Package 'mirai' reference manual

Title:	Minimalist Async Evaluation Framework for R
Description:	Designed for simplicity, a 'mirai' evaluates an R expression asynchronously in a parallel process, locally or distributed over the network, with the result automatically available upon completion. Modern networking and concurrency built on 'nanonext' and 'NNG' (Nanomsg Next Gen) ensure reliable and efficient scheduling, over fast inter-process communications or TCP/IP secured by TLS. Advantages include being inherently queued thus handling many more tasks than available processes, no storage on the file system, support for otherwise non-exportable reference objects, an event-driven promises implementation, and built-in asynchronous parallel map.
Authors:	Charlie Gao [aut, cre] , Joe Cheng [ctb], Hibiki AI Limited [cph]
Maintainer:	Charlie Gao <[email protected]>
License:	GPL (>= 3)
Version:	1.3.1.9000
Built:	2024-11-18 15:24:10 UTC
Source:	https://github.com/shikokuchuo/mirai

mirai: Minimalist Async Evaluation Framework for R

Description

Designed for simplicity, a 'mirai' evaluates an R expression asynchronously in a parallel process, locally or distributed over the network, with the result automatically available upon completion. Modern networking and concurrency built on 'nanonext' and 'NNG' (Nanomsg Next Gen) ensure reliable and efficient scheduling, over fast inter-process communications or TCP/IP secured by TLS. Advantages include being inherently queued thus handling many more tasks than available processes, no storage on the file system, support for otherwise non-exportable reference objects, an event-driven promises implementation, and built-in asynchronous parallel map.

Notes

For local mirai requests, the default transport for inter-process communications is platform-dependent: abstract Unix domain sockets on Linux, Unix domain sockets on MacOS, Solaris and other POSIX platforms, and named pipes on Windows.

This may be overriden, if desired, by specifying 'url' in the daemons interface and launching daemons using launch_local.

Reference Manual

vignette("mirai", package = "mirai")

Author(s)

Charlie Gao [email protected] (ORCID)

Make Mirai Promise

Description

Creates a ‘promise’ from a ‘mirai’.

Usage

## S3 method for class 'mirai'
as.promise(x)
## S3 method for class 'mirai'
as.promise(x)

Arguments

`x`	an object of class ‘mirai’.

Details

This function is an S3 method for the generic as.promise for class ‘mirai’.

Requires the promises package.

Allows a ‘mirai’ to be used with the promise pipe %...>%, which schedules a function to run upon resolution of the ‘mirai’.

Value

A ‘promise’ object.

Examples

if (interactive() && requireNamespace("promises", quietly = TRUE)) {

library(promises)

p <- as.promise(mirai("example"))
print(p)
is.promise(p)

p2 <- mirai("completed") %...>% identity()
p2$then(cat)
is.promise(p2)

}

if (interactive() && requireNamespace("promises", quietly = TRUE)) {

library(promises)

p <- as.promise(mirai("example"))
print(p)
is.promise(p)

p2 <- mirai("completed") %...>% identity()
p2$then(cat)
is.promise(p2)

}

mirai (Call Value)

Description

call_mirai waits for the ‘mirai’ to resolve if still in progress, storing the value at $data, and returns the ‘mirai’ object.

call_mirai_ is a variant of call_mirai that allows user interrupts, suitable for interactive use.

Usage

call_mirai(x)

call_mirai_(x)
call_mirai(x)

call_mirai_(x)

Arguments

`x`	a ‘mirai’ object, or list of ‘mirai’ objects.

Details

Both functions accept a list of ‘mirai’ objects, such as that returned by mirai_map as well as individual ‘mirai’.

They will wait for the asynchronous operation(s) to complete if still in progress (blocking).

x[] may also be used to wait for and return the value of a mirai x, and is the equivalent of call_mirai_(x)$data.

Value

The passed object (invisibly). For a ‘mirai’, the retrieved value is stored at $data.

Alternatively

The value of a ‘mirai’ may be accessed at any time at $data, and if yet to resolve, an ‘unresolved’ logical NA will be returned instead.

Using unresolved on a ‘mirai’ returns TRUE only if it has yet to resolve and FALSE otherwise. This is suitable for use in control flow statements such as while or if.

Errors

If an error occurs in evaluation, the error message is returned as a character string of class ‘miraiError’ and ‘errorValue’ (the stack trace is available at $stack.trace on the error object). is_mirai_error may be used to test for this.

If a daemon crashes or terminates unexpectedly during evaluation, an ‘errorValue’ 19 (Connection reset) is returned (when not using dispatcher or using dispatcher with retry = FALSE). Otherwise, using dispatcher with retry = TRUE, the mirai will remain unresolved and is automatically re-tried on the next daemon to connect to the particular instance. To cancel the task instead, use saisei(force = TRUE) (see saisei).

is_error_value tests for all error conditions including ‘mirai’ errors, interrupts, and timeouts.

Examples

if (interactive()) {
# Only run examples in interactive R sessions

# using call_mirai()
df1 <- data.frame(a = 1, b = 2)
df2 <- data.frame(a = 3, b = 1)
m <- mirai(as.matrix(rbind(df1, df2)), df1 = df1, df2 = df2, .timeout = 1000)
call_mirai(m)$data

# using unresolved()
m <- mirai(
  {
    res <- rnorm(n)
    res / rev(res)
  },
  n = 1e6
)
while (unresolved(m)) {
  cat("unresolved\n")
  Sys.sleep(0.1)
}
str(m$data)

}

if (interactive()) {
# Only run examples in interactive R sessions

# using call_mirai()
df1 <- data.frame(a = 1, b = 2)
df2 <- data.frame(a = 3, b = 1)
m <- mirai(as.matrix(rbind(df1, df2)), df1 = df1, df2 = df2, .timeout = 1000)
call_mirai(m)$data

# using unresolved()
m <- mirai(
  {
    res <- rnorm(n)
    res / rev(res)
  },
  n = 1e6
)
while (unresolved(m)) {
  cat("unresolved\n")
  Sys.sleep(0.1)
}
str(m$data)

}

mirai (Collect Value)

Description

collect_mirai waits for the ‘mirai’ to resolve if still in progress, and returns its value directly. It is a more efifcient version of and equivalent to call_mirai(x)$data.

Usage

collect_mirai(x)
collect_mirai(x)

Arguments

`x`	a ‘mirai’ object, or list of ‘mirai’ objects.

Details

This function will wait for the asynchronous operation(s) to complete if still in progress (blocking), and is not interruptible.

x[] may be used to wait for and return the value of a mirai x, and is the user-interruptible counterpart to collect_mirai(x).

Value

An object (the return value of the ‘mirai’), or a list of such objects (the same length as ‘x’, preserving names).

Alternatively

The value of a ‘mirai’ may be accessed at any time at $data, and if yet to resolve, an ‘unresolved’ logical NA will be returned instead.

Using unresolved on a ‘mirai’ returns TRUE only if it has yet to resolve and FALSE otherwise. This is suitable for use in control flow statements such as while or if.

Errors

is_error_value tests for all error conditions including ‘mirai’ errors, interrupts, and timeouts.

Examples

if (interactive()) {
# Only run examples in interactive R sessions

# using collect_mirai()
df1 <- data.frame(a = 1, b = 2)
df2 <- data.frame(a = 3, b = 1)
m <- mirai(as.matrix(rbind(df1, df2)), df1 = df1, df2 = df2, .timeout = 1000)
collect_mirai(m)

# using x[]
m[]

}

if (interactive()) {
# Only run examples in interactive R sessions

# using collect_mirai()
df1 <- data.frame(a = 1, b = 2)
df2 <- data.frame(a = 3, b = 1)
m <- mirai(as.matrix(rbind(df1, df2)), df1 = df1, df2 = df2, .timeout = 1000)
collect_mirai(m)

# using x[]
m[]

}

Daemon Instance

Description

Starts up an execution daemon to receive mirai requests. Awaits data, evaluates an expression in an environment containing the supplied data, and returns the value to the host caller. Daemon settings may be controlled by daemons and this function should not need to be invoked directly, unless deploying manually on remote resources.

