Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

Today we are proud to announce a major release of the RQGIS package providing an interface between R and QGIS. We have completeley rewritten RQGIS by now using reticulate to establish a tunnel to the Python QGIS API. To make RQGIS even more user-friendly, we have implemented, among others, following features:

• set_env now caches its output, so you have to call it only once
• find_algorithms now supports regular expressions
• run_qgis now accepts R named arguments as input (see example below). The load_output argument is now logical. If TRUE all specified output files will be loaded into R.
• RQGIS now supports simple features

For a detailed overview of all changes, please refer to the release news.

Let’s use RQGIS for calculating the slope and the aspect of a digital elevation model. First of all, we have to attach RQGIS and a digital elevation model (dem) to the global environment:

library("RQGIS")
data("dem", package = "RQGIS")

Next, we need to find all paths necessary to run successfully the Python QGIS API. Please note, that the output of set_env will be cached and reused in subsequent function calls. Note also, that set_env is always faster when you indicate the path to your QGIS installation, in my case e.g., set_env(root = "C:/OSGeo4W64"). Secondly, we open a QGIS Python application with open_app.

set_env()
open_app()

Now, that we have established a tunnel to the QGIS Python API, we are ready for some geoprocessing. First, we need a geoalgorithm, that computes the slope and the aspect of a digital elevation model. To find the name of such a geoalgorithm, we use a regular expression which searches for the terms slope and aspect in all available QGIS geoalgorithms.

find_algorithms("slope(.+)?aspect")

## [1] "r.slope.aspect - Generates raster layers of slope, aspect, curvatures and partial derivatives from a elevation raster layer.--->grass7:r.slope.aspect"
## [2] "Slope, aspect, curvature----------------------------->saga:slopeaspectcurvature"
## [3] "r.slope.aspect - Generates raster layers of slope, aspect, curvatures and partial derivatives from a elevation raster layer.--->grass:r.slope.aspect"

We will use grass7:r.slope.aspect. To retrieve its function parameters and arguments you can run get_usage("grass7:r.slope.aspect"). Use get_args_man to collect the parameter-argument list including all default values:

params <- get_args_man("grass7:r.slope.aspect")

As with previous RQGIS releases, you can still modify the parameter-argument list and submit it to run_qgis:

params$elevation <- dem params$slope <- "slope.tif"
params$aspect <- "aspect.tif" run_qgis(alg = "grass7:r.slope.aspect", params = params) But now you can also use R named arguments in run_qgis, i.e. you can specify the geoalgorithms parameters directly in run_qgis (adapted from package rgrass7). Here, we specify the input- and the output-rasters. For all other parameters, default values will automatically be used. For more information on the R named arguments, please refer also to the documentation by running ?run_qgis and/or ?pass_args. out <- run_qgis(alg = "grass7:r.slope.aspect", elevation = dem, slope = "slope.tif", aspect = "aspect.tif", load_output = TRUE) ##$slope
## [1] "C:\\Users\\pi37pat\\AppData\\Local\\Temp\\RtmpmsSSA4/slope.tif"
##
## $aspect ## [1] "C:\\Users\\pi37pat\\AppData\\Local\\Temp\\RtmpmsSSA4/aspect.tif" ## ##$dxy
## [1] "C:\\Users\\pi37pat\\AppData\\Local\\Temp\\processing5bb46293bfb243f092a57ce9cf50348b\\ac7b8544e8194dd1a1b8710e6091f1f3\\dxy.tif"
##
## $dxx ## [1] "C:\\Users\\pi37pat\\AppData\\Local\\Temp\\processing5bb46293bfb243f092a57ce9cf50348b\\1576d9dc93434a578b3aeb16bedb17a2\\dxx.tif" ## ##$tcurvature
## [1] "C:\\Users\\pi37pat\\AppData\\Local\\Temp\\processing5bb46293bfb243f092a57ce9cf50348b\\afea27676cc049faaa8526a486f13f70\\tcurvature.tif"
##
## $dx ## [1] "C:\\Users\\pi37pat\\AppData\\Local\\Temp\\processing5bb46293bfb243f092a57ce9cf50348b\\2d71fd26b1aa4868a5cdfd0d7ad47a0c\\dx.tif" ## ##$dy
## [1] "C:\\Users\\pi37pat\\AppData\\Local\\Temp\\processing5bb46293bfb243f092a57ce9cf50348b\\458f38f6c71947d3a37532e4bc6a6a53\\dy.tif"
##
## $pcurvature ## [1] "C:\\Users\\pi37pat\\AppData\\Local\\Temp\\processing5bb46293bfb243f092a57ce9cf50348b\\80ad6fa1843d4d3a92ed0b4c6a39dcfa\\pcurvature.tif" ## ##$dyy
## [1] "C:\\Users\\pi37pat\\AppData\\Local\\Temp\\processing5bb46293bfb243f092a57ce9cf50348b\\6c52d235c8614a719954f1f744e3fef1\\dyy.tif"

Setting load_output to TRUE automatically loads the resulting QGIS output back into R. Since we have specified two ouput files, the output will be loaded into R as a list (here named out) with two elements: a slope and an aspect raster. However, in the background, QGIS calculates all terrain attributes and derivatives provided by grass7:r.slope.aspect, and saves them to a temporary location. Of course, if you wish you can still access these layers from there.

Before running RQGIS, you need to install third-party software. We wrote a package vignette, which guides you through the installation process of QGIS, GRASS and SAGA on various platforms. To access the vignette, please run:

vignette("install_guide", package = "RQGIS")

For more information on RQGIS and examples how to use RQGIS, please refer to its github page and my blog.