Limit filetype options for tar_terra_sds() and tar_terra_sprc()

[Aariq]

It would be worthwhile testing with a driver that does not support sub-datasets first to see if the error message is informative—perhaps if it is we don't actually need to do anything

APPEND_SUBDATASET=YES works for GTiff output of both SpatRasterDataset or SpatRasterCollection objects, but other formats will not necessarily work this way. Perhaps these methods should be limited to filetype="GTiff"?

From https://gdal.org/user/raster_data_model.html#subdatasets-domain

Currently, drivers which support subdatasets are: ADRG, ECRGTOC, GEORASTER, GTiff, HDF4, HDF5, netCDF, NITF, NTv2, OGDI, PDF, PostGISRaster, Rasterlite, RPFTOC, RS2, TileDB, WCS, and WMS.

Originally posted by @brownag in #59

[Aariq]

Error message is:

── Last error message ────────────────────
    _store_ [writeRaster] cannot append datasets with this file format

That seems explicit enough to not worry about adding guard rails that we'd have to maintain as new drivers get added. If you agree @njtierney, feel free to close this issue.

[njtierney]

Can you provide a bit more context around the code that produces this error? Or is it linked in the issue above?

[Aariq]

Sure, here's an example that produces this error:

targets::tar_dir({ # tar_dir() runs code from a temporary directory.
    targets::tar_script({
    library(geotargets)
        elev_scale <- function(z = 1, projection = "EPSG:4326") {
            terra::project(
                terra::rast(system.file("ex", "elev.tif", package = "terra")) * z,
                projection
            )
        }
        list(
            tar_terra_sprc(
                raster_elevs,
                command = terra::sprc(list(
                    elev_scale(1),
                    elev_scale(2, "+proj=igh")
                )),
                filetype = "MFF2"
            )
        )
    })
    targets::tar_make()
})

The errors don't show up the same when using the reprex package for some reason, but here's what they look like with targets 1.8.0:

▶ dispatched target raster_elevs
✖ errored target raster_elevs
✖ errored pipeline [0.317 seconds]
Error:
! targets::tar_make() error

── Debug target raster_elevs ───────────────────────────────────────────────────────────────────────────────────
    tar_meta(raster_elevs)$error
    tar_workspace(raster_elevs)

── General debugging ───────────────────────────────────────────────────────────────────────────────────────────
    • tar_errored()
    • tar_meta(fields = any_of("error"), complete_only = TRUE)
    • tar_workspace()
    • tar_workspaces()

── How to ──────────────────────────────────────────────────────────────────────────────────────────────────────
    • Debug: https://books.ropensci.org/targets/debugging.html
    • Help: https://books.ropensci.org/targets/help.html

── Last error message ──────────────────────────────────────────────────────────────────────────────────────────
    _store_ [writeRaster] cannot append datasets with this file format

── Last error traceback ────────────────────────────────────────────────────────────────────────────────────────
    No traceback available.

[Aariq]

Pros of doing our own checking:

• Opportunity to write a better/different error message
• Fail faster, before command is run

Cons:

• Need to maintain our own list of GDAL drivers and versions that support APPEND_SUBDATASET=YES

After seeing your changes in #97 and realizing that APPEND_SUBDATSET requires GDAL ≥ 3.0, I'm more convinced that we should intervene here.