R/fun-clean_column_names.R

In olobiolo/siscreenr: Analyze Genome-Wide siRNA Screening Campaigns

#' clean up superfluous characters in column names
#'
#' Remove gate and object names from column names. Some reformatting to make them easier to read.
#'
#' ScanR export files are tab delimited tables whose columns contain the name of the object,
#' the gate it belongs to, the parameter in question and the measure being exported, e.g. mean or error.
#' The names are cumbersome and when loaded with \code{check.names = TRUE}, they can be quite confusing.
#' Usually most of these can be removed without any percievable loss.
#'
#' This is a simple stream of gsub calls that removes superfluous information from the column names.
#' It also improves the formatting. \code{make.names} is run first.
#'
#' @param x a \code{data.frame}
#' @param gate gate name to be cleared, given as a regular expression;
#'             defaults to "dummy", which does nothing, unless there is a gate called "dummy"
#' @param object name of object to be cleared; defaults to "Main"
#' @return data frame with modified column names
#'
#' @export
#'

clean_column_names <- function(x, gate = 'dummy', object = 'Main') {
  if (!inherits(x, what = 'data.frame')) stop('x must be a data frame')

  y <- names(x)
  y <- make.names(y)
  y <- gsub('....in.', '_in_', y, fixed = TRUE)
  y <- gsub(paste0(gate, '_in_', gate), gate, y, fixed = TRUE)
  y <- gsub(paste0(gate, '\\.'), '', y)
  y <- gsub(paste0(object, '\\.'), '', y)
  y <- gsub('\\.Mean', '', y)
  y <- gsub('\\.\\.Counts', '_count', y)
  names(x) <- y
  return(x)
}