tm_crs and raster data

[Nowosad]

@mtennekes, have you considered moving raster.warp to the tm_crs function as one of the arguments? It seems like a good home for it, and a little bit easier to explain than tm_options.

See https://tmap.geocompx.org/map-projections#raster-data for examples.

tm_shape(worldelevation) +
  tm_raster("worldelevation.tif", 
            col.scale = tm_scale(values = terrain.colors(8))) +
  tm_crs("EPSG:8857") +
  tm_options(raster.warp = FALSE)

[mtennekes]

I have the feeling that tm_crs is taking over tm_shape. Originally (v3) raster.warp was part of tm_shape, so in that sense that's a logical choice, and so would be tm_crs. Also unit may be added to tm_crs.

Overlap between tm_shape and tm_crs is not an issue but should be clarified. One core but subtle difference is that tm_shape is creating while tm_crs is just adjusting. E.g. tm_shape(bbox = "Utrecht") (in view mode) will generate a plain map, but just tm_crs(bbox = "Utrect") won't work. (Still have to improve error message on that).

[Nowosad]

After some thinking I am unsure if raster.warp fits into tm_crs() -- is the raster.warp argument only used for reprojections or is it also used for other operations, e.g., downscaling?