WKT string as srs in cube_view

[jvandoninck]

There seems to be a strange behaviour when defining the srs in gdalcubes::cube_view as a WKT. I created a data cube, but used terra:crs() to set the srs from an existing spatRaster object. In the reproducible example below a similar way to set srs using the example from the package.

library(gdalcubes)
library(terra)

#Data cube (from L8 cube gdalcubes package)
L8_files <- list.files(system.file("L8NY18", package = "gdalcubes"),
                       ".TIF", recursive = TRUE, full.names = TRUE)
create_image_collection(L8_files, "L8_L1TP", file.path(tempdir(), "L8.db"), quiet = TRUE)

L8.col = image_collection(file.path(tempdir(), "L8.db"))
v = cube_view(extent=list(left=388941.2, right=766552.4, 
                          bottom=4345299, top=4744931, t0="2018-01", t1="2018-06"),
              srs=terra::crs("EPSG:32618"), nx = 497, ny=526, dt="P1M",
              aggregation="first", resampling="bilinear")
L8.cube = raster_cube(L8.col, v) 
L8.cube = select_bands(L8.cube, c("B02", "B03", "B04", "B05", "B06"))

No problem so far. An error only occurs when using this data cube in a subsequent reduce_time() function, but only when using the FUN argument, and not the expr argument. In the example below two ways to create a median value composite. I have no clue what is causing this or whether it is machine dependent. When using the proj4 or EPSG notation both functions execute correctly.

#Median compositing function 1 - with expr
cube_composite.median <- function(cube){
  names <- gdalcubes::bands(cube)$name
  comp_args <- list(x=cube)
  comp_args <- c(comp_args, lapply(names, function(x){paste0("median","(",x,")")}))
  composite <- do.call(gdalcubes::reduce_time, comp_args)

  return(composite)
}

#Median compositing function 2 - with FUN
cube_composite.median2 <- function(cube){
  names <- gdalcubes::bands(cube)$name
  composite <- gdalcubes::reduce_time(cube, names=names, FUN=function(x){
    x_median <- apply(x, 1, median, na.rm=TRUE)
    return(x_median)
  })
  return(composite)
}

#Apply the two median composite functions 
L8.median <- cube_composite.median(L8.cube)
plot(L8.median, rgb=3:1)

L8.median2 <- cube_composite.median2(L8.cube)
plot(L8.median2, rgb=3:1)