Usage

daemon(
  url,
  asyncdial = FALSE,
  autoexit = TRUE,
  cleanup = TRUE,
  output = FALSE,
  maxtasks = Inf,
  idletime = Inf,
  walltime = Inf,
  timerstart = 0L,
  ...,
  tls = NULL,
  rs = NULL
)
daemon(
  url,
  asyncdial = FALSE,
  autoexit = TRUE,
  cleanup = TRUE,
  output = FALSE,
  maxtasks = Inf,
  idletime = Inf,
  walltime = Inf,
  timerstart = 0L,
  ...,
  tls = NULL,
  rs = NULL
)

Arguments

`url`	the character host or dispatcher URL to dial into, including the port to connect to (and optionally for websockets, a path), e.g. 'tcp://hostname:5555' or 'ws://10.75.32.70:5555/path'.
`asyncdial`	[default FALSE] whether to perform dials asynchronously. The default FALSE will error if a connection is not immediately possible (for instance if `daemons` has yet to be called on the host, or the specified port is not open etc.). Specifying TRUE continues retrying (indefinitely) if not immediately successful, which is more resilient but can mask potential connection issues.
`autoexit`	[default TRUE] logical value, whether the daemon should exit automatically when its socket connection ends. If a signal from the tools package, such as `tools::SIGINT`, or an equivalent integer value is supplied, this signal is additionally raised on exit (see 'Persistence' section below).
`cleanup`	[default TRUE] logical value, whether to perform cleanup of the global environment and restore attached packages and options to an initial state after each evaluation. For more granular control, also accepts an integer value (see ‘Cleanup Options’ section below).
`output`	[default FALSE] logical value, to output generated stdout / stderr if TRUE, or else discard if FALSE. Specify as TRUE in the ‘...’ argument to `daemons` or `launch_local` to provide redirection of output to the host process (applicable only for local daemons).
`maxtasks`	[default Inf] the maximum number of tasks to execute (task limit) before exiting.
`idletime`	[default Inf] maximum idle time, since completion of the last task (in milliseconds) before exiting.
`walltime`	[default Inf] soft walltime, or the minimum amount of real time taken (in milliseconds) before exiting.
`timerstart`	[default 0L] number of completed tasks after which to start the timer for ‘idletime’ and ‘walltime’. 0L implies timers are started upon launch.
`...`	reserved but not currently used.
`tls`	[default NULL] required for secure TLS connections over 'tls+tcp://' or 'wss://'. Either the character path to a file containing X.509 certificate(s) in PEM format, comprising the certificate authority certificate chain starting with the TLS certificate and ending with the CA certificate, or a length 2 character vector comprising [i] the certificate authority certificate chain and [ii] the empty string `''`.
`rs`	[default NULL] the initial value of .Random.seed. This is set automatically using L'Ecuyer-CMRG RNG streams generated by the host process and should not be independently supplied.

Details

The network topology is such that daemons dial into the host or dispatcher, which listens at the ‘url’ address. In this way, network resources may be added or removed dynamically and the host or dispatcher automatically distributes tasks to all available daemons.

Value

Invisible NULL.

Persistence

The ‘autoexit’ argument governs persistence settings for the daemon. The default TRUE ensures that it will exit cleanly once its socket connection has ended.

Instead of TRUE, supplying a signal from the tools package, such as tools::SIGINT, or an equivalent integer value, sets this signal to be raised when the socket connection ends. For instance, supplying SIGINT allows a potentially more immediate exit by interrupting any ongoing evaluation rather than letting it complete.

Setting to FALSE allows the daemon to persist indefinitely even when there is no longer a socket connection. This allows a host session to end and a new session to connect at the URL where the daemon is dialled in. Daemons must be terminated with daemons(NULL) in this case, which sends explicit exit instructions to all connected daemons.

Cleanup Options

The ‘cleanup’ argument also accepts an integer value, which operates an additive bitmask: perform cleanup of the global environment (1L), reset attached packages to an initial state (2L), restore options to an initial state (4L), and perform garbage collection (8L).

As an example, to perform cleanup of the global environment and garbage collection, specify 9L (1L + 8L). The default argument value of TRUE performs all actions apart from garbage collection and is equivalent to a value of 7L.

Caution: do not reset options but not loaded packages if packages set options on load.

Daemons (Set Persistent Processes)

Description

Set ‘daemons’ or persistent background processes to receive mirai requests. Specify ‘n’ to create daemons on the local machine. Specify ‘url’ for receiving connections from remote daemons (for distributed computing across the network). Specify ‘remote’ to optionally launch remote daemons via a remote configuration. By default, dispatcher ensures optimal scheduling.

Usage

daemons(
  n,
  url = NULL,
  remote = NULL,
  dispatcher = c("process", "thread", "none"),
  ...,
  force = TRUE,
  seed = NULL,
  tls = NULL,
  pass = NULL,
  .compute = "default"
)
daemons(
  n,
  url = NULL,
  remote = NULL,
  dispatcher = c("process", "thread", "none"),
  ...,
  force = TRUE,
  seed = NULL,
  tls = NULL,
  pass = NULL,
  .compute = "default"
)

Arguments

`n`	integer number of daemons to set.
`url`	[default NULL] if specified, the character URL or vector of URLs on the host for remote daemons to dial into, including a port accepting incoming connections (and optionally for websockets, a path), e.g. 'tcp://hostname:5555' or 'ws://10.75.32.70:5555/path'. Specify a URL starting 'tls+tcp://' or 'wss://' to use secure TLS connections. Auxiliary function `host_url` may be used to construct a valid host URL.
`remote`	[default NULL] required only for launching remote daemons, a configuration generated by `remote_config` or `ssh_config`.
`dispatcher`	[default 'process'] character value, one of ‘process’, ‘thread’ or ‘none’. Whether to deploy dispatcher in another process, on a thread or not at all. Dispatcher is an extension that ensures optimal scheduling, although this is not always required (for details see Dispatcher section below). Note that the option ‘thread’ is new and currently considered experimental.
`...`	(optional) additional arguments passed through to `dispatcher` if using dispatcher and/or `daemon` if launching daemons. These include ‘retry’ and ‘token’ at dispatcher and ‘asyncdial’, ‘autoexit’, ‘cleanup’, ‘output’, ‘maxtasks’, ‘idletime’, ‘walltime’ and ‘timerstart’ at daemon.
`force`	[default TRUE] logical value whether to always reset daemons and apply new settings for a compute profile, even if already set. If FALSE, applying new settings requires daemons to be explicitly reset first using `daemons(0)`.
`seed`	[default NULL] (optional) supply a random seed (single value, interpreted as an integer). This is used to inititalise the L'Ecuyer-CMRG RNG streams sent to each daemon. Note that reproducible results can be expected only for `dispatcher = 'none'`, as the unpredictable timing of task completions would otherwise influence the tasks sent to each daemon. Even for `dispatcher = 'none'`, reproducibility is not guaranteed if the order in which tasks are sent is not deterministic.
`tls`	[default NULL] (optional for secure TLS connections) if not supplied, zero-configuration single-use keys and certificates are automatically generated. If supplied, either the character path to a file containing the PEM-encoded TLS certificate and associated private key (may contain additional certificates leading to a validation chain, with the TLS certificate first), or a length 2 character vector comprising [i] the TLS certificate (optionally certificate chain) and [ii] the associated private key.
`pass`	[default NULL] (required only if the private key supplied to ‘tls’ is encrypted with a password) For security, should be provided through a function that returns this value, rather than directly.
`.compute`	[default 'default'] character value for the compute profile to use (each compute profile has its own independent set of daemons).

Details

Use daemons(0) to reset daemon connections:

All connected daemons and/or dispatchers exit automatically.
mirai reverts to the default behaviour of creating a new background process for each request.
Any unresolved ‘mirai’ will return an ‘errorValue’ 19 (Connection reset) after a reset.
Calling daemons with revised (or even the same) settings for the same compute profile resets daemons before applying the new settings if force = TRUE.

If the host session ends, all connected dispatcher and daemon processes automatically exit as soon as their connections are dropped (unless the daemons were started with autoexit = FALSE). If a daemon is processing a task, it will exit as soon as the task is complete.

To reset persistent daemons started with autoexit = FALSE, use daemons(NULL) instead, which also sends exit instructions to all connected daemons prior to resetting.

For historical reasons, daemons() with no arguments returns the value of status.

Value

If using dispatcher, the integer number of daemons set, or else the integer number of daemons launched locally (zero if using a remote launcher).

Local Daemons

Daemons provide a potentially more efficient solution for asynchronous operations as new processes no longer need to be created on an ad hoc basis.

Supply the argument ‘n’ to set the number of daemons. New background daemon processes are automatically created on the local machine connecting back to the host process, either directly or via dispatcher.

Dispatcher

By default dispatcher = "process" launches a background process running dispatcher. Dispatcher connects to daemons on behalf of the host and ensures optimal FIFO scheduling of tasks.

Specifying dispatcher = "thread" runs dispatcher logic on a new thread, a faster and more efficient alternative to using a separate process. This is a new feature and should be considered experimental.

Specifying dispatcher = "none", uses the default behaviour without additional dispatcher logic. In this case daemons connect directly to the host and tasks are distributed in a round-robin fashion. Optimal scheduling is not guaranteed as the duration of tasks cannot be known a priori, hence tasks can be queued at one daemon while other daemons remain idle. However, this provides the most resource-light approach, suited to working with similar-length tasks, or where concurrent tasks typically do not exceed available daemons.

