mirror of
https://github.com/nushell/nu_scripts.git
synced 2024-08-17 09:20:28 +03:00
Rewrite and expand background task module (#691)
Also renamed the module to task instead of job, https://www.nushell.sh/book/background_task.html will need updating. Other than that, this should work. --------- Co-authored-by: RGBCube <RGBCube@users.noreply.github.com>
This commit is contained in:
RGBCube
GitHub
parent
9fc4eb0ac5
commit
9700391563
@ -1,85 +1,65 @@
|
||||
# Background task with pueue
|
||||
make nushell "support" background task feature.
|
||||
# Background tasks with pueue
|
||||
|
||||
# Prerequisite
|
||||
Install [pueue](https://github.com/Nukesor/pueue) and make sure `pueued` is running with default config, and, `pueue` is in `PATH`
|
||||
# Usage
|
||||
```nushell
|
||||
use job.nu
|
||||
Makes Nushell "support" background tasks.
|
||||
|
||||
# spawn a new job in background.
|
||||
> job spawn { sleep 3sec ; echo gg | save gg.txt }
|
||||
## Prerequisite
|
||||
|
||||
# get job status
|
||||
> job status
|
||||
Install [pueue](https://github.com/Nukesor/pueue) and make sure `pueued` is running and that `pueue` is in `PATH`.
|
||||
|
||||
# get one job's log
|
||||
> job log $job_id
|
||||
## Usage
|
||||
|
||||
# kill one job
|
||||
> job kill $job_id
|
||||
You will get tab completions and suggestions when you install the module.
|
||||
Please check those.
|
||||
|
||||
# clean pueued finished job
|
||||
> job clean
|
||||
To install the module, copy the `task.nu` to the `$env.NU_LIB_DIRS` directory, then do:
|
||||
|
||||
```nu
|
||||
use task.nu
|
||||
```
|
||||
|
||||
# Q&A
|
||||
## How can I pass data to background job?
|
||||
In your Nushell config under `~/.config/nushell`.
|
||||
|
||||
The easiest way to do is pass it by environment variable, and you can use these variables in job.
|
||||
## Q&A
|
||||
|
||||
So don't do this, it doesn't work:
|
||||
```nushell
|
||||
let a = 3
|
||||
job spawn {
|
||||
echo $a
|
||||
### How can I pass data to a background task?
|
||||
|
||||
You can use environment variables, since they
|
||||
are inherited from the parent when spawning a process.
|
||||
|
||||
```nu
|
||||
$env.FOO = 123
|
||||
|
||||
task spawn {
|
||||
echo $env.FOO
|
||||
}
|
||||
```
|
||||
|
||||
Instead, doing this:
|
||||
```nushell
|
||||
let a = 3
|
||||
with-env {"a": 3} {
|
||||
job spawn {
|
||||
echo $env.a
|
||||
}
|
||||
If you want to pass serialized data, you can do this:
|
||||
|
||||
```nu
|
||||
let foo = { a: 1 b: 2 c: 3 }
|
||||
|
||||
with-env { FOO: ($foo | to json) } {
|
||||
task spawn {
|
||||
let foo = ($env.FOO | from json)
|
||||
|
||||
echo $foo
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
But what if I want to pass nushell's structure data to background job?
|
||||
## How can I reuse custom commands in a background task?
|
||||
|
||||
You can serialize it to env variable, and then deserialize it in background job.
|
||||
You can define these commands in a seperate module, like so:
|
||||
|
||||
```nushell
|
||||
let a = [1, 2, 3]
|
||||
with-env {"a": ($a | to json)} {
|
||||
job spawn {
|
||||
let a = ($env.a | from json)
|
||||
echo $a
|
||||
}
|
||||
```nu
|
||||
# --- in foo.nu ---
|
||||
export def bar [] { echo bar }
|
||||
|
||||
# --- in main.nu ---
|
||||
task spawn {
|
||||
use foo.nu
|
||||
|
||||
foo bar
|
||||
}
|
||||
```
|
||||
|
||||
## How can I reuse custom commands in background job?
|
||||
|
||||
You can defined these custom commands in a module, then you can use the module inside the job.
