Merge pull request #33 from ankemt/improve-design-16

Add functionality and improvements around design text file

DESCRIPTION

@@ -12,14 +12,14 @@ Encoding: UTF-8
 LazyData: true
 RoxygenNote: 7.2.1
 Imports:
-    brio,
-    dplyr,
-    magrittr,
-    robustbase (>= 0.93-6),
-    stats,
-    stringr,
-    tidyr,
-    tidyselect
+ assertthat,
+ dplyr,
+ magrittr,
+ robustbase (>= 0.93-6),
+ stats,
+ stringr,
+ tidyr,
+ tidyselect
 Suggests: 
     testthat (>= 3.1.5)
 Config/testthat/edition: 3

NAMESPACE

@@ -1,5 +1,7 @@
 # Generated by roxygen2: do not edit by hand
 
+export(categorize_wells)
+export(exp_design)
 export(find_outliers)
 export(parse_MEA_file)
 export(parse_designfile)

R/exp_design.R

@@ -0,0 +1,117 @@
+#' Generate experimental design file
+#'
+#' With information provided by the user, a file is created
+#' that formalizes the experimental design so it can be used to
+#' analyse the MEA file.
+#'
+#' @param date date (string, preferably formatted as YYYYMMDD), default is prompted
+#' @param expID identifier for the experiment (string), default is prompted
+#' @param path path to the design file (dir/name.txt)
+#'
+#' @export
+exp_design <- function(path="./design.txt",
+                       date = readline(prompt="What is the date of the experiment? (YYYYMMDD) "),
+                       expID = readline(prompt="What is the experiment ID? ")){
+  # TODO check if file can be created
+
+  nwells <- 48 # TODO make this a question also
+
+  write_meta(date, expID, nwells, path)
+
+  # TODO make this a while loop where wells are scored instead of a Q/A
+  continue <- TRUE
+  while(continue){
+    message("You can now add experimental conditions one by one:")
+    write_category(nwells=nwells, path=path)
+    check_cont <- readline(prompt = "Do you want to add another group? (y/n) ")
+    continue <- stringr::str_detect(check_cont, "y|Y")
+  }
+}
+
+
+write_meta <- function(date, expID, nwells, path){
+  directory <- dirname(path)
+  if(!dir.exists(directory)){
+    dir.create(directory, recursive=TRUE)
+  }
+  meta <- paste0("Date: ", date, "\n",
+                 "ExperimentID: ", expID, "\n",
+                 "Total_wells: ", nwells, "\n",
+                 "Groups:")
+  write(meta, file=path)
+
+  message(paste("The design metadata has been written to file at",path))
+}
+
+write_category <- function(
+    group = readline(prompt="What is the experimental condition? "),
+    start = readline(prompt="What is the first well in this category? (e.g.: A1) "),
+    end = readline(prompt="What is the last well in this category? (e.g.: F8) "),
+    dirx = readline(
+      prompt="In what direction ('LR' for left-to-right or 'TB' for top-to-bottom) is the sequence of wells? "),
+    nwells, path){
+  groupname <- paste0(group, ":") # hacky way to make sure the group question is asked first
+  wells <- categorize_wells(start, end, nwells, direction=dirx)
+  write(paste(groupname, wells), file=path, append=TRUE)
+}
+
+
+#' Create well ranges for a design file
+#'
+#' @param start First well ID (e.g. "A1")
+#' @param end Last well ID (e.g "B7")
+#' @param nwells Total number of wells on the plate (currently only 48)
+#' @param direction TB for top-to-bottom or LR for left-to-right
+#'
+#' @return string with a range of well IDs
+#' @export
+#'
+#' @examples
+#' # Wells from left to right, starting at A1 and ending at B4
+#' categorize_wells("A1", "B4")
+#' # Wells from top to bottom, starting at A1 and ending at B4
+#' categorize_wells("A1", "B4", direction="TB")
+categorize_wells <- function(start, end, nwells = 48, direction = "LR"){
+  assertthat::assert_that(direction %in% c("LR", "TB"),
+                          msg = "The direction must be LR (left-to-right), or TB (top-to-bottom).")
+
+  assertthat::assert_that(typeof(nwells) == "double",
+                          msg = "`nwells` should be a number.")
+  #TODO would be nicer if it can convert a string to a number as well
+
+  assertthat::assert_that(nwells %in% c(48), #TODO: add other options like c(6, 12, 24, 48, 96),
+                          msg = "The total number of wells on the plate (`nwells`) is incorrect.")
+
+  rowtotal <- 6
+  coltotal <- 8
+  allcols <- 1:coltotal
+  allrows <- LETTERS[1:rowtotal]
+
+  startrow = stringr::str_extract(start, "[:alpha:]")
+  startcol = as.numeric(stringr::str_extract(start, "[:digit:]"))
+  endrow = stringr::str_extract(end, "[:alpha:]")
+  endcol = as.numeric(stringr::str_extract(end, "[:digit:]"))
+
+  rows <- LETTERS[letter_as_number(startrow):letter_as_number(endrow)]
+  cols <- startcol:endcol
+
+  wells <- NULL
+
+  if(direction == "LR"){
+    for(row in rows) wells <- c(wells, paste0(row, allcols))
+    removecols <- length(wells) - (max(allcols) - endcol)
+    wells <- wells[startcol:removecols]
+  } else if(direction == "TB"){
+    for(col in cols) wells <- c(wells, paste0(allrows, col))
+    maxrow <- max(letter_as_number(allrows))
+    removerows <- length(wells) - (maxrow - letter_as_number(endrow))
+    wells <- wells[letter_as_number(startrow):removerows]
+  }
+
+  wells <- paste(wells, collapse = " ")
+  return(wells)
+}
+
+letter_as_number <- function(letter) as.numeric(stats::setNames(1:26, LETTERS)[letter])
+
+

