I am trying to convert an R script into something that a client can run in batch mode. My code uses generic functions and one snippet which is near the beginning goes like:
setGeneric("testInput", function(inputData, params = list())
standardGeneric("testInput"))
I've been using R CMD BATCH and it works fine. However I couldn't find an easy way to make my script print the output on the console, so based on that (and suggestion that Rscript.exe is the "proper" way to run R batch files) I decided to switch to Rscript. However when running the very same .R file with Rscript I get the following:
Error: could not find function "setGeneric"
Execution halted
I know there is probably a trivial reason behind this but I just cannot figure it out. Can someone please point me to where the mistake is?Any suggestions?
setGeneric is part of the methods package which is usually loaded when you start R in an interactive session but not in non interactive session using Rscript or littler.
So you need to add a require(methods) before calling setGeneric in your script.
For example, this code will not work
Rscript -e "setGeneric('mean', function(x) standardGeneric('mean'))"
Error: could not find function "setGeneric"
Execution halted
But this one will work
Rscript -e "require(methods);setGeneric('mean', function(x) standardGeneric('mean'))"
Loading required package: methods
[1] "mean"
Related
Running my working R script in the windows command line (cmd) using Rscript results in a parsing error (premature EOF).
When I run the script in RStudio, it compiles and runs as expected.
I have read the Rscript page in R documentation, and I see that the problem must be due to spaces in my script itself, which probably make it into the cmd console somehow during parsing, but that's as far as I get.
Or should I have done something with the #! functionality mentioned therein?
I am trying to run it on cmd:
Rscript .\start_app.r
I am in the right working directory, and have set the folder containing Rscript in my environment.
The script is too long to share, and I am too inexperienced to give you the parts that make it break (otherwise I wouldn't be here), but it is full of functions, if statements and the like, that use curly brackets and are indented. I also often include empty rows (someteimes indented) for readability. It makes use of the shiny-package. An example could be:
islocal = nchar(Sys.getenv("LOCAL"))>1 | interactive()
if (islocal){
source('../../path/app/variables/styling.R')
} else {
source('./variables/styling.R')
}
As the example above, it also includes other R code called via source()
Can that somehow make it to the cmd line and be incorrectly compiled?
I get the following messages:
Error: parse error: premature EOF
(right here) ------^
Execution halted
Not enough memory resources are available to process this command.
(I guess the second message is an unrelated issue, but include it here just to be sure.)
As suggested in a comment, the solution was changing the encoding.
As mentionned by the requestor himself, Using "Save with Encoding -> ISO-8895-1 (System default)" solves the issue.
I'm writing a bash script (to be called from the terminal -- on a Linux system) that creates a log-file prior to initiating an 'rscript' using some simple user input. I'm however running into problems in controlling what messages are included in the log file (or sent to the terminal), and can't find any solution for excluding one specific R package-load message:
Package WGCNA 1.66 loaded.
In other words I need a way to (only) silence this specific message, which is printed when the WGCNA package is successfully loaded.
I will try to keep the code non-specific to hopefully make it easier to follow.
The below block is a skeleton (excluding some irrelevant code), and will be followed by some different variants I've tried. Where, originally i tried controlling the output from the R script using sink() and suppressPackageStartupMessages(), which - I thought - should have been enough.
bash script:
#!/usr/bin/env bash
read RDS
DATE=`date +%F-%R`
LOG=~/path/log/$DATE.log
touch $LOG
export ALLOW_WGCNA_THREADS=4
Rscript ~/path/analysis.R $RDS $DATE $LOG
R script:
#!/usr/bin R
# object set-up
rds.path <- "~/path/data/"
temp.path <- "~/path/temp/"
pp.data <- readRDS(paste0(rds.path, commandArgs(T)[1]))
file.date <- paste0(commandArgs(T)[2], "_")
# set up error logging
log.file <- file(commandArgs(T)[3], open="a")
sink(log.file, append=TRUE, type="message")
sink(log.file, append=TRUE, type="output")
# main pkg call
if(suppressPackageStartupMessages(!require(thePKG))){
stop("\nPlease follow the below link to install the requested package (thePKG) with relevant dependencies\n https://link.address")
}
# thePKG method call
cat("> Running "method"\n", append=TRUE)
module <- method(thePKG_input = pp.data, ppi_network = ppi_network)
# reset sink and close file connection
sink(type="message")
sink(type="output")
close(log.file)
This doesn't output anything to the terminal (which is good), including the following in the log-file:
Package WGCNA 1.66 loaded.
> Running "method"
Error: added error to verify that it's correctly printed to the file
I want to keep my log files as clean and on point as possible (and avoid clutter in the terminal), and therefor wish to remove the package-load message. I've tried the following...
i. Omitting the sink() call from the R script, and adding
&>> $LOG
to the bash 'Rscript' call. Resulting in the same file output as above.
ii. Same as i but substituting suppressPackageStartupMessages() with suppressMessages(), which results in the same file output as above.
iii. Same as i but added
... require(thePKG, quietly=TRUE)
in the R script # main pkg call, with the same results.
