R/HelperUtils.R

In GreenleafLab/ArchR: Analyzing single-cell regulatory chromatin in R.

##########################################################################################
# S4Vectors/BiocGenerics Within Methods
##########################################################################################

#' Negated Value Matching
#'
#' This function is the reciprocal of %in%. See the match funciton in base R.
#'
#' @param x The value to search for in `table`.
#' @param table The set of values to serve as the base for the match function.
#' @export
"%ni%" <- function(x, table) !(match(x, table, nomatch = 0) > 0)

#' Generic matching function for S4Vector objects
#'
#' This function provides a generic matching function for S4Vector objects primarily to avoid ambiguity.
#'
#' @param x An `S4Vector` object to search for in `table`.
#' @param table The set of `S4Vector` objects to serve as the base for the match function.
#' @export
'%bcin%' <- function(x, table) S4Vectors::match(x, table, nomatch = 0) > 0

#' Negated matching function for S4Vector objects
#'
#' This function provides the reciprocal of %bcin% for S4Vector objects primarily to avoid ambiguity.
#'
#' @param x An `S4Vector` object to search for in `table`.
#' @param table The set of `S4Vector` objects to serve as the base for the match function.
#' @export
'%bcni%' <- function(x, table) !(S4Vectors::match(x, table, nomatch = 0) > 0)

##########################################################################################
# Helper to try to reformat fragment files appropriately if a bug is found
##########################################################################################

#' Reformat Fragment Files to be Tabix and Chr Sorted
#'
#' This function provides help in reformatting Fragment Files for reading in createArrowFiles.
#' It will handle weird anomalies found that cause errors in reading tabix bgzip'd fragment files.
#'
#' @param fragmentFiles A character vector the paths to fragment files to be reformatted
#' @param checkChrPrefix A boolean value that determines whether seqnames should be checked to contain
#' "chr". IF set to `TRUE`, any seqnames that do not contain "chr" will be removed from the fragment files.
#' @export
reformatFragmentFiles <- function(
  fragmentFiles = NULL,
  checkChrPrefix = getArchRChrPrefix()
  ){

  .validInput(input = fragmentFiles, name = "fragmentFiles", valid = c("character"))
  .validInput(input = checkChrPrefix, name = "checkChrPrefix", valid = c("boolean"))

  options(scipen = 999)
  .requirePackage("data.table")
  .requirePackage("Rsamtools")
  for(i in seq_along(fragmentFiles)){
    message(i, " of ", length(fragmentFiles))
    dt <- data.table::fread(fragmentFiles[i])
    dt <- dt[order(dt$V1,dt$V2,dt$V3), ]
    if(checkChrPrefix){
      idxRemove1 <- which(substr(dt$V1,1,3) != "chr")
    }else{
      idxRemove1 <- c()
    }
    idxRemove2 <- which(dt$V2 != as.integer(dt$V2))
    idxRemove3 <- which(dt$V3 != as.integer(dt$V3))
    #get all
    idxRemove <- unique(c(idxRemove1, idxRemove2, idxRemove3))
    if(length(idxRemove) > 0){
      dt <- dt[-idxRemove,]
    }
    if(nrow(dt) == 0){
      if(checkChrPrefix){
        stop("No fragments found after checking for integers and chrPrefix!")
      }else{
        stop("No fragments found after checking for integers!")
      }
    }
    #Make sure no spaces or #
    dt$V4 <- gsub(" |#", ".", dt$V4)
    fileNew <- gsub(".tsv.bgz|.tsv.gz", "-Reformat.tsv", fragmentFiles[i])
    data.table::fwrite(dt, fileNew, sep = "\t", col.names = FALSE)
    Rsamtools::bgzip(fileNew)
    file.remove(fileNew)
    .fileRename(paste0(fileNew, ".bgz"), paste0(fileNew, ".gz"))
  }
}


##########################################################################################
# Helper For cluster identity
##########################################################################################

#' Create a Confusion Matrix based on two value vectors
#'
#' This function creates a confusion matrix based on two value vectors.
#'
#' @param i A character/numeric value vector to see concordance with j.
#' @param j A character/numeric value vector to see concordance with i.
#' @export
confusionMatrix <- function(
  i = NULL, 
  j = NULL
  ){
  ui <- unique(i)
  uj <- unique(j)
  m <- Matrix::sparseMatrix(
    i = match(i, ui),
    j = match(j, uj),
    x = rep(1, length(i)),
    dims = c(length(ui), length(uj))
  )
  rownames(m) <- ui
  colnames(m) <- uj
  m
}


#' Re-map a character vector of labels from an old set of labels to a new set of labels
#'
#' This function takes a character vector of labels and uses a set of old and new labels
#' to re-map from the old label set to the new label set.
#'
#' @param labels A character vector containing lables to map.
#' @param newLabels A character vector (same length as oldLabels) to map labels to from oldLabels.
#' @param oldLabels A character vector (same length as newLabels) to map labels from to newLabels
#' @export
mapLabels <- function(labels = NULL, newLabels = NULL, oldLabels = names(newLabels)){

  .validInput(input = labels, name = "labels", valid = c("character"))
  .validInput(input = newLabels, name = "newLabels", valid = c("character"))
  .validInput(input = oldLabels, name = "oldLabels", valid = c("character"))

  if(length(newLabels) != length(oldLabels)){
    stop("newLabels and oldLabels must be equal length!")
  }

  if(!requireNamespace("plyr", quietly = TRUE)){
    labels <- paste0(labels)
    oldLabels <- paste0(oldLabels)
    newLabels <- paste0(newLabels)
    labelsNew <- labels
    for(i in seq_along(oldLabels)){
        labelsNew[labels == oldLabels[i]] <- newLabels[i]
    }
    paste0(labelsNew)
  }else{
    paste0(plyr::mapvalues(x = labels, from = oldLabels, to = newLabels))
  }

}