responding yes to terminal prompt via system2() in R - r

tl;dr: How can I invoke the system command y | conda create --name gee_interface from an R console, e.g. via system2()? I'm comfortable enough with system2('conda', c('create', '--name', 'gee_interface')), but I don't know how to handle piping in the 'y' via system2().
Details
I am trying to use an R console to run the bash command conda create --name gee_interface (OSX Mojave with Anaconda installed).
In terminal, that command executes just fine, but prompts me to answer with Proceed ([y]/n)? (I answer 'y' and everything works smoothly).
In R, I run
Sys.setenv(PATH = paste(c("/Applications/anaconda3/bin", Sys.getenv("PATH")), collapse = .Platform$path.sep)) # ensures that system2() finds conda
system2('conda', c('create', '--name', 'gee_interface')) # This is the key line for the purposes of this question
When running the second line [i.e. system2('conda', c('create', '--name', 'gee_interface'))], the process never finishes, but quickly falls to zero CPU usage. Presumably the system is waiting for my response to the prompt, but I don't know how to provide it. How does one do this via an R script? Note also that in my particular case, the number of times that I need to respond 'y' is variable, depending on whether an environment of the name gee_interface already exists or not.

The fix to your first problem is to tell conda not to ask for confirmation using -y:
system2('conda', c('create', '--name', 'gee_interface', '-y'))
As to the second part (variable times that your input is required), I'm guessing it's to overwrite the environment if it exists? In that case, you could check for its existence first with conda info --envs, and run conda remove --name gee_interface --all if necessary before creating it.
See:
https://docs.conda.io/projects/conda/en/latest/commands/create.html
https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html#removing-an-environment
You could also try your system2 call, with the argument input = "y", but that doesn't fix your second problem of needing to affirm multiple times.
See: Invoke a system command and pipe a variable as an argument

Related

How to release R's prompt when using 'system'?

I am writing an R code on a Linux system using RStudio. At some point in the code, I need to use a system call to a command that will download a few thousand of files from the lines of a text file:
down.command <- paste0("parallel --gnu -a links.txt wget")
system(down.command)
However, this command takes a little while to run (a couple of hours), and the R prompt stays locked while the command runs. I would like to keep using R while the command runs on the background.
I tried to use nohup like this:
down.command <- paste0("nohup parallel --gnu -a links.txt wget > ~/down.log 2>&1")
system(down.command)
but the R prompt still gets "locked" waiting for the end of the command.
Is there any way to circumvent this? Is there a way to submit system commands from R and keep them running on the background?
Using ‘processx’, here’s how to create a new process that redirects both stdout and stderr to the same file:
args = c('--gnu', '-a', 'links.txt', 'wget')
p = processx::process$new('parallel', args, stdout = '~/down.log', stderr = '2>&1')
This launches the process and resumes the execution of the R script. You can then interact with the running process via the p name. Notably you can signal to it, you can query its status (e.g. is_alive()), and you can synchronously wait for its completion (optionally with a timeout after which to kill it):
p$wait()
result = p$get_exit_status()
Based on the comment by #KonradRudolph, I became aware of the processx R package that very smartly deals with system process submissions from within R.
All I had to do was:
library(processx)
down.command <- c("parallel","--gnu", "-a", "links.txt", "wget", ">", "~/down.log", "2>&1")
processx::process$new("nohup", down.comm, cleanup=FALSE)
As simple as that, and very effective.

R system functions always returns error 127

I need to execute an external tool from R and process errors (if any) occurred in that tool.
I know 3 functions to do something familiar with my task:
shell, system and system2.
Trying to test those, I see that command
shell("notepad")
opens notepad. As far as I know shell doesn't allow to check errors (there's no interface to look into stderr).
When I call
system("notepad")
or
system2("notepad")
R freezes trying to make those commands.
Calling
system("start notepad")
or
system2("start notepad")
returns warning
Warning message:
running command '"start notepad"' had status 127
Adapting #DavidTseng's answer (sorry for not having enough reputation to upvote it)...
system("cmd.exe", input = "notepad")
worked for me in Windows.
As I mentioned in my comments, the R documentation reveals that in Windows the system() function does not launch a separate shell (if needed). This is why command line commands run with system(), but Notepad, which needs a separate window, does not run:
From the documentation for system():
The most important difference is that on a Unix-alike system launches a shell which then runs command. On Windows the command is run directly – use shell for an interface which runs command via a shell (by default the Windows shell cmd.exe, which has many differences from a POSIX shell).
system("bash -l", input = "notepad")
I'm not sure if there's been an update to R that allows this since the question was asked nearly four years ago, but system("\"C:\path\to\exe.exe\" args", intern = T) works for me and WILL bring up a separate child window and works on Windows 10 + R 3.6 + RStudio.
Not using the 'intern = T' was giving me a return code of 127 and did not run the process.
I had the same issue. there is an additional step in the installation process which i did not do.
refer to to the url
https://cran.r-project.org/bin/windows/Rtools/
Look for "Putting Rtools on the PATH"
writeLines('PATH="${RTOOLS40_HOME}\usr\bin;${PATH}"', con = "~/.Renviron")
for windows users
wrong: system(path("c:", "program files", "r", "anysoft.EXE"))
but works : system(path("c:", shQuote("program files"), "r", "anysoft.EXE"))
You guys are making it so complicated. I solved this problem by referring to this answer. The problem is with the PATH. type Sys.which('') in R, and you will see nothing. So you have to set the path in CMD, and then use Sys.setenv(PATH = '') in R to get this work.

