The link2GI
package provides a small linking tool to simplify the use of
GRASS GIS, SAGA GIS,
Orfeo Toolbox (OTB) and GDAL
binaries for R users, with the focus on making this software accessible
to non-operating system specialists or highly experienced GIS geeks. In
fact, it is a direct result of numerous graduate courses with R(-GIS)
novices in the hostile world of university computer pools running on
extremely limited Windows systems.
R has a lot of classes for storing and manipulating spatial data. For vector data, the sp and currently the great sf packages are well known, and the raster data world is largely covered by the terra and recently the stars packages. For more specific links, such as those needed for manipulating atmospheric models, packages like ncdf4 are very helpful.
The spatial analysis itself is often supported by wrapper packages that integrate external libraries, command line tools, or a mixture of both in an R-like syntax geosphere, Distance, igraph, or spatstat.
A comprehensive introduction to the spatial R-biotope and its backgrounds is excellently treated in Geocomputation with R, which is not only highly recommended as a reference book, but is also an indispensable basis for working and analyzing spatial data with R/Python.
Despite all these spatial analysis and data handling capabilities in
the R world, it can be said (at least from a non-R point of
view) that there is still a huge gap between R and the mature open
source Geographic Information System (GIS) and even more so the Remote
Sensing (RS) software community. QGIS, GRASS GIS, and SAGA GIS provide
an extensive, growing, and mature collection of sophisticated
algorithms. The algorithms provided are fast, stable and most of them
are well proven. Probably most R users who are somehow
connected to the GI community know that there are great wrapper packages
to bridge this gap. For GRASS GIS
7/8 it is rgrass and for SAGA GIS there are RSAGA and Rsagacmd. There is
no mature wrapper package for the [Orfeo Toolbox] (https://gitlab.orfeo-toolbox.org/orfeotoolbox/otb),
which is indispensable for remote sensing. This package not only creates
the API link, but also provides a lightweight wrapper for easy use of
OTB modules from R.
Unfortunately you will run into a lot of technical problems depending
on the chosen operating system (OS) or library dependencies or GIS
software versions. In the case of RSAGA for example, the
main problem was that the SAGA GIS developers not only
change the syntax and strategy of the command line interface (CLI), but
also within the same release the calls differ from OS to OS. So the
maintenance of RSAGA is at least tedious (but thumbs up again). Another
example is GRASS GIS, which is known for a sophisticated
setup of the environment and the spatial properties of the database. If
you “only” want to use a specific GRASS algorithm from R,
you will probably get lost in setting up all the OS dependencies that
are necessary to set up a correct temporary or permanent
GRASS environment from “outside”. This is not only due to
the strict space and projection requirements of GRASS, but
much more due to the demanding OS environments, especially Windows.
To cut a long story short, it is a bit cumbersome to deal with all
this stuff when you just want to start GRASS from the R
command line, e.g. for a powerful random walk cost analysis
(r.walk) call as provided by GRASS.
Linking simply means providing all the necessary environment
variables to allow full access to the command line APIs of the software
tools mentioned. link2GI also tries to be compliant with
the existing wrapper packages and integrate their use.
link2GI tries to analyse which software is installed in
which version in order to create a temporary or permanent environment
that meets the requirements mentioned above.
GRASS GIS has the most demanding requirements. It needs a lot of
environment and path variables as and a correct setup
of the geographic data parameters. The linkGRASS function
tries to find all installations and lets you (optionally) choose the one
you want to use and generate the necessary variables. As a result you
can use both the rgrass package and the command line
API of GRASS.
SAGA GIS is much easier to set up. Again, the linkSAGA
function will try to find all SAGA installations, let you
(optionally) choose one, and generate the necessary variables. You can
also use RSAGA, but you have to pass the result of
linkSAGA like
RSAGA::rsaga.env(path = saga$sagaPath). For easy use you
can just use the R system() call to interface
R with the saga_cmd API.
The Orfeo Toolbox (OTB) is a very powerful remote
sensing toolbox. It is widely used for classification, filtering and
machine learning applications. You will find some of the implemented
algorithms in various R packages, but always much
slower or only running on small data chunks. link2GI
searches and links all OTB installations of a given search
path and returns the result as a concise list. Due to a missing wrapper
package, a list-based OTB module and function parser is
also available, which can be piped into the runOTB function
for a convenient function call.
Although GDAL is well integrated into the R packages, in
quite a few cases it is advantageous to use system calls and to obtain
the binaries directly, especially after the discontinuation of
rgdal and some other convenience packages. In particular,
the development towards GDAL 3.x and optionally different
boxed versions of GDAL binaries that work with different
Python and proj4/proj6 libraries make it often
difficult to find the right GDAL version.
link2GI generates a list of all paths and commands of all
GDAL installations in the given search path. With this
list, you can easily use all available API calls of each
installation.