust = ifelse(vertical == TRUE, 0.5 - stat.dist, 0.5))) # beware: no need of order() for labels because bar.coord$x set the order
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::annotate(geom = "text", x = text.coord$x[bar.coord$y < 0], y = text.coord$text.min.pos[bar.coord$y < 0], label = fun_round(bar.coord$y, 2)[bar.coord$y < 0], size = stat.size, color = "black", hjust = ifelse(vertical == TRUE, 0.5, 0.5 + stat.dist), vjust = ifelse(vertical == TRUE, 0.5 + stat.dist, 0.5))) # beware: no need of order() for labels because bar.coord$x set the order
}
# end stat display
}else{
Gael  MILLOT's avatar
Gael MILLOT committed
4887
tempo.cat <- (paste0("\n\n============\n\nERROR IN ", function.name, ": CODE INCONSISTENCY 8\n\n============\n\n"))
Gael  MILLOT's avatar
Gael MILLOT committed
4888
4889
4890
4891
4892
4893
4894
4895
4896
4897
4898
4899
4900
4901
4902
4903
4904
4905
4906
4907
4908
4909
4910
4911
4912
4913
4914
4915
4916
4917
4918
4919
4920
4921
4922
4923
4924
4925
4926
4927
4928
4929
stop(tempo.cat)
}
}
# end stat display
# y scale management (cannot be before dot plot management)
if(ylog == TRUE){
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::annotation_logticks(sides = "l")) # string containing any of "trbl", for top, right, bottom, and left
if( ! is.null(y.break.nb)){
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::scale_y_continuous(breaks = fun_round(seq(ylim[1], ylim[2], length.out = y.break.nb), dec.nb = 2, after.lead.zero = TRUE)))
}
}else{
if( ! is.null(y.break.nb)){
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::scale_y_continuous(
breaks = fun_round(seq(ylim[1], ylim[2], length.out = y.break.nb), dec.nb = 2, after.lead.zero = TRUE), 
expand = c(0, 0),
limits = NA
))
}else{
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::scale_y_continuous(
expand = c(0, 0),
limits = NA
))
}
}
if(vertical == TRUE){
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::coord_cartesian(ylim = ylim))
}else{
assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::coord_flip(ylim = ylim))
}
# end y scale  management (cannot be before dot plot management)
suppressWarnings(print(eval(parse(text = paste(paste0(tempo.gg.name, 1:tempo.gg.count), collapse = " + ")))))
# end barplot
if(return == TRUE){
output <- ggplot2::ggplot_build(eval(parse(text = paste(paste0(tempo.gg.name, 1:tempo.gg.count), collapse = " + "))))
output <- list(stat = stat, removed.row.nb = removed.row.nb, removed.rows = removed.rows, data = output$data, warnings = paste0("\n", warning, "\n\n"))
return(output)
}
}


######## fun_gg_boxplot() #### ggplot2 boxplot + background dots if required

Gael  MILLOT's avatar
Gael MILLOT committed
4930

Gael  MILLOT's avatar
Gael MILLOT committed
4931
# http://www.sthda.com/english/wiki/ggplot2-box-plot-quick-start-guide-r-software-and-data-visualization
Gael  MILLOT's avatar
Gael MILLOT committed
4932
# Exmaple of boxplots: https://github.com/IndrajeetPatil/ggstatsplot
Gael  MILLOT's avatar
Gael MILLOT committed
4933
4934
4935
4936
4937
4938
4939
4940
4941
4942
4943
4944
4945
4946
4947
4948
4949
4950
4951
4952
4953
4954
4955
4956
4957
4958
4959
4960
4961
4962
4963
4964
4965
4966
4967
4968
4969
4970
4971
4972
4973
4974
4975
4976
4977
4978
4979
4980
4981
4982
4983

fun_gg_boxplot <- function(data1, y, categ, class.order = NULL, legend.name = NULL, categ.color = NULL, dot.color = "same", box.width = 0.5, whisker.width = 0.5, jitter = 0.25, ylim = NULL, ylog = FALSE, y.include.zero = FALSE, top.extra.margin = 0.05, bottom.extra.margin = 0, xlab = NULL, ylab = NULL, pt.size = 3, pt.border.size = 0.5, alpha = 0.5, show.stat = NULL, stat.size = 4, title = "", text.size = 12, break.nb = NULL, classic = FALSE, grid = FALSE, return = FALSE, path.lib = NULL){
# AIM
# ggplot2 vertical barplot representing mean values with the possibility to add error bars and to overlay dots
# for ggplot2 specifications, see: https://ggplot2.tidyverse.org/articles/ggplot2-specs.html
# WARNINGS
# rows containing NA in data1[, c(y, categ)] will be removed before processing, with a warning (see below)
# to have a single boxplot, create a factor column with a single class and specify the name of this column in categ argument as unique element (no categ2 in categ argument). For a single set of grouped boxplots, create a factor column with a single class and specify this column in categ argument as first element (categ1). See categ below
# with several single boxplots (categ argument with only one element), bar.width argument (i.e., width argument of ggplot2::geom_bar()) defines each bar width. The bar.width argument also defines the space between bars by using (1 - bar.width). In addition, xmin and xmax of the fun_gg_bar_mean() output report the bar boundaries (around x-axis unit 1, 2, 3, etc., for each bar)
# with several sets of grouped bars (categ argument with two elements), bar.width argument defines each set of grouped bar width. The bar.width argument also defines