Simple worker threads pool using Node's worker_threads module. Compatible with ES6+ Promise, Async/Await and TypeScript🚀.
- Use
StaticPoolto create a threads pool with a task from worker file or from task function provided to make use of multi-core processor. - Use
DynamicPoolto create a threads pool with different tasks provided each call. Thus you can get more flexibility thanStaticPooland make use of multi-core processor. - Gain extra controllability for the underlying threads by the power of worker_threads, like resourceLimits, SHARE_ENV, transferList and more.
- This module can only run in Node.js.
npm install node-worker-threads-pool --save
Quickly create a pool with static task:
const { StaticPool } = require('node-worker-threads-pool');
const staticPool = new StaticPool({
size: 4,
task: (n) => n + 1
});
staticPool.exec(1).then((result) => {
console.log('result from thread pool:', result); // result will be 2.
});There you go! 🎉
Create a pool with dynamic task:
const { DynamicPool } = require('node-worker-threads-pool');
const dynamicPool = new DynamicPool(4);
dynamicPool
.exec({
task: (n) => n + 1,
param: 1
})
.then((result) => {
console.log(result); // result will be 2.
});
dynamicPool
.exec({
task: (n) => n + 2,
param: 1
})
.then((result) => {
console.log(result); // result will be 3.
});About the differences between StaticPool and DynamicPool, please see this issue.
Instance of StaticPool is a threads pool with static task provided.
opt<Object>size<number>Number of workers in this pool.task<string | function>Static task to do. It can be a absolute path of worker file (usage here) or a function.⚠️ Notice: If task is a function, you can NOT access variables defined outside the task function! If you do want to use external data, use workerData to pass some cloneable data.workerData<any>Cloneable data you want to access in task function. (usage here)shareEnv<boolean>Settrueto enable SHARE_ENV for all threads in pool.resourceLimits<Object>Set resourceLimits for all threads in pool.
// Access the workerData by requiring it.
const { parentPort, workerData } = require('worker_threads');
// Something you shouldn"t run in main thread
// since it will block.
function fib(n) {
if (n < 2) {
return n;
}
return fib(n - 1) + fib(n - 2);
}
// Main thread will pass the data you need
// through this event listener.
parentPort.on('message', (param) => {
if (typeof param !== 'number') {
throw new Error('param must be a number.');
}
const result = fib(param);
// Access the workerData.
console.log('workerData is', workerData);
// return the result to main thread.
parentPort.postMessage(result);
});const { StaticPool } = require('node-worker-threads-pool');
const filePath = 'absolute/path/to/your/worker/script';
const pool = new StaticPool({
size: 4,
task: filePath,
workerData: 'workerData!'
});
for (let i = 0; i < 20; i++) {
(async () => {
const num = 40 + Math.trunc(10 * Math.random());
// This will choose one idle worker in the pool
// to execute your heavy task without blocking
// the main thread!
const res = await pool.exec(num);
console.log(`Fibonacci(${num}) result:`, res);
})();
}You can access workerData in task function using this keyword:
const pool = new StaticPool({
size: 4,
workerData: 'workerData!',
task() {
console.log(this.workerData);
}
});this.workerData, because arrow function don't have this binding.
param<any>The param your worker script or task function need.- Returns:
<Promise>
The simplest way to execute a task without considering other configurations. This will choose an idle worker in the pool to execute your heavy task with the param you provided. The Promise is resolved with the result.
- Returns:
<StaticTaskExecutor>
Create a task executor of this pool. This is used to apply some advanced settings to a task. See more details of StaticTaskExecutor.
Call worker.terminate() for every worker in the pool and release them.
Executor for StaticPool. Used to apply some advanced settings to a task.
const staticPool = new StaticPool({
size: 4,
task: (buf) => {
// do something with buf.
}
});
const buf = Buffer.alloc(1024 * 1024);
staticPool
.createExecutor() // create a StaticTaskExecutor instance.
.setTimeout(1000) // set timeout for task.
.setTransferList([buf.buffer]) // set transferList.
.exec(buf) // execute!
.then(() => console.log('done!'));t<number>timeout in millisecond.- Returns:
<StaticTaskExecutor>
Set timeout for this task.
transferList<Object[]>- Returns:
<StaticTaskExecutor>
Set transferList for this task. This is useful when you want to pass some huge data into worker thread.
param<any>- Returns:
<Promise>
Execute this task with the parameter and settings provided. The Promise is resolved with the result your task returned.
Instance of DynamicPool is a threads pool executes different task functions provided every call.
size<number>Number of workers in this pool.optshareEnv<boolean>Settrueto enable SHARE_ENV for every threads in pool.resourceLimits<Object>Set resourceLimits for all threads in pool.
opttask<function>Function as a task to do.⚠️ Notice: You can NOT access variables defined outside the task function!timeout<number>Timeout in milisecond for limiting the execution time. When timeout, the function will throw aTimeoutError, useisTimeoutErrorfunction to detect it.
- Returns:
<Promise>
Choose one idle worker in the pool to execute your task function. The Promise is resolved with the result your task returned.
taskFunctiontask function.- Returns:
<DynamicTaskExecutor>
Create a task executor of this pool. This is used to apply some advanced settings to a task. See more details of DynamicTaskExecutor.
Call worker.terminate() for every worker in the pool and release them.
Executor for DynamicPool. Used to apply some advanced settings to a task.
const dynamicPool = new DynamicPool(4);
const buf = Buffer.alloc(1024 * 1024);
dynamicPool
.createExecutor((buf) => {
// do something with buf.
})
.setTimeout(1000) // set timeout for task.
.setTransferList([buf.buffer]) // set transferList.
.exec(buf) // execute!
.then(() => console.log('done!'));t<number>timeout in millisecond.- Returns:
<DynamicTaskExecutor>
Set timeout for this task.
transferList<Object[]>- Returns:
<DynamicTaskExecutor>
Set transferList for this task. This is useful when you want to pass some huge data into worker thread.
param<any>- Returns:
<Promise>
Execute this task with the parameter and settings provided. The Promise is resolved with the result your task returned.
Detect if a thrown error is TimeoutError.
err <Error>The error you want to detect.- Returns
<boolean>trueif the error is aTimeoutError.
const { isTimeoutError } = require("node-worker-threads-pool");
// create pool.
...
// static pool exec with timeout.
const timeout = 1000;
try {
const res = await staticPool.exec(param, timeout);
} catch (err) {
if (isTimeoutError(err)) {
// deal with timeout.
} else {
// deal with other errors.
}
}
// dynamic pool exec with timeout.
const timeout = 1000;
try {
const res = await dynamicPool.exec({
task() {
// your task.
},
timeout
});
} catch (err) {
if (isTimeoutError(err)) {
// deal with timeout.
} else {
// deal with other errors.
}
}If you are using webpack in your project and want to import third-party libraries in task function, please use this.require:
const staticPool = new StaticPool({
size: 4,
task() {
const lib = this.require('lib');
// ...
}
});const dynamicPool = new DynamicPool(4);
dynamicPool
.exec({
task() {
const lib = this.require('lib');
// ...
}
})
.then((result) => {
// ...
});
![dependabot[bot] avatar](https://avatars.githubusercontent.com/in/29110?v=4 "dependabot[bot]")
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent