View source for Calling an R function

R is an open source programming language commonly used for data analysis. 
This page describes how to call an existing R function from a model running in the {{Analytica Developer}} edition.  


<center>** This is currently just a place to take notes -- I do not yet have a functioning example **</center>

== Using the RDCOMServer package ==
One approach of many is to use the [https://www.omegahat.net/RDCOMServer/ RDCOMServer] package in R.

This section includes code that enables an Analytica model to call an R function; however, it requires the installation of an R-side package called <code>RDCOMServer</code>. I was unable to get this package installed, so I have not been able to verify that the code works. I am not an R user, so this might be something an R user would know how to do.  If you know how to install the R package, please update the [[#Install packages]] section below.

=== Installation of R ===

# [https://cran.r-project.org/bin/windows/base/ Install R]
## After installer finished, launch R
## Type: <code>Sys.setenv(R_LIBS_USER="C:/Users/«yourName»/AppData/Local/R/win-library/4.3")</code>
##:where you replace «yourName» with your login name.
#:The setenv step is required because R installs into <code>"C:\Program Files"</code> by default, but then new packages can't be installed there, so this puts them into a writable location.

=== Install packages ===
''This is the hard part.''

# Install [https://www.omegahat.net/RDCOMServer/ the RDCOMServer package] for R
## Download the Windows Binary ZIP file
## Unzip it into <code>%R_HOME%\library</code>. For me this was <code>"C:\Program Files\R\R-4.3.0\library"</code>
# Install [https://www.omegahat.net/SWinRegistry/ the SWinRegistry package].
## Download the Windows Binary ZIP file
## Unzip it into <code>%R_HOME%\library</code>. For me this was <code>"C:\Program Files\R\R-4.3.0\library"</code>
## In a CMD prompt: 
##:<code>CD "C:\Program Files\R\R-4.3.0\bin\x64"</code>
##:<code>Rcmd install RWinRegistry</code>
# Install the R package [https://bioconductor.org/install/ Ruuid].
## Start R
## <code>install.packages("BiocManager")</code>
# Launch R
# Run:
#:<code>install.packages("devtools")</code>
#:<code>devtools::install_github("omegahat/RDCOMServer")</code>

-- ''I was unable to get R packages to install. As someone who doesn't use R, I'm afraid I don't know enough to get to first base to even get started.  Aborted here. Sorry.''

=== Server-side code ===

* ''Note: This part is untested, since I haven't yet been able to successful install the required R packages. ''

Here's example R-code that creates a simple COM server with a single method named multiply:
 library(RDCOMServer) 
 &nbsp;
 serverDesc <- COMCreate("R.Server")
 serverDesc$setClass("R.Server")
 &nbsp;
 # Create the `multiply` method
 serverDesc$addMethodDef(
   functionName = "multiply",
   args = c("x" = "numeric", "y" = "numeric"),
   return = "numeric",
   def = function(self, x, y) x * y
 )
 &nbsp;
 # Register the server
 serverDesc$register()
 &nbsp;
 # Start the server
 COMServer()

== Analytica-side call ==

# Create the variable:
#:Variable RServer ::= <code>COMCreateObject("R.Server")</code>
#:''Note: Launch the R-side server before evaluating <code>RServer</code>
# Call the method within a different variable:
#:Variable Y ::= <code>RServer->multiply(5,7)</code>