These where the potential solutions I came across, and tried in different variations with no positive results.
I also wondered if the WGCNA package was loaded "outside" of the !require-loop of thePKG, since it's not affected by suppressMessages() for that call. But, introducing an intentional error (which terminated the process) prior to the if-require(thePKG)-call removed the message -- hinting at its initiation inside the loop.
I also tried calling WGCNA by itself at the start of the R script, with an added suppressMessages() to it, but that didn't work either.
The export function used in the bash script doesn't effect the outcome (to my knowledge) and removing it extends the load message to include the following (truncated to save space):
Package WGCNA 1.66 loaded.
Important note: It appears that your system supports multi-threading, but it is not enabled within WGCNA in R.
To allow multi-threading within WGCNA with all available cores, use allowWGCNAThreads() within R. Use disableWGCNAThreads() to disable threading if necessary.
(...)
I'm aware that I could send the output (only) to /dev/null, but there's other output printed to the file (e.g. > Running "method") that I still want.
Does anyone have a suggestion for how to remove this message? I'm very new to programming in R, and just started using Linux (Ubuntu LTS), so please keep that in mind when answering.
Thanks in advance.
I could gather the following execution flow from your question:
Bash_Script ----> R_Script ----> Output_to_terminal (and also a log file probably)
For the requirements stated, following commands (in bash) could help you:
grep - This command helps you search for patterns
Pipe (|) - This is an operator in Linux which helps you redirect output of one command as input to another for further processing (For eg. A | B )
tee - This command takes any input and forwards it to a file + standard output (terminal)
So, you could mix and match the above commands in your bash file to get the desired output:
You can extend
Rscript ~/path/analysis.R $RDS $DATE $LOG
with
Rscript ~/path/analysis.R "$RDS" "$DATE" "$LOG" | grep -v "Package WGCNA" | tee "output.log"
What each command does ?
a) | -> This redirects output of preceding command as input to the next. It's more of a connector
b) grep -> Normally used to search for patterns. In this example, we use it with -v option to do a invert-match instead i.e. search for everything except Package WGCNA
c) tee -> This writes output to terminal + also redirects the same to a log file name, passed as argument. You could skip this command altogether, if not needed
I'm using the R package 'here' to define my working directory using the following command at the start of my script:
here::set_here(path='/path/to/my_directory/', verbose = F)
Every time I run the script it prints this to the console:
here() starts at /path/to/my_directory
Is there a way to suppress this output? I tried using the invisible() function but that didn't work...
The message you’re seeing is printed when you’re attaching the ‹here› package. Simply don’t do that (it’s unnecessary anyway) to prevent it.
Otherwise, load it as follows:
suppressPackageStartupMessages(library(here))
… yeah, not exactly elegant.
How can I call a sconstruct script from R (e.g. in Rstudio)? I'd like to call SCons and, ideally, also read the output. So that I can get the print from e.g. scons --tree=all as a string.
If I run > system("scons")
I get: sh: scons: command not found
Setting the path Sys.setenv(PATH=paste(Sys.getenv("PATH"), "/path/to/my/sconstruct", sep=":")) doesn't help.
However, any other command works. E.g. if I have a Python script (in the same directory), I can call it as: > system('python test.py')
and get the expected: Hello Rld! back. >system('ls') lists the SConstruct, so I'm in the right work directory.
Calling from a Python script also works, eg: from subprocess import call call('scons') evokes the SConstruct as expected. However, calling that Python script from R -- doesn't work.
It appears as there is something in the R environment setting that I got wrong.
I'm on OS, but a portable solution would be fantastic!
You'll need to either be in the directory where the SConstruct is, specify it's file or use -C to change directory to it.
I don't know R or what the syntax is.
Sys.setenv(PATH=paste(Sys.getenv("PATH"), "/path/to/my/sconstruct", sep=":"))
Should likely be:
Sys.setenv(PATH=paste(Sys.getenv("PATH"), "/path/to/scons", sep=":"))
I am developing a package that exposes an R interface (a bunch of functions to be used interactively) and a command line interface via Rscript. This second one works via a small launcher, for instance, at the command line:
Rscript mylauncher.R arg1 arg2 arg3
would call a function of my package.
I would like to test a couple of command lines from R. Nothing fancy, just make sure that everything runs without errors.
If I test these calls doing in an R source file:
system("Rscript mylauncher.R arg1 arg2 arg3")
How can I be sure that I called the right Rscript? In case there are multiple R installations? (which is actually the case in my setting).
Another approach would be write in the R source file:
source("mylauncher.R")
But I don't see how to specify the command line (I would avoid the trick of overwriting the function commandArgs, because I want to test also the right tokenization of the command line). Does anybody have an idea?
Thanks!
Regarding
How can I be sure that I called the right Rscript? In case there are
multiple R installations?
you would query R RHOME on the command-line and Sys.getenv("R_HOME") from wihthin R.
You then append bin/RScript and should have the Rscript corresponding to your current session. I still design my libraries in such a way that I can call them from R ...