the space between set of grouped bars by using (1 - bar.width). In addition, xmin and xmax of the fun_gg_bar_mean() output report the bar boundaries (around x-axis unit 1, 2, 3, etc., for each set of grouped bar)
# ARGUMENTS
# data1: a dataframe containing one column of values (see y argument below) and one or two columns of categories (see categ argument below)
# y: character string of the data1 column name for y-axis (containing numeric values). Numeric values will be used to generate the boxplots and will also be used to plot the dots
# categ: vector of character strings of the data1 column name for categories (column of characters or factor). Must either be one or two column names. If a single column name (further refered to as categ1), then one boxplot per class of categ1. If two column names (further refered to as categ1 and categ2), then one boxplot per class of categ2, which form a group of boxplots in each class of categ1. Beware, categ1 (and categ2 if it exists) must have a single value of y per class of categ1 (and categ2). To have a single boxplot, create a factor column with a single class and specify the name of this column in categ argument as unique element (no categ2 in categ argument). For a single set of grouped boxplots, create a factor column with a single class and specify this column in categ argument as first element (categ1)
# class.order: list indicating the order of the classes of categ1 and categ2 represented on the boxplot (the first compartment for categ1 and and the second for categ2). If class.order = NULL, classes are represented according to the alphabetical order. Some compartment can be NULL and other not
# legend.name: character string of the legend title for categ2. If legend.name = NULL, then legend.name <- categ1 if only categ1 is present and legend.name <- categ2 if categ1 and categ2 are present. Write "" if no legend required
# categ.color: vector of character color string for boxplot color. If categ.color = NULL, default colors of ggplot2, whatever categ1 and categ2. If categ.color is non null and only categ1 in categ argument, categ.color can be either: (1) a single color string (all the boxplots will have this color, whatever the classes of categ1), (2) a vector of string colors, one for each class of categ1 (each color will be associated according to class.order of categ1), (3) a vector or factor of string colors, like if it was one of the column of data1 data frame (beware: a single color per class of categ1 and a single class of categ1 per color must be respected). Integers are also accepted instead of character strings, as long as above rules about length are respected. Integers will be processed by fun_gg_palette() using the max integer value among all the integers in categ.color. If categ.color is non null and categ1 and categ2 specified, all the rules described above will apply to categ2 instead of categ1 (colors will be determined for boxplots inside a group of boxplots)
# dot.color: vector of character string. Idem as categ.color but for dots, except that in the possibility (3), the rule "a single color per class of categ1 and a single class of categ1", cannot be respected (each dot can have a different color). If NULL, no dots plotted
# box.width: numeric value (from 0 to 1) of the bar or set of grouped bar width (see warnings above)
# whisker.width: numeric value (from 0 to 1) of the whisker (error bar extremities) width, with 0 meaning no whiskers and 1 meaning a width equal to the corresponding bar width
# jitter: numeric value (from 0 to 1) of random dot horizontal dispersion, with 0 meaning no dispersion and 1 meaning a dispersion in the corresponding bar width interval
# ylim: 2 numeric values for y-axis range. If NULL, range of y in data1
# ylog: logical. Log10 scale for the y-axis? Beware: if TRUE, ylim must not contain null or negative values
# y.include.zero: logical. Does ylim range include 0? Beware: if ylog = TRUE, will be automately set to FALSE with a warning message
# top.extra.margin: single proportion (between 0 and 1) indicating if extra margins must be added to ylim. If different from 0, add the range of the axis * top.extra.margin (e.g., abs(ylim[2] - ylim[1]) * top.extra.margin) to the top of y-axis. Beware with ylog = TRUE, the range result must not overlap zero or negative values
# bottom.extra.margin: idem as top.extra.margin but to the bottom of y-axis
# xlab: a character string for x-axis legend. If NULL, character string of categ1
# ylab: a character string y-axis legend. If NULL, character string of the y argument
# pt.size: numeric value of dot size
# pt.border.size: numeric value of border dot size. Write zero for no stroke
# alpha: numeric value (from 0 to 1) of dot transparency (full transparent to full opaque, respectively)
# show.stat: add the mean number above the corresponding bar. Either NULL (no number shown), "top" (at the top of the  figure region) or "above" (above each bar)
