Skip to contents

Task queues for R, implemented using Redis.

Getting started

Create an rrq_controller object

obj <- rrq_controller("rrq:readme")
rrq_default_controller_set(obj)

Submit work to the queue:

t <- rrq_task_create_expr(runif(10))
t
#> [1] "fd08ef462c509e174338ca5920b13adf"

Query task process:

rrq_task_status(t)
#> [1] "PENDING"

Run tasks on workers in the background

rrq_worker_spawn()
#> ℹ Spawning 1 worker with prefix 'nonmetalliferous_jabiru'
#> <rrq_worker_manager>
#>   Public:
#>     clone: function (deep = FALSE)
#>     id: nonmetalliferous_jabiru_1
#>     initialize: function (controller, n, logdir = NULL, name_config = "localhost",
#>     is_alive: function (worker_id = NULL)
#>     kill: function (worker_id = NULL)
#>     logs: function (worker_id)
#>     stop: function (worker_id = NULL, ...)
#>     wait_alive: function (timeout, time_poll = 0.2, progress = NULL)
#>   Private:
#>     check_worker_id: function (worker_id)
#>     controller: rrq_controller
#>     logfile: /tmp/Rtmp4AMYNM/filee336a6e3e1369/nonmetalliferous_jabiru_1
#>     process: list
#>     worker_id_base: nonmetalliferous_jabiru

Wait for tasks to complete

rrq_task_wait(t)
#> [1] TRUE

Retrieve results from a task

rrq_task_result(t)
#>  [1] 0.044105073 0.151111529 0.047623996 0.936703515 0.719949653 0.519903127
#>  [7] 0.030698510 0.057218178 0.503331142 0.002863957

Query what workers have done

rrq_worker_log_tail(n = Inf)
#>                   worker_id child       time       command
#> 1 nonmetalliferous_jabiru_1    NA 1721033895         ALIVE
#> 2 nonmetalliferous_jabiru_1    NA 1721033895         ENVIR
#> 3 nonmetalliferous_jabiru_1    NA 1721033895         QUEUE
#> 4 nonmetalliferous_jabiru_1    NA 1721033895    TASK_START
#> 5 nonmetalliferous_jabiru_1    NA 1721033895 TASK_COMPLETE
#>                            message
#> 1
#> 2                              new
#> 3                          default
#> 4 fd08ef462c509e174338ca5920b13adf
#> 5 fd08ef462c509e174338ca5920b13adf

For more information, see vignette("rrq")

Installation

Install from the mrc-ide R universe package repository:

install.packages(
  "rrq",
  repos = c("https://mrc-ide.r-universe.dev", "https://cloud.r-project.org"))

Alternatively, install with remotes:

remotes::install_github("mrc-ide/rrq", upgrade = FALSE)

Testing

To test, we need a redis server that can be automatically connected to using the redux defaults. This is satisfied if you have an unauthenticated redis server running on localhost, otherwise you should update the environment variable REDIS_URL to point at a redis server. Do not use a production server, as the package will create and delete a lot of keys.

A suitable redis server can be started using docker with

./scripts/redis start

(and stopped with ./scripts/redis stop)

Alternatively, on Windows, a native (i.e., not depending on CygWin, MSys, or WSL) port of Redis 5.0.14.1 can be installed from here, and will run out of the box.

Testing the pkgdown site with examples

The documentation includes many executed code examples. To preview the documentation with the results of these, you’ll need a Redis server running, and the rrq package installed. Then, in an R terminal:-

rrq::rrq_worker$new("rrq:example")$loop()

to run a worker. Then in another R terminal,

pkgdown::build_site()

and the pkgdown site will be built in the docs/ folder.

License

MIT © Imperial College of Science, Technology and Medicine