This package supplies a progress bar (shamelessly borrowed from dplyr::progress_estimated
) that has options to not have output captured inside a knitr
chunk.
You want to use knitr
or rmarkdown
, but you want to see the progress of a longer running calculation in the chunk. You think that you can just use dplyr::progress_estimated
. But if you do, all the output from the progress bar will be suppressed (by design, actually).
This package has two functions, progress_estimated
, that creates a Progress
object that has a connection object associated with it, and update_progress
, that properly updates the progress object. The output from the progress will be written to that connection. This connection will be either stdout
(default within an R session), stderr
(default from within knitr
), or to a log-file.
None of these are run in this document!
library(knitrProgressBar)
# borrowed from example by @hrbrmstr
arduously_long_nchar <- function(input_var, .pb=NULL) {
update_progress(.pb)
Sys.sleep(0.5)
nchar(input_var)
}
If you want the object to decide where to put output, do nothing. Just call the progress_estimated()
function, which uses make_kpb_output_decisions()
:
pb <- progress_estimated(length(letters))
purrr::map_int(letters, arduously_long_nchar, .pb = pb)
See the help and the vignette for an explanation of how make_kpb_output_decisions()
decides where to display the progress bar output.
If you want to write the progress out to a specific connection, just pass the connection to the progress_estimated()
call:
pb <- progress_estimated(length(letters), progress_location = stdout())
purrr::map_int(letters, arduously_long_nchar, .pb = pb)
This includes specific files. You can then display the file, or use tailf
or equivalent to watch the output of the file.
pb <- progress_estimated(length(letters), progress_location = file("progress.log", open = "w"))
purrr::map_int(letters, arduously_long_nchar, .pb = pb)
Each connection will display in specific situations, notably stdout()
will not display to the terminal when run as part of a document being knit
ted.
This package (and the examples) was inspired by this post from Bob Rudis! Also, thanks to Hadley Wickham for the great Progress
object and methods!
Web accessible documentation is available here.
Please submit bug reports using the GitHub issue tracker.
Please note that this project is released with a Contributor Code of Conduct. By participating in this project you agree to abide by its terms.
This package is licensed using an MIT license, copyright Robert M Flight.