R/deannotateDetails.R

In shirewoman2/Consultancy: Standardized tools for using R within the Simcyp Consultancy Team

#' De-annotate a data.frame of experimental details that was previously
#' annotated
#'
#' \code{deannotateDetails} takes a data.frame that contains annotated
#' experimental details and converts it from having columns for each file and
#' rows for each experimental detail to having rows for each file and columns
#' for each detail.
#'
#' @param existing_exp_details the annotated output from
#'   \code{extractExpDetails} or \code{extractExpDetails_mult} where the
#'   argument "annotate_details" was set to TRUE or the output from
#'   \code{annotateDetails}
#' @param apply_class TRUE (default) or FALSE for whether to set the class of
#'   each column to numeric or character data based on the data classes listed
#'   in \code{AllExpDetails}. If set to FALSE, all data classes will be
#'   character.
#'
#' @return a list of simulator experimental details
#' @export
#'
#' @examples
#'
#' Annotated <- annotateDetails(MDZdetails)
#' Deannotated <- deannotateDetails(Annotated)
#' 

deannotateDetails <- function(existing_exp_details, 
                              apply_class = TRUE){
   
   FileOrder <- names(existing_exp_details$MainDetails)[
      str_detect(names(existing_exp_details$MainDetails), "xlsx")]
   
   CompoundNames <- existing_exp_details$MainDetails %>% 
      select(any_of(c("Compound", "CompoundID")), matches("xlsx$")) %>% 
      pivot_longer(cols = !any_of(c("Compound", "CompoundID")),
                   names_to = "File", values_to = "Value") %>%
      filter(complete.cases(Value) & complete.cases(Compound)) %>% 
      select(File, Compound, CompoundID) %>% unique() %>% 
      mutate(File = factor(File, levels = FileOrder))
   
   existing_exp_details$MainDetails <- existing_exp_details$MainDetails %>% 
      select(!any_of(c("SimulatorSection", "Sheet", "Notes",
                       "CompoundID", "Compound",
                       "All files have this value for this compound ID and compound",
                       "All files have this value for this compound ID"))) %>% 
      pivot_longer(cols = -Detail, 
                   names_to = "File", values_to = "Value") %>% 
      # Need to remove NA values here b/c they can otherwise lead to
      # multiple values for a given detail when one value is for the
      # correct compound and the other is for compounds that are present
      # in other files but DO have that particular compound ID. For
      # example, metabolite 1 might be OH-MDZ for some files and
      # OH-bupropion for others, and that will have multiple rows. Removed
      # NA values should mostly come out in the wash, I think, but there
      # is a risk that we'll lose some NA values that should be included.
      # I think that's an acceptable risk. - LSh
      filter(complete.cases(Value)) %>% unique() %>% 
      pivot_wider(names_from = Detail, values_from = Value) %>% 
      mutate(File = factor(File, levels = FileOrder))
   
   if(apply_class){
      suppressWarnings(suppressMessages(
         existing_exp_details$MainDetails <- existing_exp_details$MainDetails %>% 
            mutate(across(.cols = any_of(AllExpDetails %>% 
                                            filter(Class == "numeric") %>% pull(Detail)), 
                          .fns = as.numeric), 
                   across(.cols = matches("CLint|CLadd|CLbiliary|CLrenal|CLiv|^fu|Km|^Vmax|Jmax|RAFREF|HalfLife|^Ind|^MBI"),
                          .fns = as.numeric))
      ))
   }
   
   existing_exp_details$MainDetails <- 
      existing_exp_details$MainDetails %>% mutate(File = factor(File, levels = FileOrder)) %>% 
      arrange(File) %>% 
      mutate(File = as.character(File))
   
   return(existing_exp_details)
   
}