[R] Remote controlling R via sockets

[Tim CHURCHES]

Many thanks to the R core team for their work on R connections. I was delighted to discover that it is now possible to remote-control R-1.3.0 via a socket connection - something which Brian Ripley noted was not yet possible with R-1.2.0 back in January 2001 in the first R News newsletter. How? Just substitute source() for readLines() in last few lines of the example for a socket connection given in the R-1.3.0 help(connection) - the following works fine on a Red Hat Linux 6.2 system - under Windows the while(isIncomplete()) loop exits immediately for some reason, although the socket connection as a source for source() works fine:

     ## two R processes communicating via non-blocking sockets
     # R process 1
     con1 <- socketConnection(port = 6011, server=TRUE)
     writeLines("plot(rnorm(100))", con1)
     writeLines("plot(rnorm(1000))", con1)
     writeLines("plot(rnorm(10000))", con1)
     close(con1)

     # R process 2
     con2 <- socketConnection(Sys.info()["nodename"], port = 6011)
     # as non-blocking, may need to loop for input
     source(con2)
     while(isIncomplete(con2)) {source(con2)}
     close(con2)

However, in the above example, the R "server" is on the client side of the socket connection. Logically, it should be possible to reverse this and have the R server session acting as the socket server, but I have been unable to get this to work:

     ## two R processes communicating via non-blocking sockets
     # R process 1
     con1 <- socketConnection(port = 6011, server=TRUE)
     source(con1)
     while(isIncomplete(con1)) {source(con1)}
     close(con1)

     # R process 2
     con2 <- socketConnection(Sys.info()["nodename"], port = 6011)
     writeLines("plot(rnorm(100))", con2)
     writeLines("plot(rnorm(1000))", con2)
     writeLines("plot(rnorm(10000))", con2)
     close(con2)

Playing with various combinations of the blocking and open modes have been to no avail. Can anyone suggest some alternative strategies? BTW, my ultimate aim is to use Python to control the R "server". Substituting a Python script for "R process 1" in the first example above works a treat - you can type R commands at the Python prompt and they execute in an R session on a machine on the other side of the room. The only problem is that with R as the socket client the Python process can't connect and disconnect from it as it could if R were the socket server. As it is I think this capability is already very exciting, and the prospect of an R socket server daemon seems tantalisingly close. I don't have enough experience with socket programming to try this myself but I would be very happy to collaborate with anyone who wanted to attack this problem.

Tim Churches





*********************************************************************
This message is intended for the addressee named and may
contain confidential information. If you are not the intended
recipient, please delete it and notify the sender. Views 
expressed in this message are those of the individual sender,
and are not necessarily the views of NSW Health.
*********************************************************************

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._