|
||||
|
||||
So don't do this, it won't work:
|
||||
```nushell
|
||||
def custom [] { echo 3 }
|
||||
job spawn {
|
||||
custom # custom is not defined
|
||||
}
|
||||
```
|
||||
|
||||
Instead, do this:
|
||||
```nushell
|
||||
# a.nu
|
||||
export def custom [] { echo 3 }
|
||||
|
||||
# main.nu
|
||||
job spawn {
|
||||
use a.nu
|
||||
a custom
|
||||
}
|
||||
```
|
||||
@ -1,78 +0,0 @@
|
||||
# spawn task to run in the background
|
||||
#
|
||||
# please note that a fresh nushell is spawned to execute the given command
|
||||
# So it doesn't inherit current scope's variables, custom commands, alias definition, except env variables which value can convert to string.
|
||||
#
|
||||
# Note that the closure to spawn can't take arguments. And it only supports something like this: { echo 3 }, it have no parameter list.
|
||||
#
|
||||
# e.g:
|
||||
# spawn { echo 3 }
|
||||
export def spawn [
|
||||
command: closure # the command to spawn
|
||||
--label(-l): string # the label of comand
|
||||
--group(-g): string # the group name to spawn
|
||||
] {
|
||||
let config_path = $nu.config-path
|
||||
let env_path = $nu.env-path
|
||||
let source_code = (view source $command | str trim -l -c '{' | str trim -r -c '}')
|
||||
mut args = [
|
||||
$"nu --config \"($config_path)\" --env-config \"($env_path)\" -c '($source_code)'",
|
||||
]
|
||||
if $label != null {
|
||||
$args = ($args | prepend ["--label", $label])
|
||||
}
|
||||
if $group != null {
|
||||
$args = ($args | prepend ["--group", $group])
|
||||
}
|
||||
let job_id = (pueue add -p $args)
|
||||
|
||||
{"job_id": $job_id}
|
||||
}
|
||||
|
||||
export def log [
|
||||
id: int # id to fetch log
|
||||
] {
|
||||
pueue log $id -f --json
|
||||
| from json
|
||||
| transpose -i info
|
||||
| flatten --all
|
||||
| flatten --all
|
||||
| flatten status
|
||||
}
|
||||
|
||||
# get job's stdout.
|
||||
export def output [
|
||||
id: int # id to fetch job's stdout
|
||||
] {
|
||||
log $id | get output.0
|
||||
}
|
||||
|
||||
# get job running status
|
||||
export def status [
|
||||
--detailed(-d) # need to get detailed stauts?
|
||||
] {
|
||||
let output = (
|
||||
pueue status --json
|
||||
| from json
|
||||
| get tasks
|
||||
| transpose -i status
|
||||
| flatten
|
||||
| flatten status
|
||||
)
|
||||
|
||||
if not $detailed {
|
||||
$output | select id label group Done? status? start? end?
|
||||
} else {
|
||||
$output
|
||||
}
|
||||
}
|
||||
|
||||
# kill specific job
|
||||
export def kill (id: int) {
|
||||
pueue kill $id
|
||||
}
|
||||
|
||||
# clean job log
|
||||
export def clean () {
|
||||
pueue clean
|
||||
}
|
||||
431
modules/background_task/task.nu
Normal file
431
modules/background_task/task.nu
Normal file
@ -0,0 +1,431 @@
|
||||
# Spawn a task to run in the background, even when the shell is closed.
|
||||
#
|
||||
# Note that a fresh Nushell interpreter is spawned to execute the
|
||||
# given task, so it won't inherit the current scope's variables,
|
||||
# custom commands and alias definitions.
|
||||
#
|
||||
# It will only inherit environment variables which can be converted to a string.
|
||||
#
|
||||
# Note that the closure can't take arguments.
|
||||
#
|
||||
# Example usage: task spawn { echo "Hello, World!" }
|
||||
export def spawn [
|
||||
command: closure # The closure to run.