Distributed Computing

Specifying ‘url’ allows tasks to be distributed across the network. This should be a character string such as ‘tcp://10.75.32.70:5555’ at which daemon processes should connect to. Switching the URL scheme to ‘tls+tcp://’ or ‘wss://’ automatically upgrades the connection to use TLS. The auxiliary function host_url may be used to automatically construct a valid host URL based on the computer's hostname.

Specify ‘remote’ with a call to remote_config or ssh_config to launch daemons on remote machines. Otherwise, launch_remote may be used to generate the shell commands to deploy daemons manually on remote resources.

IPv6 addresses are also supported and must be enclosed in square brackets [ ] to avoid confusion with the final colon separating the port. For example, port 5555 on the IPv6 loopback address ::1 would be specified as ‘tcp://[::1]:5555’.

Specifying the wildcard value zero for the port number e.g. ‘tcp://[::1]:0’ or ‘ws://[::1]:0’ will automatically assign a free ephemeral port. Use status to inspect the actual assigned port at any time.

With Dispatcher

When using dispatcher, it is recommended to use a websocket URL rather than TCP, as this requires only one port to connect to all daemons: a websocket URL supports a path after the port number, which can be made unique for each daemon.

Specifying a single host URL such as ‘ws://10.75.32.70:5555’ with n = 6 will automatically append a sequence to the path, listening to the URLs ‘ws://10.75.32.70:5555/1’ through ‘ws://110.75.32.70:5555/6’.

Alternatively, specify a vector of URLs to listen to arbitrary port numbers / paths. In this case it is optional to supply ‘n’ as this can be inferred by the length of vector supplied.

Individual daemons then dial in to each of these host URLs. At most one daemon can be dialled into each URL at any given time.

Dispatcher automatically adjusts to the number of daemons actually connected. Hence it is possible to dynamically scale up or down the number of daemons as required, subject to the maximum number initially specified.

Alternatively, supplying a single TCP URL will listen at a block of URLs with ports starting from the supplied port number and incrementing by one for ‘n’ specified e.g. the host URL ‘tcp://10.75.32.70:5555’ with n = 6 listens to the contiguous block of ports 5555 through 5560.

Without Dispatcher

A TCP URL may be used in this case as the host listens at only one address, utilising a single port.

The network topology is such that daemons (started with daemon) or indeed dispatchers (started with dispatcher) dial into the same host URL.

‘n’ is not required in this case, and disregarded if supplied, as network resources may be added or removed at any time. The host automatically distributes tasks to all connected daemons and dispatchers in a round-robin fashion.

Compute Profiles

By default, the ‘default’ compute profile is used. Providing a character value for ‘.compute’ creates a new compute profile with the name specified. Each compute profile retains its own daemons settings, and may be operated independently of each other. Some usage examples follow:

local / remote daemons may be set with a host URL and specifying ‘.compute’ as ‘remote’, which creates a new compute profile. Subsequent mirai calls may then be sent for local computation by not specifying the ‘.compute’ argument, or for remote computation to connected daemons by specifying the ‘.compute’ argument as ‘remote’.

cpu / gpu some tasks may require access to different types of daemon, such as those with GPUs. In this case, daemons() may be called to set up host URLs for CPU-only daemons and for those with GPUs, specifying the ‘.compute’ argument as ‘cpu’ and ‘gpu’ respectively. By supplying the ‘.compute’ argument to subsequent mirai calls, tasks may be sent to either ‘cpu’ or ‘gpu’ daemons as appropriate.

Note: further actions such as resetting daemons via daemons(0) should be carried out with the desired ‘.compute’ argument specified.

Examples

if (interactive()) {
# Only run examples in interactive R sessions

# Create 2 local daemons (using dispatcher)
daemons(2)
status()
# Reset to zero
daemons(0)

# Create 2 local daemons (not using dispatcher)
daemons(2, dispatcher = "none")
status()
# Reset to zero
daemons(0)

# 2 remote daemons via dispatcher using WebSockets
daemons(2, url = host_url(ws = TRUE))
status()
# Reset to zero
daemons(0)

# Set host URL for remote daemons to dial into
daemons(url = host_url(), dispatcher = "none")
status()
# Reset to zero
daemons(0)

# Use with() to evaluate with daemons for the duration of the expression
with(
  daemons(2),
  {
    m1 <- mirai(Sys.getpid())
    m2 <- mirai(Sys.getpid())
    cat(call_mirai(m1)$data, call_mirai(m2)$data, "\n")
  }
)

}

## Not run: 
# Launch 2 daemons on remotes 'nodeone' and 'nodetwo' using SSH
# connecting back directly to the host URL over a TLS connection:

daemons(url = host_url(tls = TRUE),
        remote = ssh_config(c('ssh://nodeone', 'ssh://nodetwo')),
        dispatcher = "none")

# Launch 4 daemons on the remote machine 10.75.32.90 using SSH tunnelling
# over port 5555 ('url' hostname must be 'localhost' or '127.0.0.1'):

daemons(n = 4,
        url = 'ws://localhost:5555',
        remote = ssh_config('ssh://10.75.32.90', tunnel = TRUE))


## End(Not run)

if (interactive()) {
# Only run examples in interactive R sessions

# Create 2 local daemons (using dispatcher)
daemons(2)
status()
# Reset to zero
daemons(0)

# Create 2 local daemons (not using dispatcher)
daemons(2, dispatcher = "none")
status()
# Reset to zero
daemons(0)

# 2 remote daemons via dispatcher using WebSockets
daemons(2, url = host_url(ws = TRUE))
status()
# Reset to zero
daemons(0)

# Set host URL for remote daemons to dial into
daemons(url = host_url(), dispatcher = "none")
status()
# Reset to zero
daemons(0)

# Use with() to evaluate with daemons for the duration of the expression
with(
  daemons(2),
  {
    m1 <- mirai(Sys.getpid())
    m2 <- mirai(Sys.getpid())
    cat(call_mirai(m1)$data, call_mirai(m2)$data, "\n")
  }
)

}

## Not run: 
# Launch 2 daemons on remotes 'nodeone' and 'nodetwo' using SSH
# connecting back directly to the host URL over a TLS connection:

daemons(url = host_url(tls = TRUE),
        remote = ssh_config(c('ssh://nodeone', 'ssh://nodetwo')),
        dispatcher = "none")

# Launch 4 daemons on the remote machine 10.75.32.90 using SSH tunnelling
# over port 5555 ('url' hostname must be 'localhost' or '127.0.0.1'):

daemons(n = 4,
        url = 'ws://localhost:5555',
        remote = ssh_config('ssh://10.75.32.90', tunnel = TRUE))


## End(Not run)

Dispatcher

Description

Dispatches tasks from a host to daemons for processing, using FIFO scheduling, queuing tasks as required. Daemon / dispatcher settings may be controlled by daemons and this function should not need to be invoked directly.

Usage

dispatcher(
  host,
  url = NULL,
  n = NULL,
  ...,
  retry = FALSE,
  token = FALSE,
  tls = NULL,
  pass = NULL,
  rs = NULL,
  monitor = NULL
)
dispatcher(
  host,
  url = NULL,
  n = NULL,
  ...,
  retry = FALSE,
  token = FALSE,
  tls = NULL,
  pass = NULL,
  rs = NULL,
  monitor = NULL
)

Arguments

`host`	the character host URL to dial (where tasks are sent from), including the port to connect to (and optionally for websockets, a path), e.g. 'tcp://hostname:5555' or 'ws://10.75.32.70:5555/path'.
`url`	(optional) the character URL or vector of URLs dispatcher should listen at, including the port to connect to (and optionally for websockets, a path), e.g. 'tcp://hostname:5555' or 'ws://10.75.32.70:5555/path'. Specify 'tls+tcp://' or 'wss://' to use secure TLS connections. Tasks are sent to daemons dialled into these URLs. If not supplied, ‘n’ local inter-process URLs will be assigned automatically.
`n`	(optional) if specified, the integer number of daemons to listen for. Otherwise ‘n’ will be inferred from the number of URLs supplied in ‘url’. Where a single URL is supplied and ‘n’ > 1, ‘n’ unique URLs will be automatically assigned for daemons to dial into.
`...`	(optional) additional arguments passed through to `daemon`. These include ‘asyncdial’, ‘autoexit’, ‘cleanup’, ‘maxtasks’, ‘idletime’, ‘walltime’ and ‘timerstart’.
`retry`	[default FALSE] logical value, whether to automatically retry tasks where the daemon crashes or terminates unexpectedly on the next daemon instance to connect. If TRUE, the mirai will remain unresolved but `status` will show ‘online’ as 0 and ‘assigned’ > ‘complete’. To cancel a task in this case, use `saisei(force = TRUE)`. If FALSE, such tasks will be returned as ‘errorValue’ 19 (Connection reset).
`token`	[default FALSE] if TRUE, appends a unique 24-character token to each URL path the dispatcher listens at (not applicable for TCP URLs which do not accept a path).
`tls`	[default NULL] (required for secure TLS connections) either the character path to a file containing the PEM-encoded TLS certificate and associated private key (may contain additional certificates leading to a validation chain, with the TLS certificate first), or a length 2 character vector comprising [i] the TLS certificate (optionally certificate chain) and [ii] the associated private key.
`pass`	[default NULL] (required only if the private key supplied to ‘tls’ is encrypted with a password) For security, should be provided through a function that returns this value, rather than directly.
`rs`	[default NULL] the initial value of .Random.seed. This is set automatically using L'Ecuyer-CMRG RNG streams generated by the host process and should not be independently supplied.
`monitor`	(for package internal use only) do not set this parameter.