Calling external program from R with multiple commands in system

I am new to programming and mainly I am able to do some scripts within R, but for my work I need to call an external program. For this program to work on the ubuntu's terminal I have to first use setenv and then execute the program. Googling I've found the system () and Sys.setenv() functions, but unfortunately I can make it function.
This is the code that does work in the ubuntu terminal:
$ export PATH=/home/meme/bin:$PATH
$ mast "/home/meme/meme.txt" "/home/meme/seqs.txt" -o "/home/meme/output" -comp
Where the first two arguments are input files, the -o argument is the output directory and the -comp is another parameter for the program to run.
The reason that I need to do it in R despite it already works in the terminal is because I need to run the program 1000 times with 1000 different files so I want to make a for loop where the input name changes in every loop and then analyze every output in R.
I have already tried to use:
Sys.setenv(PATH="/home/meme/bin"); system(mast "/home/meme/meme.txt" "/home/meme/seqs.txt" -o "/home/meme/output" -comp )
and
system(Sys.setenv(PATH="/home/meme/bin") && mast "/home/meme/meme.txt" "/home/meme/seqs.txt" -o "/home/meme/output" -comp )
but always received:
Error: unexpected constant string in "system(mast "/home/meme/meme.txt""
or
Error: unexpected symbol in "system(Sys.setenv(PATH="/home/meme/bin") && mast "/home/meme/meme.txt""
At this point I have run out of ideas to make this work. If this has already been answered, then my googling have just been poor and I would appreciate any links to its response.
Thank you very much for your time.
Carlos
Additional details:
I use Ubuntu 12.04 64-bits version, RStudio version 0.97.551, R version 3.0.2 (2013-09-25) -- "Frisbee Sailing" Platform: x86_64-pc-linux-gnu (64-bit).
The program I use (MAST) finds a sequence pattern in a list of letters and is part of the MEME SUIT version 4.9.1 found in http://meme.nbcr.net/meme/doc/meme-install.html and run through command line. The command-line usage for mast is:
mast <motif file> <sequence file> [options]
Construct the string you want to execute with paste and feed that to system:
for(i in 1:10){
cmd=paste("export FOO=",i," ; echo \"$FOO\" ",sep='')
system(cmd)
}
Note the use of sep='' to stop paste putting spaces in, and back-quoting quote marks in the string to preserve them.
Test before running by using print(cmd) instead of system(cmd) to make sure you are getting the right command built. Maybe do:
if(TESTING){print(cmd)}else{system(cmd)}
and set TESTING=TRUE or FALSE in R before running.
If you are going to be running more than one shell command per system call, it might be better to put them all in one shell script file and call that instead, passing parameters from R. Something like:
cmd = paste("/home/me/bin/dojob.sh ",i,i+1)
system(cmd)
and then dojob.sh is a shell script that parses the args. You'll need to learn a bit more shell scripting.