|
||||
--working-directory (-w): directory # Specify the working directory the task will be run in.
|
||||
--immediate (-i) # Immediately start the task.
|
||||
--stashed (-s) # Create the task in Stashed state. Useful to avoid immediate execution if the queue is empty
|
||||
--delay (-d): duration # Queue the task for execution only after the duration.
|
||||
--group (-g): string # The group to spawn the task under.
|
||||
--after (-a): int # Start the task once all specified tasks have successfully finished. As soon as one of the dependencies fails, this task will fail as well.
|
||||
--priority (-o): string # Start this task with a higher priority. The higher the number, the faster it will be processed.
|
||||
--label (-l): string # Label the task. This string will be shown in the `status` column of `task status`.
|
||||
] -> int {
|
||||
mut args = []
|
||||
|
||||
if $working_directory != null {
|
||||
$args = ($args | prepend ["--working-directory", $working_directory])
|
||||
}
|
||||
if $immediate {
|
||||
$args = ($args | prepend "--immediate")
|
||||
}
|
||||
if $stashed {
|
||||
$args = ($args | prepend "--stashed")
|
||||
}
|
||||
if $delay != null {
|
||||
$args = ($args | prepend ["--delay" ($delay | format duration sec | parse "{secs} {_}" | get 0.secs)])
|
||||
}
|
||||
if $group != null {
|
||||
$args = ($args | prepend ["--group" $group])
|
||||
}
|
||||
if $after != null {
|
||||
$args = ($args | prepend ["--after" $after])
|
||||
}
|
||||
if $priority != null {
|
||||
$args = ($args | prepend ["--priority" $priority])
|
||||
}
|
||||
if $label != null {
|
||||
$args = ($args | prepend ["--label" $label])
|
||||
}
|
||||
|
||||
let source_code = (
|
||||
view source $command
|
||||
| str trim --left --char "{"
|
||||
| str trim --right --char "}"
|
||||
)
|
||||
|
||||
(pueue add --print-task-id $args $"nu --config '($nu.config-path)' --env-config '($nu.env-path)' --commands '($source_code)'")
|
||||
}
|
||||
|
||||
# Remove tasks from the queue.
|
||||
# Running or paused tasks need to be killed first.
|
||||
export def remove [
|
||||
...ids: int # IDs of the tasks to remove from the status list.
|
||||
] {
|
||||
pueue remove $ids
|
||||
}
|
||||
|
||||
# Switches the queue position of two tasks.
|
||||
# Only works for queued or stashed tasks.
|
||||
export def switch [
|
||||
task_id_1: int # The first task ID.
|
||||
task_id_2: int # The second task ID.
|
||||
] {
|
||||
pueue switch $task_id_1 $task_id_2
|
||||
}
|
||||
|
||||
# Stash a task that is not currently running.
|
||||
#
|
||||
# Stashed tasks won't be automatically started.
|
||||
# You will have to queue them or start them by hand.
|
||||
export def stash [
|
||||
...ids: int # IDs of the tasks to stash.
|
||||
] {
|
||||
pueue stash $ids
|
||||
}
|
||||
|
||||
# Queue stashed tasks for execution.
|
||||
export def queue [
|
||||
...ids: int # IDs of the tasks to queue.
|
||||
--delay (-d): duration # Queue only after the specified delay.
|
||||
] {
|
||||
let args = if $delay != null {
|
||||
["--delay" ($delay | format duration sec | parse '{secs} {_}' | get 0.secs)]
|
||||
} else {
|
||||
[]
|
||||
}
|
||||
|
||||
pueue enqueue $args $ids
|
||||
}
|
||||
|
||||
# Resume operation of specific tasks or groups of tasks.
|
||||
#
|
||||
# By default, this resumes the default group and all its tasks.
|
||||
# It can also be used force-start specific tasks or start whole groups.
|
||||
export def start [
|
||||
...ids: int # IDs of the tasks to start. By default all the tasks in the default group will be started.
