Documentation

callr

Call R from R

Linux Build Status Windows Build status CRAN RStudio mirror downloads Coverage Status

It is sometimes useful to perform a computation in a separate R process, without affecting the current R process at all. This packages does exactly that.


Features

  • Calls an R function, with arguments, in a subprocess.
  • Copies function arguments to the subprocess and copies the return value of the function back, seamlessly.
  • Copies error objects back from the subprocess, including a stack trace.
  • Shows and/or collects the standard output and standard error of the subprocess.
  • Supports both one-off and persistent R subprocesses.
  • Calls the function synchronously or asynchronously (in the background).
  • Can call R CMD commands, synchronously or asynchronously.
  • Can call R scripts, synchronously or asynchronously.
  • Provides extensible r_process, rcmd_process and rscript_process R6 classes, based on processx::process.

Installation

Install the stable version from CRAN:

install.packages("callr")

Synchronous, one-off R processes

Use r() to run an R function in a new R process. The results are passed back seamlessly:

library(callr)
r(function() var(iris[, 1:4]))
#>              Sepal.Length Sepal.Width Petal.Length Petal.Width
#> Sepal.Length    0.6856935  -0.0424340    1.2743154   0.5162707
#> Sepal.Width    -0.0424340   0.1899794   -0.3296564  -0.1216394
#> Petal.Length    1.2743154  -0.3296564    3.1162779   1.2956094
#> Petal.Width     0.5162707  -0.1216394    1.2956094   0.5810063

Passing arguments

You can pass arguments to the function by setting args to the list of arguments. This is often necessary as these arguments are explicitly copied to the child process, whereas the evaluated function cannot refer to variables in the parent. For example, the following does not work:

mycars <- cars
r(function() summary(mycars))
#> Error: callr subprocess failed: object 'mycars' not found

But this does:

r(function(x) summary(x), args = list(mycars))
#>      speed           dist       
#>  Min.   : 4.0   Min.   :  2.00  
#>  1st Qu.:12.0   1st Qu.: 26.00  
#>  Median :15.0   Median : 36.00  
#>  Mean   :15.4   Mean   : 42.98  
#>  3rd Qu.:19.0   3rd Qu.: 56.00  
#>  Max.   :25.0   Max.   :120.00

Note that the arguments will be serialized and saved to a file, so if they are large R objects, it might take a long time for the child process to start up.

Using packages

You can use any R package in the child process, just make sure to refer to it explicitly with the :: operator. For example, the following code creates an igraph graph in the child, and calculates some metrics of it.

r(function() { g <- igraph::sample_gnp(1000, 4/1000); igraph::diameter(g) })
#> [1] 11

Error handling

callr` copies errors from the child process back to the main R session:

r(function() 1 + "A")
#> Error: callr subprocess failed: non-numeric argument to binary operator

callr sets the .Last.error variable, and after an error you can inspect this for more details about the error, including stack traces both from the main R process and the subprocess.

.Last.error
#> <callr_status_error: callr subprocess failed: non-numeric argument to binary operator>
#> -->
#> <callr_remote_error in 1 + "A": non-numeric argument to binary operator>

The error objects has two parts. The first belongs to the main process, and the second belongs to the subprocess.

.Last.error also includes a stack trace, that includes both the main R process and the subprocess:

.Last.error.trace
#> 
#>  ERROR TRACE for callr_status_error, callr_error, rlib_error
#> 
#>  Process 31908:
#>  38. callr:::r(function() 1 + "A")
#>  39. callr:::get_result(output = out, options)
#>     R/eval.R:149:3
#>  40. callr:::throw(new_callr_error(output, msg), parent = err[[2]])
#>     R/result.R:73:5
#> 
#>  x callr subprocess failed: non-numeric argument to binary operator 
#> 
#>  Process 32006:
#>  52. (function ()  ...
#>  53. base:::.handleSimpleError(function (e)  ...
#>     R/<text>:1:3
#>  54. h(simpleError(msg, call))
#> 
#>  x non-numeric argument to binary operator

The top part of the trace contains the frames in the main process, and the bottom part contains the frames in the subprocess, starting with the anonymous function.

Standard output and error

By default, the standard output and error of the child is lost, but you can request callr to redirect them to files, and then inspect the files in the parent:

x <- r(function() { print("hello world!"); message("hello again!") },
  stdout = "/tmp/out", stderr = "/tmp/err"
)
readLines("/tmp/out")
#> [1] "[1] \"hello world!\""
readLines("/tmp/err")
#> [1] "hello again!"

With the stdout option, the standard output is collected and can be examined once the child process finished. The show = TRUE options will also show the output of the child, as it is printed, on the console of the parent.

Background R processes

r_bg() is similar to r() but it starts the R process in the background. It returns an r_process R6 object, that provides a rich API:

rp <- r_bg(function() Sys.sleep(.2))
rp
#> PROCESS 'R', running, pid 32028.

