The hardware and bandwidth for this mirror is donated by METANET, the Webhosting and Full Service-Cloud Provider.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]metanet.ch.
The txtq
package helps parallel R processes send
messages to each other. Let’s say Process A and Process B are working on
a parallel task together. First, both processes grab the queue.
<- tempfile() # Define a path to your queue.
path # In real life, temp files go away when the session exits, so be careful.
path #> [1] "/var/folders/k3/q1f45fsn4_13jbn0742d4zj40000gn/T//RtmpezZ1r3/file6f34c3ca3b4"
<- txtq(path) # Create a new queue or recover an existing one.
q $validate() # Check if the queue is corrupted. q
The queue uses text files to keep track of your data.
list.files(q$path()) # The queue's underlying text files live in this folder.
#> [1] "db" "head" "lock" "total"
$list() # You have not pushed any messages yet.
q#> [1] title message time
#> <0 rows> (or 0-length row.names)
Then, Process A sends instructions to Process B.
$push(title = "Hello", message = "process B.")
q$push(
qtitle = c("Calculate", "Calculate"),
message = c("sqrt(4)", "sqrt(16)")
)$push(title = "Send back", message = "the sum.") q
You can inspect the contents of the queue from either process.
$list()
q#> title message time
#> 1 Hello process B. 2021-01-08 08:16:52.610826 -0500 GMT
#> 2 Calculate sqrt(4) 2021-01-08 08:16:52.625710 -0500 GMT
#> 3 Calculate sqrt(16) 2021-01-08 08:16:52.625710 -0500 GMT
#> 4 Send back the sum. 2021-01-08 08:16:52.628082 -0500 GMT
$list(1) # You can specify the number of messages to list.
q#> title message time
#> 1 Hello process B. 2021-01-08 08:16:52.610826 -0500 GMT
$count()
q#> [1] 4
As Process A is pushing the messages, Process B can consume them.
$pop(2) # If you pass 2, you are assuming the queue has >=2 messages.
q#> title message time
#> 1 Hello process B. 2021-01-08 08:16:52.610826 -0500 GMT
#> 2 Calculate sqrt(4) 2021-01-08 08:16:52.625710 -0500 GMT
Those popped messages are not technically in the queue any longer.
$list()
q#> title message time
#> 1 Calculate sqrt(16) 2021-01-08 08:16:52.625710 -0500 GMT
#> 2 Send back the sum. 2021-01-08 08:16:52.628082 -0500 GMT
$count() # Number of messages technically in the queue.
q#> [1] 2
But we still have a full log of all the messages that were ever sent.
$log()
q#> title message time
#> 1 Hello process B. 2021-01-08 08:16:52.610826 -0500 GMT
#> 2 Calculate sqrt(4) 2021-01-08 08:16:52.625710 -0500 GMT
#> 3 Calculate sqrt(16) 2021-01-08 08:16:52.625710 -0500 GMT
#> 4 Send back the sum. 2021-01-08 08:16:52.628082 -0500 GMT
$total() # Number of messages that were ever queued.
q#> [1] 4
Let’s let Process B get the rest of the instructions.
$pop() # q$pop() with no arguments just pops one message.
q#> title message time
#> 1 Calculate sqrt(16) 2021-01-08 08:16:52.625710 -0500 GMT
$pop() # Call q$pop(-1) to pop all the messages at once.
q#> title message time
#> 1 Send back the sum. 2021-01-08 08:16:52.628082 -0500 GMT
Now let’s say Process B follows the instructions in the messages. The last step is to send the results back to Process A.
$push(title = "Results", message = as.character(sqrt(4) + sqrt(16))) q
Process A can now see the results.
$pop()
q#> title message time
#> 1 Results 6 2021-01-08 08:16:52.690375 -0500 GMT
The queue can grow large if you are not careful. Popped messages are kept in the database file.