|
||||
--group (-g): string # Resume a specific group and all paused tasks in it. The group will be set to running and its paused tasks will be resumed.
|
||||
--all (-a) # Resume all groups. All groups will be set to running and paused tasks will be resumed.
|
||||
] {
|
||||
mut args = []
|
||||
|
||||
if $group != null {
|
||||
$args = ($args | prepend ["--group" $group])
|
||||
}
|
||||
if $all {
|
||||
$args = ($args | prepend "--all")
|
||||
}
|
||||
|
||||
pueue start $args
|
||||
}
|
||||
|
||||
# Restart failed or successful task(s).
|
||||
#
|
||||
# By default, identical tasks will be created and
|
||||
# enqueued, but it's possible to restart in-place.
|
||||
#
|
||||
# You can also edit a few properties, such as
|
||||
# the path and the command of the task, before restarting.
|
||||
export def restart [
|
||||
...ids: int # IDs of the tasks to restart.
|
||||
--all-failed (-a) # Restart all failed tasks across all groups. Nice to use in combination with `--in-place/i`.
|
||||
--failed-in-group (-g): string # Like `--all-failed`, but only restart tasks failed tasks of a specific group. The group will be set to running and its paused tasks will be resumed.
|
||||
--start-immediately (-k) # Immediately start the tasks, no matter how many open slots there are. This will ignore any dependencies tasks may have.
|
||||
--stashed (-s) # Set the restarted task to a "Stashed" state. Useful to avoid immediate execution.
|
||||
--in-place (-i) # Restart the tasks by reusing the already existing tasks.
|
||||
--not-in-place (-n) # Opposite of `--in-place`. This is already the default unless you have `restart_in_place` set to true.
|
||||
--edit (-e) # Edit the tasks' commands before restarting
|
||||
--edit-path (-p) # Edit the tasks' paths before restarting
|
||||
--edit-label (-l) # Edit the tasks' labels before restarting
|
||||
] {
|
||||
mut args = []
|
||||
|
||||
if $all_failed {
|
||||
$args = ($args | prepend "--all-failed")
|
||||
}
|
||||
if $failed_in_group != null {
|
||||
$args = ($args | prepend "--failed-in-group")
|
||||
}
|
||||
if $start_immediately {
|
||||
$args = ($args | prepend "--start-immediately")
|
||||
}
|
||||
if $stashed {
|
||||
$args = ($args | prepend "--stashed")
|
||||
}
|
||||
if $in_place {
|
||||
$args = ($args | prepend "--in-place")
|
||||
}
|
||||
if $not_in_place {
|
||||
$args = ($args | prepend "--not-in-place")
|
||||
}
|
||||
if $edit {
|
||||
$args = ($args | prepend "--edit")
|
||||
}
|
||||
if $edit_path {
|
||||
$args = ($args | prepend "--edit-path")
|
||||
}
|
||||
if $edit_label {
|
||||
$args = ($args | prepend "--edit-label")
|
||||
}
|
||||
|
||||
pueue restart $args $ids
|
||||
}
|
||||
|
||||
# Either pause a running tasks or a specific groups of tasks.
|
||||
#
|
||||
# By default, pauses the default group and all its tasks.
|
||||
#
|
||||
# A paused group won't start any new tasks automatically.
|
||||
export def pause [
|
||||
...ids: int # IDs of the tasks to pause.
|
||||
--group (-g) # Pause a specific group
|
||||
--all (-a) # Pause all groups.
|
||||
--wait (-w) # Only pause the specified group and let already running tasks finish by themselves
|
||||
] {
|
||||
mut args = []
|
||||
|
||||
if $group != null {
|
||||
$args = ($args | prepend "--group")
|
||||
}
|
||||
if $all != null {
|
||||
$args = ($args | prepend "--all")
|
||||
}
|
||||
if $wait != null {
|
||||
$args = ($args | prepend "--wait")
|
||||
}
|
||||
|
||||
pueue pause $args $ids
|
||||
}
|
||||
|
||||
# Kill specific running tasks or whole task groups.
|
||||
#
|
||||
# Kills all tasks of the default group when no ids or a specific group are provided.
|
||||
export def kill [
|
||||
...ids: int # IDs of the tasks to kill.
|
||||
--group (-g): string # Kill all running tasks in a group. This also pauses the group.
|
||||
--all (-a) # Kill all running tasks across ALL groups. This also pauses all groups.
|
||||
--signal (-s): string # Send a UNIX signal instead of simply killing the process. DISCLAIMER: This bypasses Pueue's process handling logic! You might enter weird invalid states, use at your own descretion.
|
||||
] {
|
||||
mut args = []
|
||||
|
||||
if $group != null {
|
||||
$args = ($args | prepend ["--group" $group])
|
||||
}
|
||||
if $all {
|
||||
$args = ($args | prepend "--all")
|
||||
}
|
||||
if $signal != null {
|
||||
$args = ($args | prepend ["--signal" $signal])
|
||||
}
|
||||
|
||||
pueue kill $args $ids
|
||||
}
|
||||
|
||||
# Send something to a task. Useful for sending confirmations such as "y\n".
|
||||
export def send [
|
||||
id: int # ID of the task to send something to.
|
||||
input: string # The input that should be sent to the process.
|
||||
] {
|
||||
pueue send $id $input
|
||||
}
|
||||
|
||||
# Edit the command, path or label of a stashed or queued task.
|
||||
#
|
||||
# By default only the command is edited.
|
||||
#
|
||||
# Multiple properties can be added in one go.
|
||||
export def edit [
|
||||
id: int # ID of the task to edit.
|
||||
--command (-c) # Edit the task's command
|
||||
--path (-p) # Edit the task's path
|
||||
--label (-l) # Edit the task's label
|
||||
] {
|
||||
mut args = []
|
||||
|
||||
if $command {
|
||||
$args = ($args | prepend "--command")
|
||||
}
|
||||
if $path {
|
||||
$args = ($args | prepend "--path")
|
||||
}
|
||||
if $label {
|
||||
$args = ($args | prepend "--label")
|
||||
}
|
||||
|
||||
pueue edit $args $id
|
||||
}
|
||||
|
||||
# Use this to add or remove groups.
|
||||
#
|
||||
# By default, this will simply display all known groups.
|
||||
export def group [] {
|
||||
pueue group --json | from json
|
||||
}
|
||||
|
||||
# Create a new group with a name.
|
||||
export def "group add" [
|
||||
name: string # The name of the new group.
|
||||
--parallel (-p): int # The amount of parallel tasks the group can run at one time.
|
||||
] {
|
||||
let args = if $parallel != null {
|
||||
["--parallel" $parallel]
|
||||
} else {
|
||||
[]
|
||||
}
|
||||
|
||||
pueue group add $args $name
|
||||
}
|
||||
|
||||
# Remove a group with a name.
|
||||
export def "group remove" [
|
||||
name: string # The name of the group to be removed.
|
||||
] {
|
||||
pueue group remove $name
|
||||
}
|
||||
|
||||
# Display the current status of all tasks.
|
||||
export def status [
|
||||
--detailed (-d) # Return a table with more detailed information.
|
||||
] {
|
||||
let output = (
|
||||
pueue status --json
|
||||
| from json
|
||||
| get tasks
|
||||
| transpose --ignore-titles status
|
||||
| flatten
|
||||
)
|
||||
|
||||
# TODO: Rename the Done column to done.
|
||||
if not $detailed {
|
||||
$output | select id label group Done? status? start? end?
|
||||
} else {
|
||||
$output
|
||||
}
|
||||
}
|
||||
|
||||
# Display the log output of finished tasks.
|
||||
#
|
||||
# Only the last few lines will be shown by default for multiple tasks.
|
||||
# If you want to follow the output, use `--tail/-t`.
|
||||
export def log [
|
||||
...ids: int # The tasks to check the outputs of.
|
||||
--last (-l): int # Only print the last N lines of each task's output. This is done by default if you're looking at multiple tasks.