Details

The network topology is such that a dispatcher acts as a gateway between the host and daemons, ensuring that tasks received from the host are dispatched on a FIFO basis for processing. Tasks are queued at the dispatcher to ensure tasks are only sent to daemons that can begin immediate execution of the task.

Value

Invisible NULL.

Evaluate Everywhere

Description

Evaluate an expression ‘everywhere’ on all connected daemons for the specified compute profile. Designed for performing setup operations across daemons by loading packages, exporting common data, or registering custom serialization functions. Resultant changes to the global environment, loaded packages and options are persisted regardless of a daemon's ‘cleanup’ setting.

Usage

everywhere(.expr, ..., .args = list(), .serial = NULL, .compute = "default")
everywhere(.expr, ..., .args = list(), .serial = NULL, .compute = "default")

Arguments

`.expr`	an expression to evaluate asynchronously (of arbitrary length, wrapped in { } where necessary), or else a pre-constructed language object.
`...`	(optional) either named arguments (name = value pairs) specifying objects referenced, but not defined, in ‘.expr’, or an environment containing such objects. See ‘evaluation’ section below.
`.args`	(optional) either a named list specifying objects referenced, but not defined, in ‘.expr’, or an environment containing such objects. These objects will remain local to the evaluation environment as opposed to those supplied in ‘...’ above - see ‘evaluation’ section below.
`.serial`	[default NULL] (optional) a configuration created by `serial_config` to register serialization and unserialization functions for normally non-exportable reference objects, such as Arrow Tables or torch tensors. Updating with a new configuration replaces any existing registered functions. To remove the configuration, provide an empty list.
`.compute`	[default 'default'] character value for the compute profile to use (each compute profile has its own independent set of daemons).

Value

Invisible NULL. Will error if the specified compute profile is not found, i.e. not yet set up.

Evaluation

The expression ‘.expr’ will be evaluated in a separate R process in a clean environment (not the global environment), consisting only of the objects in the list or environment supplied to ‘.args’, with the named objects passed as ‘...’ (from the environment if one was supplied) assigned to the global environment of that process.

For evaluation to occur as if in your global environment, supply objects to ‘...’ rather than ‘.args’. For stricter scoping, use ‘.args’, which limits, for example, where variables not explicitly passed as arguments to functions are found.

As evaluation occurs in a clean environment, all undefined objects must be supplied though ‘...’ and/or ‘.args’, including self-defined functions. Functions from a package should use namespaced calls such as mirai::mirai(), or else the package should be loaded beforehand as part of ‘.expr’.

Examples

if (interactive()) {
# Only run examples in interactive R sessions

daemons(1)
# export common data by a super-assignment expression:
everywhere(y <<- 3)
# '...' variables are assigned to the global environment
# '.expr' may be specified as an empty {} in such cases:
everywhere({}, a = 1, b = 2)
m <- mirai(a + b - y == 0L)
call_mirai(m)$data
daemons(0)

# loading a package on all daemons and also
# registering custom serialization functions:
cfg <- serial_config("cls_name", function(x) serialize(x, NULL), unserialize)
daemons(1, dispatcher = "none")
everywhere(library(parallel), .serial = cfg)
m <- mirai("package:parallel" %in% search())
call_mirai(m)$data
daemons(0)

}

if (interactive()) {
# Only run examples in interactive R sessions

daemons(1)
# export common data by a super-assignment expression:
everywhere(y <<- 3)
# '...' variables are assigned to the global environment
# '.expr' may be specified as an empty {} in such cases:
everywhere({}, a = 1, b = 2)
m <- mirai(a + b - y == 0L)
call_mirai(m)$data
daemons(0)

# loading a package on all daemons and also
# registering custom serialization functions:
cfg <- serial_config("cls_name", function(x) serialize(x, NULL), unserialize)
daemons(1, dispatcher = "none")
everywhere(library(parallel), .serial = cfg)
m <- mirai("package:parallel" %in% search())
call_mirai(m)$data
daemons(0)

}

URL Constructors

Description

host_url constructs a valid host URL (at which daemons may connect) based on the computer's hostname. This may be supplied directly to the ‘url’ argument of daemons.

local_url constructs a random URL suitable for local daemons.

Usage

host_url(ws = FALSE, tls = FALSE, port = 0)

local_url()
host_url(ws = FALSE, tls = FALSE, port = 0)

local_url()

Arguments

`ws`	[default FALSE] logical value whether to use a WebSockets 'ws://' or else TCP 'tcp://' scheme.
`tls`	[default FALSE] logical value whether to use TLS in which case the scheme used will be either 'wss://' or 'tls+tcp://' accordingly.
`port`	[default 0] numeric port to use. This should be open to connections from the network addresses the daemons are connecting from. ‘0’ is a wildcard value that automatically assigns a free ephemeral port.

Details

host_url relies on using the host name of the computer rather than an IP address and typically works on local networks, although this is not always guaranteed. If unsuccessful, substitute an IPv4 or IPv6 address in place of the hostname.

local_url generates a random URL for the platform's default inter-process communications transport: abstract Unix domain sockets on Linux, Unix domain sockets on MacOS, Solaris and other POSIX platforms, and named pipes on Windows.

Value

A character string comprising a valid URL.

Examples

host_url()
host_url(ws = TRUE)
host_url(tls = TRUE)
host_url(ws = TRUE, tls = TRUE, port = 5555)

local_url()

host_url()
host_url(ws = TRUE)
host_url(tls = TRUE)
host_url(ws = TRUE, tls = TRUE, port = 5555)

local_url()

Is mirai / mirai_map

Description

Is the object a ‘mirai’ or ‘mirai_map’.

Usage

is_mirai(x)

is_mirai_map(x)
is_mirai(x)

is_mirai_map(x)

Arguments

`x`	an object.

Value

Logical TRUE if ‘x’ is of class ‘mirai’ or ‘mirai_map’ respectively, FALSE otherwise.

Examples

if (interactive()) {
# Only run examples in interactive R sessions

daemons(1, dispatcher = "none")
df <- data.frame()
m <- mirai(as.matrix(df), df = df)
is_mirai(m)
is_mirai(df)

mp <- mirai_map(1:3, runif)
is_mirai_map(mp)
is_mirai_map(mp[])
daemons(0)

}

if (interactive()) {
# Only run examples in interactive R sessions

daemons(1, dispatcher = "none")
df <- data.frame()
m <- mirai(as.matrix(df), df = df)
is_mirai(m)
is_mirai(df)

mp <- mirai_map(1:3, runif)
is_mirai_map(mp)
is_mirai_map(mp[])
daemons(0)

}

Error Validators

Description

Validator functions for error value types created by mirai.

Usage

is_mirai_error(x)

is_mirai_interrupt(x)

is_error_value(x)
is_mirai_error(x)

is_mirai_interrupt(x)

is_error_value(x)

Arguments

`x`	an object.

Details

Is the object a ‘miraiError’. When execution in a ‘mirai’ process fails, the error message is returned as a character string of class ‘miraiError’ and ‘errorValue’. The stack trace is available at $stack.trace on the error object.

Is the object a ‘miraiInterrupt’. When an ongoing ‘mirai’ is sent a user interrupt, it will resolve to an empty character string classed as ‘miraiInterrupt’ and ‘errorValue’.

Is the object an ‘errorValue’, such as a ‘mirai’ timeout, a ‘miraiError’ or a ‘miraiInterrupt’. This is a catch-all condition that includes all returned error values.

Value

Logical value TRUE or FALSE.

