update

DESCRIPTION

@@ -14,18 +14,13 @@ Imports:
 	units,
 	sf
 Suggests: 
+	ggplot2,
+	ggthemes,
+	viridis,
     testthat,
     knitr,
     covr,
-    tibble,
-    dplyr,
-    tidyr,
-    geosphere (>= 1.5-5), 
-    maptools, 
-    maps,
-	ggplot2,
-	ggthemes,
-	viridis
+	rmarkdown
 LinkingTo: Rcpp
 VignetteBuilder: knitr
 SystemRequirements: GDAL (>= 2.1.0), PROJ.4 (>= 4.8.0)

R/stars.R

@@ -142,17 +142,22 @@ st_stars.list = function(x, ..., dimensions = NULL) {
 #' image(x)
 image.stars = function(x, ..., band = 1, attr = 1, asp = 1, rgb = NULL, maxColorValue = 1,
 		xlab = names(dims)[1], ylab = names(dims)[2]) {
+
+	if (any(dim(x) == 1))
+		x = adrop(x)
+
 	dims = expand_dimensions(x)
-	x = x[[attr]]
+	x = unclass(x[[ attr ]])
 	x = if (length(dim(x)) == 3) {
 			if (is.null(rgb))
 				x[ , rev(seq_len(dim(x)[2])), band]
 			else {
+				stop("not yet supported")
 				xy = dim(x)[1:2]
 				x = structure(x[ , , rgb], dim = c(prod(xy), 3)) # flattens x/y
 				x = rgb(x, maxColorValue = maxColorValue) # FIXME: deal with NAs
 				dim(x) = xy
-				return(rasterImage(x[ , rev(seq_len(dim(x)[2]))], 0, 0, 1, 1, interpolate = FALSE))
+				#return(rasterImage(x[ , rev(seq_len(dim(x)[2]))], 0, 0, 1, 1, interpolate = FALSE))
 			}
 		} else
 			x[ , rev(seq_len(dim(x)[2]))]

documents/stars_data.md

@@ -1,14 +1,14 @@
 # An Earth data processing backend for testing and evaluating stars
 
 [Edzer Pebesma](https://github.com/edzer/), 
-[Michael Sumner](https://github.com/mdsumer/) 
+[Michael Sumner](https://github.com/mdsumner/) 
 
 ## Summary
 
 The amount of freely available data generated by either direct observation from satellites or from weather or climate model calculations exceeds our capacity to download and process locally. The only feasible approach to analyse such data is to send compute jobs to cloud servers with direct access to this data, and have the small summaries (maps, time series, numbers) returned. R lacks software to analyse such datasets on compute servers, and software to serve such processes.
 
-The [stars](https://github.com/r-spatial/stars/) project foresees in
-processing such data, and doing this from a client on a remote server, without
+The [stars](https://github.com/r-spatial/stars/) project will enable the
+processing such data, locally as well as on a remote server, without
 downloading the data. This project, `stars:backend`, will (i) create software
 to run a back-end, (ii) develop scripts and tutorials that explain how such
 a data server and processing backend can be set up, and (iii) create an instance
@@ -42,7 +42,7 @@ The [rOpenSci](http://ropensci.org/) project has also developed several packages
 
 As mentioned, [Google Earth Engine](https://earthengine.google.com/) provides remote processing on Tb-Pb sized dateset, but is not open source, and hence to some extent a black box. In addition, users have limited control over which data it serves, and which software it runs (e.g., it has a python interface but does not allow using numpy)
 
-The Spatial data at the Australian Antarctic Division has [R tools](https://github.com/AustralianAntarcticDivision/raadtools) to organise large amounts of Earth science data and serve it to researchers.
+The Australian Antarctic Division has [R tools](https://github.com/AustralianAntarcticDivision/raadtools) to organise large amounts of Earth science data and serve it to researchers.
 
 ## The Plan: