Minimal reticulate rioxarray dask implementation

rio_xarray_dask_inout.R

library(reticulate)
py_require(c("dask", "distributed", "rioxarray"))

dask.distributed <- reticulate::import("dask.distributed")
rioxarray <- reticulate::import("rioxarray")
Sys.setenv(GDAL_VRT_ENABLE_PYTHON = "YES")


input_vrt <- "..."
output_tif <- "..."

with(dask.distributed$LocalCluster(n_workers = 5L) %as% cluster, {
  with(dask.distributed$Client(cluster) %as% client, {
    cli::cli_alert_info("Dask dashboard located at: {client$dashboard_link}")
    xds <- rioxarray$open_rasterio(
      input_vrt,
      chunks = TRUE,
      lock = FALSE,
      # lock=Lock("rio-read", client=client), # when too many file handles open
    )
    xds$rio$to_raster(
      output_tif,
      tiled = TRUE,
      lock = dask.distributed$Lock("rio", client = client),
    )
  })
})