Examples

if (interactive()) {
# Only run examples in interactive R sessions

m <- mirai(stop())
call_mirai(m)
is_mirai_error(m$data)
is_mirai_interrupt(m$data)
is_error_value(m$data)
m$data$stack.trace

m2 <- mirai(Sys.sleep(1L), .timeout = 100)
call_mirai(m2)
is_mirai_error(m2$data)
is_mirai_interrupt(m2$data)
is_error_value(m2$data)

}

if (interactive()) {
# Only run examples in interactive R sessions

m <- mirai(stop())
call_mirai(m)
is_mirai_error(m$data)
is_mirai_interrupt(m$data)
is_error_value(m$data)
m$data$stack.trace

m2 <- mirai(Sys.sleep(1L), .timeout = 100)
call_mirai(m2)
is_mirai_error(m2$data)
is_mirai_interrupt(m2$data)
is_error_value(m2$data)

}

Launch Daemon

Description

launch_local spawns a new background Rscript process calling daemon with the specified arguments.

launch_remote returns the shell command for deploying daemons as a character vector. If a configuration generated by remote_config or ssh_config is supplied then this is used to launch the daemon on the remote machine.

Usage

launch_local(url, ..., tls = NULL, .compute = "default")

launch_remote(
  url,
  remote = remote_config(),
  ...,
  tls = NULL,
  .compute = "default"
)
launch_local(url, ..., tls = NULL, .compute = "default")

launch_remote(
  url,
  remote = remote_config(),
  ...,
  tls = NULL,
  .compute = "default"
)

Arguments

`url`	the character host URL or vector of host URLs, including the port to connect to (and optionally for websockets, a path), e.g. 'tcp://hostname:5555' or 'ws://10.75.32.70:5555/path' or integer index value, or vector of index values, of the dispatcher URLs, or 1L for the host URL (when not using dispatcher). or for `launch_remote` only, a ‘miraiCluster’ or ‘miraiNode’.
`...`	(optional) additional arguments passed through to `daemon`. These include ‘autoexit’, ‘cleanup’, ‘output’, ‘maxtasks’, ‘idletime’, ‘walltime’ and ‘timerstart’.
`tls`	[default NULL] required for secure TLS connections over tls+tcp or wss. Zero-configuration TLS certificates generated by `daemons` are automatically passed to the daemon, without requiring to be specified here. Otherwise, supply either the character path to a file containing X.509 certificate(s) in PEM format, comprising the certificate authority certificate chain, or a length 2 character vector comprising [i] the certificate authority certificate chain and [ii] the empty character ”.
`.compute`	[default 'default'] character value for the compute profile to use (each compute profile has its own independent set of daemons).
`remote`	required only for launching remote daemons, a configuration generated by `remote_config` or `ssh_config`. An empty `remote_config` does not effect any daemon launches but returns the shell commands for deploying manually on remote machines.

Details

These functions may be used to re-launch daemons that have exited after reaching time or task limits.

Daemons must already be set for launchers to work.

The generated command contains the argument ‘rs’ specifying the length 7 L'Ecuyer-CMRG random seed supplied to the daemon. The values will be different each time the function is called.

Value

For launch_local: Invisible NULL.

For launch_remote: A character vector of daemon launch commands, classed as ‘miraiLaunchCmd’. The printed output may be copy / pasted directly to the remote machine.

Examples

if (interactive()) {
# Only run examples in interactive R sessions

daemons(url = host_url(ws = TRUE), dispatcher = "none")
status()
launch_local(status()$daemons, maxtasks = 10L)
launch_remote(1L, maxtasks = 10L)
Sys.sleep(1)
status()
daemons(0)

daemons(n = 2L, url = host_url(tls = TRUE))
status()
launch_local(1:2, idletime = 60000L, timerstart = 1L)
launch_remote(1:2, idletime = 60000L, timerstart = 1L)
Sys.sleep(1)
status()
daemons(0)

}

if (interactive()) {
# Only run examples in interactive R sessions

daemons(url = host_url(ws = TRUE), dispatcher = "none")
status()
launch_local(status()$daemons, maxtasks = 10L)
launch_remote(1L, maxtasks = 10L)
Sys.sleep(1)
status()
daemons(0)

daemons(n = 2L, url = host_url(tls = TRUE))
status()
launch_local(1:2, idletime = 60000L, timerstart = 1L)
launch_remote(1:2, idletime = 60000L, timerstart = 1L)
Sys.sleep(1)
status()
daemons(0)

}

Make Mirai Cluster

Description

make_cluster creates a cluster of type ‘miraiCluster’, which may be used as a cluster object for any function in the parallel base package such as clusterApply or parLapply.

stop_cluster stops a cluster created by make_cluster.

Usage

make_cluster(n, url = NULL, remote = NULL, ...)

stop_cluster(cl)
make_cluster(n, url = NULL, remote = NULL, ...)

stop_cluster(cl)

Arguments

`n`	integer number of nodes (automatically launched on the local machine unless ‘url’ is supplied).
`url`	[default NULL] (specify for remote nodes) the character URL on the host for remote nodes to dial into, including a port accepting incoming connections, e.g. 'tcp://10.75.37.40:5555'. Specify a URL with the scheme ‘tls+tcp://’ to use secure TLS connections.
`remote`	[default NULL] (specify to launch remote nodes) a remote launch configuration generated by `remote_config` or `ssh_config`. If not supplied, nodes may be deployed manually on remote resources.
`...`	additional arguments passed onto `daemons`.
`cl`	a ‘miraiCluster’.

Value

For make_cluster: An object of class ‘miraiCluster’ and ‘cluster’. Each ‘miraiCluster’ has an automatically assigned ID and ‘n’ nodes of class ‘miraiNode’. If ‘url’ is supplied but not ‘remote’, the shell commands for deployment of nodes on remote resources are printed to the console.

For stop_cluster: invisible NULL.

Remote Nodes

Specify ‘url’ and ‘n’ to set up a host connection for remote nodes to dial into. ‘n’ defaults to one if not specified.

Also specify ‘remote’ to launch the nodes using a configuration generated by remote_config or ssh_config. In this case, the number of nodes is inferred from the configuration provided and ‘n’ is disregarded.

If ‘remote’ is not supplied, the shell commands for deploying nodes manually on remote resources are automatically printed to the console.

launch_remote may be called at any time on a ‘miraiCluster’ to return the shell commands for deployment of all nodes, or on a ‘miraiNode’ to return the command for a single node.

Status

Call status on a ‘miraiCluster’ to check the number of currently active connections as well as the host URL.

Errors

Errors are thrown by the ‘parallel’ mechanism if one or more nodes failed (quit unexpectedly). The resulting ‘errorValue’ returned is 19 (Connection reset). Other types of error, e.g. in evaluation, should result in the usual ‘miraiError’ being returned.

Note

The default behaviour of clusters created by this function is designed to map as closely as possible to clusters created by the parallel package. However, ‘...’ arguments are passed onto daemons for additional customisation if desired, although resultant behaviour may not always be supported.

Examples

if (interactive()) {
# Only run examples in interactive R sessions

cl <- make_cluster(2)
cl
cl[[1L]]

Sys.sleep(0.5)
status(cl)

stop_cluster(cl)

}

if (interactive()) {
# Only run examples in interactive R sessions

cl <- make_cluster(2)
cl
cl[[1L]]

Sys.sleep(0.5)
status(cl)

stop_cluster(cl)

}

mirai (Evaluate Async)

Description

Evaluate an expression asynchronously in a new background R process or persistent daemon (local or remote). This function will return immediately with a ‘mirai’, which will resolve to the evaluated result once complete.

Usage

mirai(.expr, ..., .args = list(), .timeout = NULL, .compute = "default")
mirai(.expr, ..., .args = list(), .timeout = NULL, .compute = "default")

Arguments

`.expr`	an expression to evaluate asynchronously (of arbitrary length, wrapped in { } where necessary), or else a pre-constructed language object.
`...`	(optional) either named arguments (name = value pairs) specifying objects referenced, but not defined, in ‘.expr’, or an environment containing such objects. See ‘evaluation’ section below.
`.args`	(optional) either a named list specifying objects referenced, but not defined, in ‘.expr’, or an environment containing such objects. These objects will remain local to the evaluation environment as opposed to those supplied in ‘...’ above - see ‘evaluation’ section below.
`.timeout`	[default NULL] for no timeout, or an integer value in milliseconds. A mirai will resolve to an ‘errorValue’ 5 (timed out) if evaluation exceeds this limit.
`.compute`	[default 'default'] character value for the compute profile to use (each compute profile has its own independent set of daemons).

Details

This function will return a ‘mirai’ object immediately.

The value of a mirai may be accessed at any time at $data, and if yet to resolve, an ‘unresolved’ logical NA will be returned instead.

unresolved may be used on a mirai, returning TRUE if a ‘mirai’ has yet to resolve and FALSE otherwise. This is suitable for use in control flow statements such as while or if.

Alternatively, to call (and wait for) the result, use call_mirai on the returned ‘mirai’. This will block until the result is returned.

Specify ‘.compute’ to send the mirai using a specific compute profile (if previously created by daemons), otherwise leave as ‘default’.

Value

A ‘mirai’ object.