# stat.size: numeric value of the number size (in points)
# title: character string of the graph title
# text.size: numeric value of the text size (in points)
# break.nb: number of desired values on the y-axis
# classic: logical. Use the classic theme (article like)?
# grid: logical. draw horizontal lines in the background to better read the boxplot values? Not considered if classic = FALSE
# return: logical. Return the graph parameters?
# path.lib: absolute path of the required packages, if not in the default folders
# REQUIRED PACKAGES
# ggplot2
# REQUIRED FUNCTIONS FROM CUTE_LITTLE_R_FUNCTION
# fun_param_check()
# fun_pack_import()
# fun_gg_palette()
# fun_round()
# fun_2D_comp()
# RETURN
# a boxplot
# a list of the graph info if return argument is TRUE:
Gael  MILLOT's avatar
Gael MILLOT committed
4984
4985
4986
4987
4988
# $stat: the graphic statistics
# $removed.row.nb: which rows have been removed due to NA detection in y and categ columns (NULL if no row removed)
# $removed.rows: removed rows containing NA (NULL if no row removed)
# $data: the graphic info coordinates
# $warnings: the warning messages. Use cat() for proper display. NULL if no warning
Gael  MILLOT's avatar
Gael MILLOT committed
4989
4990
4991
4992
4993
4994
4995
4996
4997
4998
4999
5000
# EXAMPLES
# obs1 <- data.frame(a = 1:20, group1 = rep(c("G", "H"), times = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "a", categ = "group1", categ.color = "white") # separate bars, modification of bar color 1 (a single value)
# obs1 <- data.frame(a = 1:20, group1 = rep(c("G", "H"), times = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "a", categ = "group1", categ.color = c("red", "blue")) # separate bars, modification of bar color 2 (one value par class of categ2)
# obs1 <- data.frame(a = 1:20, group1 = rep(c("G", "H"), times = 10), bar.color = rep(c("brown", "orange"), time = 10)) ; obs1 ; fun_gg_bar_mean(data1 = obs1, y = "a", categ = "group1", categ.color = obs1$bar.color) # separate bars, modification of bar color 3 (one value per line of obs1, with respect of the correspondence between categ2 and bar.color columns)
# obs1 <- data.frame(a = 1:20, group1 = rep(c("G", "H"), times = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "a", categ = "group1", dot.color = "same") # separate bars, modification of dot color 1 (same dot color as the corresponding bar)
# obs1 <- data.frame(a = 1:20, group1 = rep(c("G", "H"), times = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "a", categ = "group1", dot.color = "green") # separate bars, modification of dot color 2 (single color for all the dots)
# obs1 <- data.frame(a = 1:20, group1 = rep(c("G", "H"), times = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "a", categ = "group1", dot.color = c("green", "brown")) # separate bars, modification of dot color 3 (one value par class of categ2)
# obs1 <- data.frame(a = 1:10, group1 = rep(c("G", "H"), times = 5)) ; fun_gg_bar_mean(data1 = obs1, y = "a", categ = "group1", dot.color = hsv(h = (1:nrow(obs1)) / nrow(obs1))) # separate bars, modification of dot color 4 (any color for each dot)
# obs1 <- data.frame(a = 1:20, group1 = rep(c("G", "H"), times = 10), group2 = rep(c("A", "B"), each = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "a", categ = c("group1", "group2")) # grouped bars, default arguments
# obs1 <- data.frame(a = 1:24, group1 = rep(c("G", "H"), times = 12), group2 = rep(c("A", "B", "C", "D"), each = 6)) ; fun_gg_bar_mean(data1 = obs1, y = "a", categ = c("group1", "group2"), return = TRUE) # more grouped bars
# obs1 <- data.frame(a = log10((1:20) * 100), group1 = rep(c("G", "H"), times = 10), group2 = rep(c("A", "B"), each = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "a", categ = c("group1", "group2"), ylog = TRUE) # grouped bars, log scale. Beware, y column must be log, otherwise incoherent scale
# obs1 <- data.frame(a = 1:20, group1 = rep(c("G", "H"), times = 10), group2 = rep(c("A", "B"), each = 10)) ; fun_gg_bar_mean(data1 = obs1, y = "a", categ = c("group1", "group2"), dot.color = NULL) # grouped bars, no dots