#' Download GEO dataset and preprocess it

#'

#' Functionality for automatically downloading, RMA preprocessing the array

#' data and formatting the phenotype data, and saving the results in a

#' \code{.Rds} file.

#'

#' @param geo_nbr The GEO ascession number.

#' @param destdir The destination dir of the downloaded files.

#' @param ... Arguments passed to \code{\link{preprocessCELFiles}}

#' @param clean Should the strictly unnessesary files be deleted?

#' @param verbose Signal the process.

#' @return

#'   Creates a folder with the downloaded and processed files.

#'   The processed files are saved as an \code{.Rds} object named

#'   after the used parameters.

#'   Invisibly returns the saved object.

#' @note The function will overwrite existing files in the \code{destdir}.

#'

#'   Arguments after \code{\dots} must be named.

#' @author

#'   Anders Ellern Bilgrau,

#'   Steffen Falgreen Larsen

#' @examples

#' \dontrun{

#' print(DLBCL_overview)

#' geo_nbr <- DLBCL_overview[6,1]

#' res <- downloadAndProcessGEO(geo_nbr = geo_nbr, cdf = "brainarray",

#'                              target = "ENSG", clean = FALSE)

#' head(exprs(res$es$Batch1))

#' }

#' @export

downloadAndProcessGEO <- function(geo_nbr,

                                  destdir = getwd(),

                                  ...,

                                  clean = FALSE,

                                  verbose = TRUE) {

  # Download metadata

  meta_data <- downloadAndPrepareMetadata(geo_nbr = geo_nbr, destdir = destdir,

                                          clean = clean, verbose = verbose)

   # Process metadata

  clean_meta_data <- cleanMetadata(meta_data)

  GSM_met <- basenameSansCEL(rownames(clean_meta_data))

  stopifnot(all(GSM_met == clean_meta_data$GSM))

  # Download array data

  cel_files <- downloadAndPrepareCELFiles(geo_nbr = geo_nbr, destdir = destdir,

                                          clean = clean, verbose = verbose)

  GSM_cel <- gsub("(GSM[0-9]+).*$", "\\1", basenameSansCEL(cel_files))

  # Add local filenames to cleaned data

  clean_meta_data$file <- cel_files[pmatch(GSM_met, GSM_cel)]

  # Checks and warnings

  if (!all(GSM_cel %in% GSM_met)) {

    warning("Not all downloaded CEL files are in the metadata")

  }

  if (!all(GSM_met %in% GSM_cel)) {

    warning("Not all GSM numbers in the metadata have CEL files")

  }

  # Preprocess array data by RMA for each batch

  if (is.null(clean_meta_data$Batch)) {

    es <- preprocessCELFiles(cel_files, ...)

  } else {

    if (verbose) {

      message("Batches detected. RMA normalizing each of the batches: ",

              paste0(unique(clean_meta_data$Batch), collapse = ", "))

    }

    batch_list <- with(clean_meta_data, split(file, Batch))

    es <- lapply(batch_list, preprocessCELFiles, ...)

  }

  # Save Rds file

  a         <- list()

  a$cdf     <- tolower(finfo(es, "cdf"))

  a$target  <- finfo(es, "target")

  a$version <- finfo(es, "version")

  file_name <-

    paste0(geo_nbr, "_", a$cdf,

           ifelse(a$target != "", "_", ""),

           tolower(a$target),

           ifelse(a$cdf == "affy", "", paste0("_", a$version)),

           ".Rds")

  output <- list(es = es, metadata = clean_meta_data, call = match.call())

  saveRDS(output, file = file.path(destdir, geo_nbr, file_name))

  # Clean if wanted

  if (clean) file.remove(cel_files)

  if (verbose) message("done.\n")

  return(invisible(output))

}

finfo <- function(x, y) {

  if (!is.list(x)) {

    x <- list(x)

  }

  info <- unique(unlist(lapply(x, function(e) attributes(e)[[y]])))

  ans <- paste(info, collapse = "-")

  return(ans)

}