R/gdalio.R

In hypertidy/gdalio: Read Data to via GDAL Warper to an Assumed Grid

#' Read GDAL raster numeric data
#'
#' Data may be one band (the default, first band) or many.
#'
#' [gdalio_data()] returns a list of vectors, [gdalio_matrix()] and [gdalio_array()] and
#' [gdalio_graphics()] return
#' a matrix, array, matrix of the necessary shape, as used by [image()] and [plot()].
#'
#' The matrix of hex values returned by [gdalio_graphics()] cannot really be placed on a spatial
#' plot window without a lot of extra work, but it's good for fast visuals to 'plot()' the output.
#' We can write helpers to plot this thing better but WIP atm.
#'
#' @param dsn character string, raster source understood by GDAL
#' @param ... arguments passed to 'vapour::vapour_warp_raster'
#' @param bands default 1L, but can be more, duplicated in different order, or 'NULL' for all
#' @return list of numeric vectors
#' @export
#' @examples
#' \dontrun{
#'  f <- system.file("extdata/sst.tif", package = "vapour", mustWork = TRUE)
#'
#'  g <- list(extent = c(130, 160, -60, -30), dimension = c(180, 100),
#'     projection = "+proj=longlat")
#'  gdalio_set_default_grid(g)
#'  v <- gdalio_data(f, band_output_type = "int")
#'  image(seq(130, 160, length.out = 181), seq(-60, -30, length.out = 101),
#'     matrix(v[[1]], g$dimension[1])[,g$dimension[2]:1], asp = 1.5)
#' }
gdalio_data <- function(dsn, ..., bands = 1L) {
 UseMethod("gdalio_data")
}
#' @export
gdalio_data.vrt_simple <- function(dsn, ..., bands = 1L) {
  ## these have to be NULL, or 4 numbers and a string
  src_extent <- .vrt_extent(dsn)
  if (is.null(src_extent)) src_extent <- 0  ## that's what vapour expects, not NULL
  src_proj <- .vrt_projection(dsn)
  gdalio_data(unclass(dsn), source_extent = src_extent, source_projection = src_proj, bands = bands, ...)
}
#' @export
#' @importFrom vapour vapour_warp_raster
gdalio_data.default <- function(dsn, ..., bands = 1L) {
  g <- gdalio_get_default_grid()
  if (utils::packageVersion("vapour") <= "0.8.0") {
    ## catch this old case, it keeps confusing me ...
   out <-  vapour::vapour_warp_raster(dsn, extent = g$extent, dimension = g$dimension, wkt = g$projection, bands = bands,  ...)

  } else {
    out <- vapour::vapour_warp_raster(dsn, extent = g$extent, dimension = g$dimension, projection = g$projection, bands = bands,  ...)
  }
  out
 }

#' Read GDAL raster data as RGB triples or hex colours
#'
#' `gdalio_data_hex` and `gdalio_data_rgb` are a little strange in that they return a
#' vector of character strings and a list of numeric values respectively.
#'
#' @param dsn character string, raster source understood by GDAL
#' @param ... arguments passed to 'vapour::vapour_warp_raster'
#' @param bands bands to read, assumes 1:3 (can be 1:4 or any ordering)
#'
#' @return 'gdalio_data_rgb()' a list of integer vectors, 'gdalio_data_hex()' a character vector of "#" colours
#' @export
#' @name gdalio_data_rgb
#' @export
gdalio_data_rgb <- function(dsn, bands = 1:3, ...) {
 gdalio_data(dsn, bands = bands, band_output_type = "Int32", ...)
}

#' @name gdalio_data_rgb
#' @importFrom grDevices rgb
#' @export
gdalio_data_hex <- function(dsn, bands = 1:3, ...) {
  v <- gdalio_graphics(dsn, bands = bands, ...)
  as.vector(v)
}


#' @name gdalio_data
#' @export
gdalio_graphics <- function(dsn, bands = 1:3, ... ) {
  v <- gdalio_data(dsn, bands = bands, band_output_type = "Byte", ...)
  g <- gdalio_get_default_grid()
  dm <- c(g$dimension, length(v))
  ap <- c(2, 1, 3)
  if (length(v) == 1L) {
    dm <- dm[1:2]
    ap <- ap[1:2]
  }
  a <- aperm(array(do.call(c, v), dm) , ap)
  grDevices::as.raster(a)
}

#' @name gdalio_data
#' @export
gdalio_matrix <- function(dsn, ...) {
  v <- gdalio_data(dsn, ...)
  g <- gdalio_get_default_grid()
  matrix(v[[1]], g$dimension[1])[,g$dimension[2]:1, drop = FALSE]
}
#' @name gdalio_data
#' @export
gdalio_array <- function(dsn, ...) {
  v <- gdalio_data(dsn, ...)
  g <- gdalio_get_default_grid()

  array(v[[1]], c(g$dimension, length(v)))[,g$dimension[2]:1, , drop = FALSE]
}