Evaluation

Errors

is_error_value tests for all error conditions including ‘mirai’ errors, interrupts, and timeouts.

Examples

if (interactive()) {
# Only run examples in interactive R sessions

# specifying objects via '...'
n <- 3
m <- mirai(x + y + 2, x = 2, y = n)
m
m$data
Sys.sleep(0.2)
m$data

# passing the calling environment to '...'
df1 <- data.frame(a = 1, b = 2)
df2 <- data.frame(a = 3, b = 1)
m <- mirai(as.matrix(rbind(df1, df2)), environment(), .timeout = 1000)
call_mirai(m)$data

# using unresolved()
m <- mirai(
  {
    res <- rnorm(n)
    res / rev(res)
  },
  n = 1e6
)
while (unresolved(m)) {
  cat("unresolved\n")
  Sys.sleep(0.1)
}
str(m$data)

# evaluating scripts using source() in '.expr'
n <- 10L
file <- tempfile()
cat("r <- rnorm(n)", file = file)
m <- mirai({source(file); r}, file = file, n = n)
call_mirai(m)[["data"]]
unlink(file)

# use source(local = TRUE) when passing in local variables via '.args'
n <- 10L
file <- tempfile()
cat("r <- rnorm(n)", file = file)
m <- mirai({source(file, local = TRUE); r}, .args = list(file = file, n = n))
call_mirai(m)[["data"]]
unlink(file)

# passing a language object to '.expr' and a named list to '.args'
expr <- quote(a + b + 2)
args <- list(a = 2, b = 3)
m <- mirai(.expr = expr, .args = args)
call_mirai(m)$data

}

if (interactive()) {
# Only run examples in interactive R sessions

# specifying objects via '...'
n <- 3
m <- mirai(x + y + 2, x = 2, y = n)
m
m$data
Sys.sleep(0.2)
m$data

# passing the calling environment to '...'
df1 <- data.frame(a = 1, b = 2)
df2 <- data.frame(a = 3, b = 1)
m <- mirai(as.matrix(rbind(df1, df2)), environment(), .timeout = 1000)
call_mirai(m)$data

# using unresolved()
m <- mirai(
  {
    res <- rnorm(n)
    res / rev(res)
  },
  n = 1e6
)
while (unresolved(m)) {
  cat("unresolved\n")
  Sys.sleep(0.1)
}
str(m$data)

# evaluating scripts using source() in '.expr'
n <- 10L
file <- tempfile()
cat("r <- rnorm(n)", file = file)
m <- mirai({source(file); r}, file = file, n = n)
call_mirai(m)[["data"]]
unlink(file)

# use source(local = TRUE) when passing in local variables via '.args'
n <- 10L
file <- tempfile()
cat("r <- rnorm(n)", file = file)
m <- mirai({source(file, local = TRUE); r}, .args = list(file = file, n = n))
call_mirai(m)[["data"]]
unlink(file)

# passing a language object to '.expr' and a named list to '.args'
expr <- quote(a + b + 2)
args <- list(a = 2, b = 3)
m <- mirai(.expr = expr, .args = args)
call_mirai(m)$data

}

mirai Map

Description

Asynchronous parallel map of a function over a list or vector using mirai, with optional promises integration. Performs multiple map over the rows of a dataframe or matrix.

Usage

mirai_map(.x, .f, ..., .args = list(), .promise = NULL, .compute = "default")
mirai_map(.x, .f, ..., .args = list(), .promise = NULL, .compute = "default")

Arguments

`.x`	a list or atomic vector. Also accepts a matrix or dataframe, in which case multiple map is performed over its rows.
`.f`	a function to be applied to each element of `.x`, or row of `.x` as the case may be.
`...`	(optional) named arguments (name = value pairs) specifying objects referenced, but not defined, in `.f`.
`.args`	(optional) further constant arguments to `.f`, provided as a list.
`.promise`	(optional) if supplied, registers a promise against each mirai. Either a function, supplied to the ‘onFulfilled’ argument of `promises::then()` or a list of 2 functions, supplied respectively to ‘onFulfilled’ and ‘onRejected’ for `promises::then()`. Using this argument requires the promises package.
`.compute`	[default 'default'] character value for the compute profile to use (each compute profile has its own independent set of daemons).

Details

Sends each application of function .f on an element of .x (or row of .x) for computation in a separate mirai call.

This simple and transparent behaviour is designed to make full use of mirai scheduling to minimise overall execution time.

Facilitates recovery from partial failure by returning all ‘miraiError’ / ‘errorValue’ as the case may be, thus allowing only the failures to be re-run.

Note: requires daemons to have previously been set. If not, then one local daemon is set before the function proceeds.

Value

A ‘mirai_map’ (list of ‘mirai’ objects).

Results

x[] collects the results of a ‘mirai_map’ x and returns a list. This will wait for all asynchronous operations to complete if still in progress, blocking but user-interruptible.

x[.flat] collects and flattens map results to a vector, checking that they are of the same type to avoid coercion. Note: errors if an ‘errorValue’ has been returned or results are of differing type.

x[.progress] collects map results whilst showing a simple text progress indicator of parts completed of the total.

x[.progress_cli] collects map results whilst showing a progress bar from the cli package, if available, with completion percentage and ETA.

x[.stop] collects map results applying early stopping, which stops at the first failure and cancels remaining operations. Note: operations already in progress continue to completion, although their results are not collected.

The options above may be combined in the manner of:
x[.stop, .progress] which applies early stopping together with a progress indicator.

Multiple Map

Multiple map is performed automatically over the rows of an object with ‘dim’ attributes such as a matrix or dataframe. This is most often the desired behaviour in these cases.

To map over columns instead, first wrap a dataframe in as.list, or transpose a matrix using t.

Examples

if (interactive()) {
# Only run examples in interactive R sessions

daemons(4, dispatcher = "none")

# map with constant args specified via '.args'
mirai_map(1:3, rnorm, .args = list(mean = 20, sd = 2))[]

# flatmap with function definition passed via '...'
mirai_map(1:3, function(x) func(1L, x, x + 1L), func = stats::runif)[.flat]

# sum rows of a dataframe
(df <- data.frame(a = 1:3, b = c(4, 3, 2)))
mirai_map(df, sum)[.flat]

# sum rows of a matrix
(mat <- matrix(1:4, nrow = 2L))
mirai_map(mat, sum)[.flat]

# map over rows of a dataframe
df <- data.frame(a = c("Aa", "Bb"), b = c(1L, 4L))
mirai_map(df, function(...) sprintf("%s: %d", ...))[.flat]

# indexed map over a vector
v <- c("egg", "got", "ten", "nap", "pie")
mirai_map(
  data.frame(1:length(v), v),
  sprintf,
  .args = list(fmt = "%d_%s")
)[.flat]

# return a 'mirai_map' object, check for resolution, collect later
mp <- mirai_map(
  c(a = 2, b = 3, c = 4),
  function(x, y) do(x, as.logical(x %% y)),
  do = nanonext::random,
  .args = list(y = 2)
)
unresolved(mp)
mp
mp[]
unresolved(mp)

# progress indicator counts up from 0 to 4 seconds
res <- mirai_map(1:4, Sys.sleep)[.progress]

daemons(0)

# generates warning as daemons not set
# stops early when second element returns an error
tryCatch(
  mirai_map(list(1, "a", 3), sum)[.stop],
  error = identity
)

# promises example that outputs the results, including errors, to the console
if (requireNamespace("promises", quietly = TRUE)) {
daemons(1, dispatcher = "none")
ml <- mirai_map(
  1:30,
  function(x) {Sys.sleep(0.1); if (x == 30) stop(x) else x},
  .promise = list(
    function(x) cat(paste(x, "")),
    function(x) { cat(conditionMessage(x), "\n"); daemons(0) }
  )
)
}

}

if (interactive()) {
# Only run examples in interactive R sessions

daemons(4, dispatcher = "none")

# map with constant args specified via '.args'
mirai_map(1:3, rnorm, .args = list(mean = 20, sd = 2))[]

# flatmap with function definition passed via '...'
mirai_map(1:3, function(x) func(1L, x, x + 1L), func = stats::runif)[.flat]

# sum rows of a dataframe
(df <- data.frame(a = 1:3, b = c(4, 3, 2)))
mirai_map(df, sum)[.flat]

# sum rows of a matrix
(mat <- matrix(1:4, nrow = 2L))
mirai_map(mat, sum)[.flat]

# map over rows of a dataframe
df <- data.frame(a = c("Aa", "Bb"), b = c(1L, 4L))
mirai_map(df, function(...) sprintf("%s: %d", ...))[.flat]

# indexed map over a vector
v <- c("egg", "got", "ten", "nap", "pie")
mirai_map(
  data.frame(1:length(v), v),
  sprintf,
  .args = list(fmt = "%d_%s")
)[.flat]

# return a 'mirai_map' object, check for resolution, collect later
mp <- mirai_map(
  c(a = 2, b = 3, c = 4),
  function(x, y) do(x, as.logical(x %% y)),
  do = nanonext::random,
  .args = list(y = 2)
)
unresolved(mp)
mp
mp[]
unresolved(mp)

# progress indicator counts up from 0 to 4 seconds
res <- mirai_map(1:4, Sys.sleep)[.progress]

daemons(0)

# generates warning as daemons not set
# stops early when second element returns an error
tryCatch(
  mirai_map(list(1, "a", 3), sum)[.stop],
  error = identity
)

# promises example that outputs the results, including errors, to the console
if (requireNamespace("promises", quietly = TRUE)) {
daemons(1, dispatcher = "none")
ml <- mirai_map(
  1:30,
  function(x) {Sys.sleep(0.1); if (x == 30) stop(x) else x},
  .promise = list(
    function(x) cat(paste(x, "")),
    function(x) { cat(conditionMessage(x), "\n"); daemons(0) }
  )
)
}

}

Generic and SSH Remote Launch Configuration

Description

remote_config provides a flexible generic framework for generating the shell commands to deploy daemons remotely.

ssh_config generates a remote configuration for launching daemons over SSH, with the option of SSH tunnelling.