This is a list of all r_process methods:

ls(rp)
#>  [1] "as_ps_handle"          "clone"                
#>  [3] "finalize"              "format"               
#>  [5] "get_cmdline"           "get_cpu_times"        
#>  [7] "get_error_connection"  "get_error_file"       
#>  [9] "get_exe"               "get_exit_status"      
#> [11] "get_input_connection"  "get_input_file"       
#> [13] "get_memory_info"       "get_name"             
#> [15] "get_output_connection" "get_output_file"      
#> [17] "get_pid"               "get_poll_connection"  
#> [19] "get_result"            "get_start_time"       
#> [21] "get_status"            "get_username"         
#> [23] "get_wd"                "has_error_connection" 
#> [25] "has_input_connection"  "has_output_connection"
#> [27] "has_poll_connection"   "initialize"           
#> [29] "interrupt"             "is_alive"             
#> [31] "is_incomplete_error"   "is_incomplete_output" 
#> [33] "is_supervised"         "kill"                 
#> [35] "kill_tree"             "poll_io"              
#> [37] "print"                 "read_all_error"       
#> [39] "read_all_error_lines"  "read_all_output"      
#> [41] "read_all_output_lines" "read_error"           
#> [43] "read_error_lines"      "read_output"          
#> [45] "read_output_lines"     "resume"               
#> [47] "signal"                "supervise"            
#> [49] "suspend"               "wait"                 
#> [51] "write_input"

These include all methods of the processx::process superclass and the new get_result() method, to retrieve the R object returned by the function call. Some of the handiest methods are:

  • get_exit_status() to query the exit status of a finished process.
  • get_result() to collect the return value of the R function call.
  • intertupt() to send an interrupt to the process. This is equivalent to a CTRL+C key press, and the R process might ignore it.
  • is_alive() to check if the process is alive.
  • kill() to terminate the process.
  • poll_io() to wait for any standard output, standard error, or the completion of the process, with a timeout.
  • read_*() to read the standard output or error.
  • suspend() and resume() to stop and continue a process.
  • wait() to wait for the completion of the process, with a timeout.

Multiple background R processes and poll()

Multiple background R processes are best managed with the processx::poll() function that waits for events (standard output/error or termination) from multiple processes. It returns as soon as one process has generated an event, or if its timeout has expired. The timeout is in milliseconds.

rp1 <- r_bg(function() { Sys.sleep(1/2); "1 done" })
rp2 <- r_bg(function() { Sys.sleep(1/1000); "2 done" })
processx::poll(list(rp1, rp2), 1000)
#> [[1]]
#>   output    error  process 
#> "silent" "silent" "silent" 
#> 
#> [[2]]
#>   output    error  process 
#> "silent" "silent"  "ready"
rp2$get_result()
#> [1] "2 done"
processx::poll(list(rp1), 1000)
#> [[1]]
#>   output    error  process 
#> "silent" "silent"  "ready"
rp1$get_result()
#> [1] "1 done"

Persistent R sessions

r_session is another processx::process subclass that represents a persistent background R session:

rs <- r_session$new()
rs
#> R SESSION, alive, idle, pid 32095.

r_session$run() is a synchronous call, that works similarly to r(), but uses the persistent session. r_session$call() starts the function call and returns immediately. The r_session$poll_process() method or processx::poll() can then be used to wait for the completion or other events from one or more R sessions, R processes or other processx::process objects.

Once an R session is done with an asynchronous computation, its poll_process() method returns "ready" and the r_session$read() method can read out the result.

rs$run(function() runif(10))
#>  [1] 0.7896924 0.1770236 0.9931798 0.2850374 0.8582676 0.4489599 0.1556783
#>  [8] 0.4240582 0.4260408 0.5561292
rs$call(function() rnorm(10))
rs
#> R SESSION, alive, busy, pid 32095.
rs$poll_process(2000)
#> [1] "ready"
rs$read()
#> $code
#> [1] 200
#> 
#> $message
#> [1] "done file7ca436de78be"
#> 
#> $result
#>  [1] -1.0339966 -0.7015408  1.2220644 -0.9691123 -0.6991386  1.2698753
#>  [7]  0.2595250  1.0871667  0.5148076  0.4015015
#> 
#> $stdout
#> [1] ""
#> 
#> $stderr
#> [1] ""
#> 
#> $error
#> NULL
#> 
#> attr(,"class")
#> [1] "callr_session_result"

Running R CMD commands

The rcmd() function calls an R CMD command. For example, you can call R CMD INSTALL, R CMD check or R CMD config this way:

rcmd("config", "CC")
#> $status
#> [1] 0
#> 
#> $stdout
#> [1] "clang\n"
#> 
#> $stderr
#> [1] ""
#> 
#> $timeout
#> [1] FALSE
#> 
#> $command
#> [1] "/Library/Frameworks/R.framework/Resources/bin/R"
#> [2] "CMD"                                            
#> [3] "config"                                         
#> [4] "CC"
#>$stdout
#>[1] "clang\n"
#>
#>$stderr
#>[1] ""
#>
#>$status
#>[1] 0

This returns a list with three components: the standard output, the standard error, and the exit (status) code of the R CMD command.

License

MIT © Mango Solutions, RStudio