|
||||
--tail (-t) # Follow the output as it is printing. Only works with 1 task. When used in conjunction with `--last`, the last N lines will be printed before starting to wait for output.
|
||||
--detailed (-d) # Include all fields, don't simplify output.
|
||||
] {
|
||||
def process_raw [raw: string] {
|
||||
let full = (
|
||||
$raw
|
||||
| from json
|
||||
| transpose -i info
|
||||
| flatten --all
|
||||
| flatten --all
|
||||
)
|
||||
|
||||
if $detailed {
|
||||
$full
|
||||
} else {
|
||||
$full | select id label group Done? status? start? end?
|
||||
}
|
||||
}
|
||||
|
||||
if (($ids | length) == 1) {
|
||||
if $tail {
|
||||
let args = if $last != null {
|
||||
["--lines" $last]
|
||||
} else {
|
||||
[]
|
||||
}
|
||||
|
||||
pueue follow $ids
|
||||
} else {
|
||||
let args = if $last != null {
|
||||
["--lines" $last]
|
||||
} else {
|
||||
[]
|
||||
}
|
||||
|
||||
process_raw (pueue log --full --json $args $ids)
|
||||
| first
|
||||
}
|
||||
} else {
|
||||
if $tail {
|
||||
echo $"(ansi red)--tail can only be used with one task.(ansi reset)"
|
||||
return
|
||||
}
|
||||
|
||||
let args = if $last != null {
|
||||
["--lines" $last]
|
||||
} else {
|
||||
[]
|
||||
}
|
||||
|
||||
process_raw (pueue log --full --json $args $ids)
|
||||
}
|
||||
}
|
||||
|
||||
# Wait until the provided tasks are finished.
|
||||
#
|
||||
# This is like join() or await in many languages.
|
||||
export def wait [
|
||||
...ids: int # IDs of the tasks to wait for.
|
||||
--group (-g): string # Wait for all tasks in a specific group.
|
||||
--all (-a) # Wait for all tasks across all groups and the default group.
|
||||
--quiet (-q) # Don't show any log output while waiting.
|
||||
--status (-s): string # Wait for tasks to reach a specific task status.
|
||||
] {
|
||||
mut args = []
|
||||
|
||||
if $group != null {
|
||||
$args = ($args | prepend ["--group" $group])
|
||||
}
|
||||
if $all {
|
||||
$args = ($args | prepend $all)
|
||||
}
|
||||
if $quiet {
|
||||
$args = ($args | prepend $quiet)
|
||||
}
|
||||
if $status != null {
|
||||
$args = ($args | prepend ["--status" $status])
|
||||
}
|
||||
|
||||
pueue wait $args $ids
|
||||
}
|
||||
|
||||
# Remove tasks from the status list.
|
||||
export def clean [
|
||||
--successful-only (-s) # Only clean tasks that finished successfully
|
||||
--group (-g): string # Only clean tasks of a specific group
|
||||
] {
|
||||
mut args = []
|
||||
|
||||
if $successful_only {
|
||||
$args = ($args | prepend "--successful-only")
|
||||
}
|
||||
if $group != null {
|
||||
$args = ($args | prepend ["--group" $group])
|
||||
}
|
||||
|
||||
pueue clean $args
|
||||
}
|
||||
|
||||
# Shutdown pueue and thus this module.
|
||||
export def shutdown [] {
|
||||
pueue shutdown
|
||||
}
|
||||
|
||||
# Set the maximum parallel tasks for a group.
|
||||
#
|
||||
# Note that no tasks will be stopped if the number is lowered.
|
||||
# The limit only applies when schelduing.
|
||||
export def set-parallel-limit [
|
||||
max: int # The maximum parallel tasks allowed for a group when schelduing.
|
||||
--group (-g): string # The group to set the limit for. By default this is `default`.
|
||||
] {
|
||||
let args = if $group != null {
|
||||
["--group" $group]
|
||||
} else {
|
||||
[]
|
||||
}
|
||||
|
||||
pueue parallel $args $max
|
||||
}
|
||||
Loading…
Reference in New Issue
Block a user