Document merge option

R/download.R

@@ -80,6 +80,12 @@ rsi_download_rasters <- function(items,
     gdalwarp_options <- set_gdalwarp_extent(gdalwarp_options, aoi, NULL)
   }
 
+  # Which loop gets parallelized is determined based on which has more steps,
+  # working from the assumption that all downloads take about as long
+  #
+  # so if we aren't merging, or if there's more assets than tiles, we'll
+  # parallelize the outside loop that walks over assets (which, unless the user 
+  # has set up wild nested futures, turns the inside one into a serial process)
   asset_iterator <- ifelse(
     merge || (n_tiles_out < ncol(download_locations)),
     function(...) future.apply::future_lapply(..., future.seed = TRUE),

R/get_stac_data.R

@@ -132,7 +132,10 @@
 #' @param composite_function Character of length 1: The name of a
 #' function used to combine downloaded images into a single composite
 #' (i.e., to aggregate pixel values from multiple images into a single value).
-#' Must be one of of "sum", "mean", "median", "min", "max".
+#' Options include "merge", which 'stamps' images on top of one another such that
+#' the "last" value downloaded for a pixel -- which isn't guaranteed to be the most
+#' recent one -- will be the only value used, or any of "sum", "mean", "median", 
+#' "min", or "max", which consider all values available at each pixel.
 #' Set to `NULL` to not composite
 #' (i.e., to rescale and save each individual file independently).
 #' @inheritParams rstac::stac_search