grd_data_generic: Wrap data sources with non-standard ordering
In paleolimbot/grd: Raster and Grid Geometry

Whereas grd() objects always index grd_data() using y, x axis order with y values decreasing with increasing index value, data are often stored in other configurations. The grd_data_generic() class wraps the common case where axis order or axis direction differs from the default. The resulting object retains a reference to the underlying data but takes care of rearranging the calls to dim(), [, and [<- such that the object can be indexed and modified in the same way.

grd_data_generic(
  grid_data,
  data_order = grd_data_order(grid_data),
  ptype = grd_data_ptype(grid_data)
)

`grid_data`	The `data` member of a `grd()`. This is typically an array but can also be an S3 object with the proper methods implemented (e.g., `grd_data_generic()`).
`data_order`	A character vector with the same length as `dim(grid_data)` specifying the axis order and axis direction of indices in the x y direction. The default `c("y", "x")` indicates column-major ordering with y values decreasing in the positive i index direction and x values increasing in the positive j index direction. Use `"-y"` or `"-x"` to switch axis directions. Use `NA` to indicate a non-xy dimension.
`ptype`	The R object type that should be used to represent the data.