Usage

remote_config(
  command = NULL,
  args = c("", "."),
  rscript = "Rscript",
  quote = FALSE
)

ssh_config(
  remotes,
  tunnel = FALSE,
  timeout = 10,
  command = "ssh",
  rscript = "Rscript",
  host
)
remote_config(
  command = NULL,
  args = c("", "."),
  rscript = "Rscript",
  quote = FALSE
)

ssh_config(
  remotes,
  tunnel = FALSE,
  timeout = 10,
  command = "ssh",
  rscript = "Rscript",
  host
)

Arguments

`command`	the command used to effect the daemon launch on the remote machine as a character string (e.g. `'ssh'`). Defaults to ‘ssh’ for `ssh_config`, although may be substituted for the full path to a specific SSH application. The default NULL for `remote_config` does not effect any launches, but causes `launch_remote` to return the shell commands for manual deployment on remote machines.
`args`	(optional) arguments passed to ‘command’, as a character vector that must include `"."` as an element, which will be substituted for the daemon launch command. Alternatively, a list of such character vectors to effect multiple launches (one for each list element).
`rscript`	(optional) name / path of the Rscript executable on the remote machine. The default assumes ‘Rscript’ is on the executable search path. Prepend the full path if necessary. If launching on Windows, ‘Rscript’ should be replaced with ‘Rscript.exe’.
`quote`	[default FALSE] logical value whether or not to quote the daemon launch command (not required for Slurm ‘srun’ for example, but required for ‘ssh’ or Slurm ‘sbatch’).
`remotes`	the character URL or vector of URLs to SSH into, using the 'ssh://' scheme and including the port open for SSH connections (defaults to 22 if not specified), e.g. 'ssh://10.75.32.90:22' or 'ssh://nodename'.
`tunnel`	[default FALSE] logical value whether to use SSH reverse tunnelling. If TRUE, a tunnel is created between the same ports on the local and remote machines. See the ‘SSH Tunnelling’ section below for how to correctly specify required settings.
`timeout`	[default 10] maximum time allowed for connection setup in seconds.
`host`	(optional) only applicable for reverse tunnelling. Should be specified if creating a standalone configuration object. If calling this function directly as an argument to `daemons`, this is not required and can be inferred from the ‘url’ supplied (see ‘SSH Tunnelling’ section below).

Value

A list in the required format to be supplied to the ‘remote’ argument of launch_remote, daemons, or make_cluster.

SSH Direct Connections

The simplest use of SSH is to execute the daemon launch command on a remote machine, for it to dial back to the host / dispatcher URL.

It is assumed that SSH key-based authentication is already in place. The relevant port on the host must also be open to inbound connections from the remote machine.

SSH Tunnelling

Use of SSH tunnelling provides a convenient way to launch remote daemons without requiring the remote machine to be able to access the host. Often firewall configurations or security policies may prevent opening a port to accept outside connections.

In these cases SSH tunnelling offers a solution by creating a tunnel once the initial SSH connection is made. For simplicity, this SSH tunnelling implementation uses the same port on both the side of the host and that of the daemon. SSH key-based authentication must also already be in place.

Tunnelling requires the hostname for the ‘host’ argument (or the ‘url’ argument to daemons if called directly in that context) to be either ‘127.0.0.1’ or ‘localhost’. This is as the tunnel is created between 127.0.0.1:port or equivalently localhost:port on each machine. The host listens to port on its machine and the remotes each dial into port on their own respective machines.

Examples

# Slurm srun example
remote_config(
  command = "srun",
  args = c("--mem 512", "-n 1", "."),
  rscript = file.path(R.home("bin"), "Rscript")
)

# Slurm sbatch requires 'quote = TRUE'
remote_config(
  command = "sbatch",
  args = c("--mem 512", "-n 1", "--wrap", "."),
  rscript = file.path(R.home("bin"), "Rscript"),
  quote = TRUE
)

# SSH also requires 'quote = TRUE'
remote_config(
  command = "/usr/bin/ssh",
  args = c("-fTp 22 10.75.32.90", "."),
  quote = TRUE
)

# can be used to start local dameons with special configurations
remote_config(
  command = "Rscript",
  rscript = "--default-packages=NULL --vanilla"
)

# simple SSH example
ssh_config(
  remotes = c("ssh://10.75.32.90:222", "ssh://nodename"),
  timeout = 5
)

# SSH tunnelling example
ssh_config(
  remotes = c("ssh://10.75.32.90:222", "ssh://nodename"),
  tunnel = TRUE,
  host = "tls+tcp://127.0.0.1:5555"
)

## Not run: 

# launch 2 daemons on the remote machines 10.75.32.90 and 10.75.32.91 using
# SSH, connecting back directly to the host URL over a TLS connection:

daemons(
  url = host_url(tls = TRUE),
  remote = ssh_config(
    remotes = c("ssh://10.75.32.90:222", "ssh://10.75.32.91:222"),
    timeout = 1
  )
)

# launch 2 nodes on the remote machine 10.75.32.90 using SSH tunnelling over
# port 5555 ('url' hostname must be 'localhost' or '127.0.0.1'):

cl <- make_cluster(
  url = "tcp://localhost:5555",
  remote = ssh_config(
    remotes = c("ssh://10.75.32.90", "ssh://10.75.32.90"),
    tunnel = TRUE,
    timeout = 1
  )
)

## End(Not run)

# Slurm srun example
remote_config(
  command = "srun",
  args = c("--mem 512", "-n 1", "."),
  rscript = file.path(R.home("bin"), "Rscript")
)

# Slurm sbatch requires 'quote = TRUE'
remote_config(
  command = "sbatch",
  args = c("--mem 512", "-n 1", "--wrap", "."),
  rscript = file.path(R.home("bin"), "Rscript"),
  quote = TRUE
)

# SSH also requires 'quote = TRUE'
remote_config(
  command = "/usr/bin/ssh",
  args = c("-fTp 22 10.75.32.90", "."),
  quote = TRUE
)

# can be used to start local dameons with special configurations
remote_config(
  command = "Rscript",
  rscript = "--default-packages=NULL --vanilla"
)

# simple SSH example
ssh_config(
  remotes = c("ssh://10.75.32.90:222", "ssh://nodename"),
  timeout = 5
)

# SSH tunnelling example
ssh_config(
  remotes = c("ssh://10.75.32.90:222", "ssh://nodename"),
  tunnel = TRUE,
  host = "tls+tcp://127.0.0.1:5555"
)

## Not run: 

# launch 2 daemons on the remote machines 10.75.32.90 and 10.75.32.91 using
# SSH, connecting back directly to the host URL over a TLS connection:

daemons(
  url = host_url(tls = TRUE),
  remote = ssh_config(
    remotes = c("ssh://10.75.32.90:222", "ssh://10.75.32.91:222"),
    timeout = 1
  )
)

# launch 2 nodes on the remote machine 10.75.32.90 using SSH tunnelling over
# port 5555 ('url' hostname must be 'localhost' or '127.0.0.1'):

cl <- make_cluster(
  url = "tcp://localhost:5555",
  remote = ssh_config(
    remotes = c("ssh://10.75.32.90", "ssh://10.75.32.90"),
    tunnel = TRUE,
    timeout = 1
  )
)

## End(Not run)

Saisei (Regenerate Token)

Description

When using daemons with dispatcher, regenerates the token for the URL a dispatcher socket listens at.