$push(title = "not", message = "popped")
q$count()
q#> [1] 1
$total()
q#> [1] 6
$list()
q#> title message time
#> 1 not popped 2021-01-08 08:16:52.708969 -0500 GMT
$log()
q#> title message time
#> 1 Hello process B. 2021-01-08 08:16:52.610826 -0500 GMT
#> 2 Calculate sqrt(4) 2021-01-08 08:16:52.625710 -0500 GMT
#> 3 Calculate sqrt(16) 2021-01-08 08:16:52.625710 -0500 GMT
#> 4 Send back the sum. 2021-01-08 08:16:52.628082 -0500 GMT
#> 5 Results 6 2021-01-08 08:16:52.690375 -0500 GMT
#> 6 not popped 2021-01-08 08:16:52.708969 -0500 GMT
To keep the database file from getting too big, you can clean out the popped messages.
$clean()
q$count()
q#> [1] 1
$total()
q#> [1] 1
$list()
q#> title message time
#> 1 not popped 2021-01-08 08:16:52.737766 -0500 GMT
$log()
q#> title message time
#> 1 not popped 2021-01-08 08:16:52.737766 -0500 GMT
You can also reset the queue to remove all messages, popped or not.
$reset()
q$count()
q#> [1] 0
$total()
q#> [1] 0
$list()
q#> [1] title message time
#> <0 rows> (or 0-length row.names)
$log()
q#> [1] title message time
#> <0 rows> (or 0-length row.names)
When you are done, you can destroy the files in the queue.
$destroy()
qfile.exists(q$path())
#> [1] FALSE
This entire time, the queue was locked when either process was trying to create, access, or modify it. That way, the results stay correct even when multiple processes try to read or change the data at the same time.
You can import a txtq
into another txtq
.
The unpopped messages are grouped together and sorted by timestamp. Same
goes for the popped messages.
<- txtq(tempfile())
q_from <- txtq(tempfile())
q_to $push(title = "from", message = "popped")
q_from$push(title = "from", message = "unpopped")
q_from$push(title = "to", message = "popped")
q_to$push(title = "to", message = "unpopped")
q_to
$pop()
q_from#> title message time
#> 1 from popped 2021-01-08 08:16:52.791970 -0500 GMT
$pop()
q_to#> title message time
#> 1 to popped 2021-01-08 08:16:52.796093 -0500 GMT
$import(q_from)
q_to
$list()
q_to#> title message time
#> 1 from unpopped 2021-01-08 08:16:52.814312 -0500 GMT
#> 2 to unpopped 2021-01-08 08:16:52.814312 -0500 GMT
$log()
q_to#> title message time
#> 1 from popped 2021-01-08 08:16:52.814312 -0500 GMT
#> 2 to popped 2021-01-08 08:16:52.814312 -0500 GMT
#> 3 from unpopped 2021-01-08 08:16:52.814312 -0500 GMT
#> 4 to unpopped 2021-01-08 08:16:52.814312 -0500 GMT
As an interprocess communication tool, txtq
relies on
the filelock
package to prevent race conditions. Unfortunately, filelock
cannot prevent race conditions on network file systems (NFS), which
means neither can txtq
. In other words, on certain common
kinds of clusters, txtq
cannot reliably manage interprocess
communication for processes on different computers. However, it can
still serve as a low-tech replacement for a simple non-threadsafe
database.
Gábor Csárdi’s liteq
package
offers essentially the same functionality implemented with SQLite. It
has a some additional features, such as the ability to detect crashed
workers and re-queue failed messages, but it was in an early stage of
development at the time txtq
was released.
There is a plethora of message queues beyond R, most notably ZeroMQ and RabbitMQ. In fact, Jeroen Ooms and Whit Armstrong maintain rzmq
, a package
to work with ZeroMQ from R. Even in
this landscape, txtq
has advantages.
txtq
user interface is friendly, and its internals
are simple. No prior knowledge of sockets or message-passing is
required.txtq
is lightweight, R-focused, and easy to install. It
only depends on R and a few packages on CRAN.txtq
it is file-based,
q$log()
and q$list()
.q$list()
while your work is running.These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.