Using python to execute .R script

After seven hours of googling and rereading through somewhat similar questions, and then lots of trial and error, I'm now comfortable asking for some guidance.
To simplify my actual task, I created a very basic R script (named test_script):
x <- c(1,2,3,4,5)
avg <- mean(x)
write.csv(avg, file = "output.csv")
This works as expected.
I'm new to python and I'm just trying to figure out how to execute the R script so that the same .csv file is created.
Notable results come from:
subprocess.call(["C:/Program Files/R/R-2.15.2/bin/R", 'C:/Users/matt/Desktop/test_script.R'])
This opens a cmd window with the typical R start-up verbiage, except there is a message which reads, "ARGUMENT 'C:/Users/matt/Desktop/test_script.R' __ ignored __"
And:
subprocess.call(['C:/Program Files/R/R-2.15.2/bin/Rscript', 'C:/Users/matt/Desktop/test_script.r'])
This flashes a cmd window and returns a 0, but no .csv file is created.
Otherwise, I've tried every suggestion I could identify on this site or any other. Any insight will be greatly appreciated. Thanks in advance for your time and efforts.
Running R --help at the command prompt prints:
Usage: R [options] [< infile] [> outfile]
or: R CMD command [arguments]
Start R, a system for statistical computation and graphics, with the
specified options, or invoke an R tool via the 'R CMD' interface.
Options:
-h, --help Print short help message and exit
--version Print version info and exit
...
-f FILE, --file=FILE Take input from 'FILE'
-e EXPR Execute 'EXPR' and exit
FILE may contain spaces but not shell metacharacers.
Commands:
BATCH Run R in batch mode
COMPILE Compile files for use with R
...
Try
call(["C:/Program Files/R/R-2.15.2/bin/R", '-f', 'C:/Users/matt/Desktop/test_script.R'])
There are also some other command-line arguments you can pass to R that may be helpful. Run R --help to see the full list.
It might be too late, but hope it helps for others:
Just add --vanilla in the call list.
subprocess.call(['C:/Program Files/R/R-2.15.2/bin/Rscript', '--vanilla', 'C:/Users/matt/Desktop/test_script.r'])

R and System calls

I have used R in the past to do very basic calls to the commmand line. The example can be found here.
This time around, I am looking to mimic this code which runs successfully from the command line in Windows:
> cd C:\Documents and Settings\BTIBERT\My Documents\My Dropbox\Eclipse\Projects\R\MLB\retrosheet\rawdata
> bgame -y 2010 2010bos.eva >2010bos.txt
This is the code I am trying to run inside of R. I have already set the working directory inside of R.
dir <- paste("cd", getwd(), sep=" ")
system(dir)
system("bgame -y 2010 2010bos.eva >2010bos.txt")
I am sure this is user error, but what am I doing wrong? It appears to work initially, but returns the following error. I very well could be doing something wrong, but I believe I am using the same commands.
Expanded game descriptor, version 109(185) of 05/08/2008.
Type 'bgame -h' for help.
Copyright (c) 2001 by DiamondWare.
[Processing file 2010bos.eva.]
>2010bos.txt: can't open.
Warning message:
running command 'bgame -y 2010 2010bos.eva >2010bos.txt' had status 2
Any help you can provide will be appreciated.
You need to issue all commands in one system() call:
system(paste("cd",getwd() "&& bgame -y 2010 2010bos.eva >2010bos.txt",sep=" "))
You should already be in your working directory, so I'm not sure the cd getwd() is necessary. And you may need quotes around your path because it contains spaces. The error may be resolved by putting spaces around >.
If I were in your shoes, I would try this:
system("bgame -y 2010 2010bos.eva > 2010bos.txt")
UPDATE:
And you should probably heed this advice in the "Differences between Unix and Windows" section of ?system that says you should use shell:
• The most important difference is that on a Unix-alike
‘system’ launches a shell which then runs ‘command’. On
Windows the command is run directly - use ‘shell’ for an
interface which runs ‘command’ _via_ a shell (by default the
Windows shell ‘cmd.exe’, which has many differences from the
POSIX shell).
This means that it cannot be assumed that redirection or
piping will work in ‘system’ (redirection sometimes does, but
we have seen cases where it stopped working after a Windows
security patch), and ‘system2’ (or ‘shell’) must be used on
Windows.
Has no-one else found that system("dir", intern = T) for example doesn't work, but that you need system("cmd.exe /c dir", intern = T)? Only the latter works for me. I found this at the discussion site here (William Dunlap's post, about a third of the way down).
Also, it doesn't work with the "cd" command, but you can use the setwd() function within R and then the command will be executed within that directory.
I created the following functions for convenience, for executing programmes and running commands:
#the subject is an input file that a programme might require
execute <- function(programme, subject.spec = "", intern = FALSE, wait = FALSE){
if(!identical(subject.spec, "")){subject.spec <- paste0(" ", subject.spec)} #put space before the subject if it exists
system(paste0("cmd.exe /c ", programme, subject.spec), intern = intern, wait = wait)
}
command <- function(command, intern = TRUE, wait = FALSE){
system(paste("cmd.exe /c", command), intern = T, wait = wait)
}
Does it break your code when you get error 1 or does execution continue?
Whenever executing system commands through another language it is useful to print the system call before you call it to see exactly what is happening, pull up the shell you are intending to use and check for the same error. As the command is executed correctly this could be a hickup in bgame or R.
If you look at http://astrostatistics.psu.edu/datasets/R/html/base/html/shell.html you can see the variable flag passed to the system call."flag the switch to run a command under the shell. If the shell is bash or tcsh the default is changed to "-c"."
Also "the shell to be used can be changed by setting the configure variable R_SHELL to a suitable value (a full path to a shell, e.g. /usr/local/bin/bash)."

Resources