R/storeLandsat.R

In LSFE/LSFE-R: Lehrstuhl für Fernerkundung

#' @title storeLandsat
#'
#' @description Temporal linear interpolation of raster data.
#' @param zpPath Path where the zipped landsat data is stored
#' @param dPath Path where the unzipped landsat data should be stored.
#' @param c1 Is the data from USGS collection 1. Optional logicla argument. Default is TRUE.
#' @param remove.files Should the zipped files be deleted after unpacking? Optional logical argument. Default is FALSE.
#' @import raster sp
#' @return A folder structure containing the input Landsat data
#' @details {Local storing of landsat data as downloaded from the ESPA downloading service. A folder
#' named \emph{LANDSAT}} will be create in the target directory and the output will be split between
#' two sub-folder:
#' \itemize{
#' \item{\emph{"sr"} - Surface reflectances.}
#' \item{\emph{"infos"} - Metadata.}}
#' This function unzips the downloaded files and buils a standardized data structure from it with \emph{sr}
#' (i.e. TILE -> ACQUISITION -> FILES). Additionally, the function translates thebit quality information into
#' an FMASK equivalent with labels for Water (1), Cloud Shadow (2),Snow (3), Cloud (4) and No Data (255) and
#' provides metadata reporting on:
#' \itemize{
#' \item{\emph{"Date"} - Acquisition date.}
#' \item{\emph{"Tile"} - Acquisition tile (i.e. path/row.}
#' \item{\emph{"Path"} - Path to where the data is stored.}
#' \item{\emph{"Cover"} - Cloud cover percent.}
#' \item{\emph{"Processed"} - Date when the file was processed.}}
#' @examples \dontrun{
#'
#' }
#' @export
#' @seealso \code{\link{storeLandsat}}

#---------------------------------------------------------------------------------------------------------------------#

storeLandsat <- function(zpPath, dPath, c1=TRUE, remove.files=FALSE) {

#---------------------------------------------------------------------------------------------------------------------#
# 0. check input variables
#---------------------------------------------------------------------------------------------------------------------#

  # test source/output directories
  if (!is.logical(c1)) {stop('"c1" is not a logical argument')}
  if (!is.logical(remove.files)) {stop('"remove.files" is not a logical argument')}
  if (!exists('zpPath')) {stop('"zpPath" missing')} else {zpPath <- file.path(zpPath)}
  if (!exists('dPath')) {stop('"dPath" missing')} else {dPath <- file.path(dPath)}

#---------------------------------------------------------------------------------------------------------------------#
# 1. create base directories
#---------------------------------------------------------------------------------------------------------------------#

  # create base directory
  ltPath <- paste0(dPath, '/LANDSAT/')
  if (!dir.exists(rPath)) {dir.create(ltPath)}

  # create SR/infos folder
  mPath <- paste0(ltPath, '/infos/') #  infos
  if (!dir.exists(mPath)) {dir.create(mPath)}
  rPath <- paste0(ltPath, '/sr/') # reflectances
  if (!dir.exists(rPath)) {dir.create(rPath)}

#---------------------------------------------------------------------------------------------------------------------#
# 2. extract file info
#---------------------------------------------------------------------------------------------------------------------#

  # list zip files
  files <- list.files(zpPath, 'tar.gz', full.names=T)
  cc <- vector('numeric', length(files)) # clear pixels %

  # extract date information
  adate <- as.Date(paste0(substr(basename(files), 11, 14), '-',
                          substr(basename(files), 15, 16), '-',
                          substr(basename(files), 17, 18)))
  pdate <- sapply(files, function(x){strsplit(basename(x), '-')[[1]][2]})
  pdate <- as.Date(paste0(substr(basename(pdate), 3, 6), '-',
                          substr(basename(pdate), 7, 8), '-',
                          substr(basename(pdate), 9, 10)))

  # determine tiles
  tiles <- unname(sapply(files, function(x) {substr(strsplit(basename(x), '-')[[1]][1], 5, 10)}))
  ut <- unique(tiles)

#---------------------------------------------------------------------------------------------------------------------#
# 3. unzip and store files
#---------------------------------------------------------------------------------------------------------------------#

  # determine output directories
  odr <- as.character(sapply(1:length(files), function(x){paste0(rPath, tiles[x], '/', strsplit(basename(files[x]), '-')[[1]][1])}))
  
  for (t in 1:length(ut)) {

    # make/check target directory
    tPath <- paste0(rPath, ut[t], '/')
    if(!dir.exists(tPath)) {dir.create(tPath)}
    ind <- which(tiles==ut[t])

    for (f in 1:length(ind)) {
      
      if(!dir.exists(odr[ind[f]])) {dir.create(odr[ind[f]])}
      
      # unzip file
      untar(files[ind[f]], exdir=odr[ind[f]], tar = "internal")

      # if dealing with collection 1 translate quality layer
      if (c1) {
        cc[ind[f]] <- ltBit(odr[ind[f]])
      } else {
        r <- raster(list.files(odr[ind[f]], 'fmask.tif'))
        cc[ind[f]] <- cellStats(r==0) / cellStats(r!=255, sum) * 100
        rm(r)
      }
      
    }
  }

#---------------------------------------------------------------------------------------------------------------------#
# 4. write metadata
#---------------------------------------------------------------------------------------------------------------------#

  # save metadata
  df <- data.frame(Date=adate, Path=odr, Tile=tiles, Cover=cc, Processed=pdate, stringsAsFactors=F)
  ofile <- paste0(mPath, 'metadata.csv')
  if (file.exists(ofile)) {write.csv(df, opath, append=TRUE)} else {write.csv(df, opath)}

  # if prompted, delete zipped files
  if (remove.files) {for (f in 1:length(files)) {file.remove(files[f])}}

}