jobstatus
lets you pass live progress, status, and other information
between functions and processes in R, so that you can keep an eye on how
complex and long-running jobs are progressing. jobstatus
uses the
future
package so you can
even get live progress information back from jobs running in parallel.
jobstatus
passes status information between functions and processes in
R so you can monitor what’s happening and how much progress the code has
made. There are three main functions in jobstatus
:
jobstatus
creates a jobstatus object, which lets you record your
progress on a single taskwith_jobstatus
runs a chunk of code that runs other jobs, and
keeps track of the status informationsubjob_future
is a wrapper function for future::future
that lets
you write code that can be executed either in sequence or in
parallel.with_jobstatus
can also be used to display the progress information,
for example with a progress bar.
A jobstatus
set up might look something like this:
library (jobstatus)
library (progress)
# a function to run a single task
some_big_task <- function (iterations = 100) {
# create a jobstatus object to track progress
# do the work, incrementing the progress counter each time
status <- jobstatus$new(iterations)
result <- 0
for (i in seq_len(iterations)) {
result <- result + 1
Sys.sleep(runif(1, 0, 0.2))
status$tick()
}
# tidy up when we're done, and return the results
status$finish()
result
}
# load the future package and set it to run the jobs in parallel
library (future)
plan(multiprocess)
# dispatch some jobs, tracking their status information and displaying multiple progress bars
with_jobstatus({
# create some futures (possibly in parallel)
f1 <- subjob_future(some_big_task())
f2 <- subjob_future(some_big_task())
f3 <- subjob_future(some_big_task())
f4 <- subjob_future(some_big_task())
# get their values
v1 <- value(f1)
v2 <- value(f2)
v3 <- value(f3)
v4 <- value(f4)
}, display = percentage)
This is very much a work in progress and the current implementation is quite limited (and probably quite buggy). Ideally, we’d like it to support various different parallel backends and interfaces, handle other types of job status information, and provide different types of progress bars and displays.
This prototype was developed over two days at the 2018 rOpenSci unconference and the maintainers won’t have much time to extend and improve the package. We’d love to have help, so if you’re keen please let us know in the issues tracker!
You can install the development version from GitHub with:
# install.packages("devtools")
devtools::install_github("ropenscilabs/jobstatus")
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.