Usage

saisei(i, force = FALSE, .compute = "default")
saisei(i, force = FALSE, .compute = "default")

Arguments

`i`	integer index number URL to regenerate at dispatcher.
`force`	[default FALSE] logical value whether to regenerate the URL even when there is an existing active connection.
`.compute`	[default 'default'] character value for the compute profile to use (each compute profile has its own independent set of daemons).

Details

When a URL is regenerated, the listener at the specified socket is closed and replaced immediately, hence this function will only be successful if there are no existing connections at the socket (i.e. ‘online’ status shows 0), unless the argument ‘force’ is specified as TRUE.

If ‘force’ is specified as TRUE, the socket is immediately closed and regenerated. If this happens while a mirai task is still ongoing, it will be returned as an ‘errorValue’ 7 (Object closed). This may be used to cancel a task that consistently hangs or crashes to prevent it from failing repeatedly when new daemons connect.

Value

The regenerated character URL upon success, or else NULL.

Timeouts

Specifying the ‘.timeout’ argument to mirai ensures that the mirai always resolves. However, the task may not have completed and still be ongoing in the daemon process. In such situations, dispatcher ensures that queued tasks are not assigned to the busy process, however overall performance may still be degraded if they remain in use.

If a process hangs and cannot be restarted otherwise, saisei specifying force = TRUE may be used to cancel the task and regenerate any particular URL for a new daemon to connect to.

Examples

if (interactive()) {
# Only run examples in interactive R sessions

daemons(1L)
Sys.sleep(1L)
status()
saisei(i = 1L, force = TRUE)
status()

daemons(0)

}

if (interactive()) {
# Only run examples in interactive R sessions

daemons(1L)
Sys.sleep(1L)
status()
saisei(i = 1L, force = TRUE)
status()

daemons(0)

}

Create Serialization Configuration

Description

Returns a serialization configuration, which may be set to perform custom serialization and unserialization of normally non-exportable reference objects, allowing these to be used seamlessly between different R sessions. This feature utilises the 'refhook' system of R native serialization. Once set, the functions apply to all mirai requests for a specific compute profile.

Usage

serial_config(class, sfunc, ufunc, vec = FALSE)
serial_config(class, sfunc, ufunc, vec = FALSE)

Arguments

`class`	character string of the class of object custom serialization functions are applied to, e.g. ‘ArrowTabular’ or ‘torch_tensor’.
`sfunc`	a function that accepts a reference object inheriting from ‘class’ (or a list of such objects) and returns a raw vector.
`ufunc`	a function that accepts a raw vector and returns a reference object (or list of such objects).
`vec`	[default FALSE] whether or not the serialization functions are vectorized. If FALSE, they should accept and return reference objects individually e.g. `arrow::write_to_raw` and `arrow::read_ipc_stream`. If TRUE, they should accept and return a list of reference objects, e.g. `torch::torch_serialize` and `torch::torch_load`.

Value

A list comprising the configuration. This should be passed to the ‘.serial’ argument of everywhere.

Examples

cfg <- serial_config("test_cls", function(x) serialize(x, NULL), unserialize)
cfg

cfg <- serial_config("test_cls", function(x) serialize(x, NULL), unserialize)
cfg

Status Information

Description

Retrieve status information for the specified compute profile, comprising current connections and daemons status.

Usage

status(.compute = "default")
status(.compute = "default")

Arguments

.compute

[default 'default'] character compute profile (each compute profile has its own set of daemons for connecting to different resources).

or a ‘miraiCluster’ to obtain its status.

Value

A named list comprising:

connections - integer number of active connections.
Using dispatcher: Always 1L as there is a single connection to dispatcher, which connects to the daemons in turn.
daemons - of variable type.
Using dispatcher: a status matrix (see Status Matrix section below), or else an integer ‘errorValue’ if communication with dispatcher failed.
Not using dispatcher: the character host URL.
Not set: 0L.

Status Matrix

When using dispatcher, $daemons comprises an integer matrix with the following columns:

i - integer index number.
online - shows as 1 when there is an active connection, or else 0 if a daemon has yet to connect or has disconnected.
instance - increments by 1 every time there is a new connection at a URL. This counter is designed to track new daemon instances connecting after previous ones have ended (due to time-outs etc.). The count becomes negative immediately after a URL is regenerated by saisei, but increments again once a new daemon connects.
assigned - shows the cumulative number of tasks assigned to the daemon.
complete - shows the cumulative number of tasks completed by the daemon.

The dispatcher URLs are stored as row names to the matrix.

Examples

if (interactive()) {
# Only run examples in interactive R sessions

status()
daemons(n = 2L, url = "wss://[::1]:0")
status()
daemons(0)

}

if (interactive()) {
# Only run examples in interactive R sessions

status()
daemons(n = 2L, url = "wss://[::1]:0")
status()
daemons(0)

}

mirai (Stop)

Description

Stops a ‘mirai’ if still in progress, causing it to resolve immediately to an ‘errorValue’ 20 (Operation canceled).

Usage

stop_mirai(x)
stop_mirai(x)

Arguments

`x`	a ‘mirai’ object, or list of ‘mirai’ objects.

Details

Forces the ‘mirai’ to resolve immediately. Has no effect if the ‘mirai’ has already resolved.

If cancellation was successful, the value at $data will be an ‘errorValue’ 20 (Operation canceled). Note that in such a case, the ‘mirai’ has been aborted and the value not retrieved - but any ongoing evaluation in the daemon process will continue to completion and is not interrupted.

Value

Invisible NULL.

Examples

if (interactive()) {
# Only run examples in interactive R sessions

m <- mirai(Sys.sleep(n), n = 5)
stop_mirai(m)
m$data

}

if (interactive()) {
# Only run examples in interactive R sessions

m <- mirai(Sys.sleep(n), n = 5)
stop_mirai(m)
m$data

}

Query if a mirai is Unresolved

Description

Query whether a ‘mirai’, ‘mirai’ value or list of ‘mirai’ remains unresolved. Unlike call_mirai, this function does not wait for completion.

Usage

unresolved(x)
unresolved(x)

Arguments

`x`	a ‘mirai’ object or list of ‘mirai’ objects, or a ‘mirai’ value stored at `$data`.

Details

Suitable for use in control flow statements such as while or if.

Note: querying resolution may cause a previously unresolved ‘mirai’ to resolve.

Value

Logical TRUE if ‘aio’ is an unresolved ‘mirai’ or ‘mirai’ value or the list contains at least one unresolved ‘mirai’, or FALSE otherwise.

Examples

if (interactive()) {
# Only run examples in interactive R sessions

m <- mirai(Sys.sleep(0.1))
unresolved(m)
Sys.sleep(0.3)
unresolved(m)

}

if (interactive()) {
# Only run examples in interactive R sessions

m <- mirai(Sys.sleep(0.1))
unresolved(m)
Sys.sleep(0.3)
unresolved(m)

}

With Mirai Daemons

Description

Evaluate an expression with daemons that last for the duration of the expression.

Usage

## S3 method for class 'miraiDaemons'
with(data, expr, ...)
## S3 method for class 'miraiDaemons'
with(data, expr, ...)

Arguments

`data`	a call to `daemons`.
`expr`	an expression to evaluate.
`...`	not used.

Details

This function is an S3 method for the generic with for class 'miraiDaemons'.

Value

The return value of ‘expr’.

Examples

if (interactive()) {
# Only run examples in interactive R sessions

with(
  daemons(2),
  {
    m1 <- mirai(Sys.getpid())
    m2 <- mirai(Sys.getpid())
    cat(call_mirai(m1)$data, call_mirai(m2)$data, "\n")
  }
)

status()

}

if (interactive()) {
# Only run examples in interactive R sessions

with(
  daemons(2),
  {
    m1 <- mirai(Sys.getpid())
    m2 <- mirai(Sys.getpid())
    cat(call_mirai(m1)$data, call_mirai(m2)$data, "\n")
  }
)

status()

}

Package 'mirai'

Help Index

mirai: Minimalist Async Evaluation Framework for R

Description

Notes

Reference Manual

Author(s)

See Also

Make Mirai Promise

Description

Usage

Arguments

Details

Value

Examples

mirai (Call Value)

Description

Usage

Arguments

Details

Value

Alternatively

Errors

Examples

mirai (Collect Value)

Description

Usage

Arguments

Details

Value

Alternatively

Errors

Examples

Daemon Instance

Description

Usage

Arguments

Details

Value

Persistence

Cleanup Options

Daemons (Set Persistent Processes)

Description

Usage

Arguments

Details

Value

Local Daemons

Dispatcher

Distributed Computing

Compute Profiles

Examples

Dispatcher

Description

Usage

Arguments

Details

Value

Evaluate Everywhere

Description

Usage

Arguments

Value

Evaluation

Examples

URL Constructors

Description

Usage

Arguments

Details

Value

Examples

Is mirai / mirai_map

Description

Usage

Arguments

Value

Examples

Error Validators

Description