R/parse_designfile.R

@@ -9,31 +9,19 @@
 #'
 #' @export
 parse_designfile <- function(path){
-  info <- brio::readLines(path)
+  data <- utils::read.csv(path, header=F, sep=":")
 
-  metadata_names <- c("Date", "ExperimentID", "Total_wells")
+  # remove hashed lines, and trim spaces before and after the text
+  data <- data[grep("^#", data$V1, invert=T),]
+  data$V1<- stringr::str_trim(data$V1, side = "both")
+  data$V2 <- stringr::str_trim(data$V2, side = "both")
 
-  # make objects to fill
-  designlist <- NULL
-  metadata <- NULL
-  design <- NULL
+  # the groups start after Groups
+  groups_start <- grep("Groups", data$V1)
+  design <- data[(groups_start+1):nrow(data),]
 
-  for(i in info){
-    if(stringr::str_starts(i, "#")){next} # remove any lines that start with #
-
-    i <- stringr::str_split(i, " ", simplify=TRUE)
-
-    # get date, experimentID, total_wells
-    if(i[1] %in% metadata_names){
-      metadata <- rbind(metadata,i)
-    } else{
-      if(length(i) > 2){
-        design <- rbind(design, i)
-      }
-    }
-    # get groups
-
-  }
+  # metadata is everything before Groups
+  metadata <- data[(1:groups_start-1),]
 
   metadata <- rotate_to_df(metadata)
   design <- make_design_df(design)
@@ -56,20 +44,25 @@ rotate_to_df <- function(df){
 
 
 make_design_df <- function(df){
-  # ensure the first two columns are empty, before removing them
-  if(paste(df[,1],collapse="") != "" |
-     paste(df[,2],collapse="") != ""){
-    stop("Experimental design is unclear. Ensure that every group is preceded by two spaces.")
+  cat_wells <- stringr::str_split(df$V2, '[:blank:]*[:punct:]+[:blank:]*|[:blank:]+')
+  #names(cat_wells) <- df$V1
+
+  allwells <- NULL
+  allcats <- NULL
+
+  for(n in 1:length(cat_wells)){
+    wells <- cat_wells[[n]]
+    allwells <- c(allwells, wells)
+    cats <- rep(df$V1[n], length(wells))
+    allcats <- c(allcats, cats)
   }
-  df <- df[,-c(1,2)]
 
-  df <- rotate_to_df(df)
+  df <- data.frame(Well = allwells, Group = allcats)
+
+  if(anyDuplicated(df$Well)>0){
+    stop("Duplicate well names are not allowed.")
+  }
 
-  # pivot to two columns and sort
-  df <- tidyr::pivot_longer(df,
-                            cols = tidyr::everything(),
-                            names_to = "Group",
-                            values_to = "Well")
   #TODO: make_design_df: no visible binding for global variable ‘Well’
   #TODO: make_design_df: no visible binding for global variable ‘Group’
   df <- dplyr::select(df, Well, Group)

R/treatment_ratio.R

@@ -70,6 +70,7 @@ return(df)
 
 
 create_path <- function(path){
+  # can the following code not be replaced by `dirname`?
   path_vector <- stringr::str_split(path, "[\\\\/]")[[1]]
   lv <- length(path_vector)
   if(path_vector[lv] == ""){

inst/extdata/design_testfile.txt

@@ -1,13 +1,15 @@
-Date 20220905
-ExperimentID test
-Total_wells 48
+Date: 20220905
+ExperimentID: test
+Total_wells: 48
 # Specify the groups and the different wells below.
 # Always use two spaces before the group name
-# and separate the group and the well names with a space.
-Groups
-  Control D1 D2 D3 D4 D5 D6 D7 D8
-  0.01 F1 F2 F3 F4 F5 F6 F7 F8
-  0.1 E1 E2 E3 E4 E5 E6 E7 E8
-  1 C1 C2 C3 C4 C5 C6 C7 C8
-  10 B1 B2 B3 B4 B5 B6 B7 B8
-  30 A1 A2 A3 A4 A5 A6 A7 A8
+# and separate the group and the well names with a space or comma.
+# This design testfile has some nice weird punctuation
+# to check that everything is still parsed correctly
+Groups:
+  Control: D1 D2 D3 D4 D5 D6 D7 D8
+  0.01: F1,F2,F3 F4 F5 F6 F7 F8
+  0.1: E1 E2,E3 E4 E5 E6 E7 E8
+1: C1 C2 C3 C4  C5 C6   , C7 C8
+  10: B1 B2 B3 B4 B5 B6 B7 B8
+  30: A1 A2 A3 A4 A5 A6 A7  A8

inst/extdata/design_testfile_duplicates.txt

@@ -0,0 +1,6 @@
+Date: 20230407
+ExperimentID: PR
+Total_wells: 48
+Groups:
+control: A3 B3 C3 D3 E3 F3 A4
+0.1: A3 B3 C3 D3 E3 F3 A4

inst/extdata/design_testfile_unequalrows.txt

@@ -0,0 +1,6 @@
+Date: 20230407
+ExperimentID: PR
+Total_wells: 48
+Groups:
+control: B6 B7 B8 C1 C2 C6 C4 B5
+0.1: A3 B3 C3 D3 E3 F3 A4