https://github.com/kumavis/obj-multiplex

obj-multiplex多路复用

simple stream multiplexing for objectMode

其实就是一个多路复用流能够使用name来区分各个子流,以达到一个parent流下其实有多个子流在运行,可以通过多个子流来读入写出数据,效率更高。而且parent流结束了,则所有子流也会被销毁

usage

// create multiplexer
const mux = new ObjMultiplex() // setup substreams
const streamA = mux.createStream('hello')
const streamB = mux.createStream('world') // pipe over transport (and back)
mux.pipe(transport).pipe(mux) // send values over the substreams
streamA.write({ thisIsAn: 'object' })
streamA.write() // or pipe together normally
streamB.pipe(evilAiBrain).pipe(streamB)

obj-multiplex/index.js

const { Duplex } = require('readable-stream')
const endOfStream = require('end-of-stream')//看博客mafintosh/end-of-stream
const once = require('once')
const noop = () => {} const IGNORE_SUBSTREAM = {} class ObjectMultiplex extends Duplex { constructor(_opts = {}) {
const opts = Object.assign({}, _opts, {
objectMode: true,//流可传各类形式数据
})
super(opts)//生成这个流 this._substreams = {}
} createStream (name) {//就是创建两个流,一个是这个流,另一个是parent是这个流的一个子流,并返回子流
// validate name
if (!name) throw new Error('ObjectMultiplex - name must not be empty')//name不能为空
if (this._substreams[name]) throw new Error('ObjectMultiplex - Substream for name "${name}" already exists')//name不能重复 // create substream
const substream = new Substream({ parent: this, name: name })
this._substreams[name] = substream // listen for parent stream to end
anyStreamEnd(this, (err) => {//定义当parent流结束,则相应的所有子流也要被销毁
substream.destroy(err)//substream被destroy,如果出错返回的错误信息即err
}) return substream
} // ignore streams (dont display orphaned data warning)
ignoreStream (name) {//就是将之前创建的name的子流的内容清空
// validate name
if (!name) throw new Error('ObjectMultiplex - name must not be empty')
if (this._substreams[name]) throw new Error('ObjectMultiplex - Substream for name "${name}" already exists')
// set
this._substreams[name] = IGNORE_SUBSTREAM
} // stream plumbing
//下面就是parent流能够做的一系列读写操作
_read () {} _write(chunk, encoding, callback) {//当调用 writable.write(chunk) 时,数据会被缓冲在可写流中。
// parse message,就是当parent流write时,将根据其传入的name来决定该数据是写到哪个子流上的
const name = chunk.name
const data = chunk.data
if (!name) {//name不能为空,否则不知道是哪个子流
console.warn(`ObjectMultiplex - malformed chunk without name "${chunk}"`)
return callback()
} // get corresponding substream
const substream = this._substreams[name]//然后根据name得到子流
if (!substream) {//如果为空则warn
console.warn(`ObjectMultiplex - orphaned data for stream "${name}"`)
return callback()
} // push data into substream
if (substream !== IGNORE_SUBSTREAM) {//只有当子流不为{}时,才将data压入
substream.push(data)//当调用 stream.push(chunk) 时,数据会被缓冲在可读流中。 如果流的消费者没有调用 stream.read(),则数据会保留在内部队列中直到被消费
} callback()
}//_write }//class class Substream extends Duplex { constructor ({ parent, name }) {
super({
objectMode: true,
}) this._parent = parent
this._name = name
} _read () {}//读入的操作即Duplex的定义 _write (chunk, enc, callback) {//当子流被写入时,其实是将数据压入流parent中
this._parent.push({
name: this._name,
data: chunk,
})
callback()//然后调用回调函数
} } module.exports = ObjectMultiplex // util function anyStreamEnd(stream, _cb) {//就是当stream结束的时候就会调用cb回调函数
const cb = once(_cb)
endOfStream(stream, { readable: false }, cb)
endOfStream(stream, { writable: false }, cb)
}

通过测试学习该库的使用:

obj-multiplex/test/index.js

const test = require('tape')
const once = require('once')
const { PassThrough, Transform } = require('readable-stream')//PassThrough本质也是Transform流,是最简单的Transform流,只是将数据从此处传过
// a passthrough stream.
// basically just the most minimal sort of Transform stream.
// Every written chunk gets output as-is
const endOfStream = require('end-of-stream')
const pump = require('pump')
const ObjMultiplex = require('../index.js') test('basic - string', (t) => {
t.plan() const {
inTransport, outTransport,
inMux, outMux,
inStream, outStream,
} = basicTestSetup() bufferToEnd(outStream, (err, results) => {
t.error(err, 'should not error')
t.deepEqual(results, ['haay', 'wuurl'], 'results should match')
t.end()
}) // pass in messages
inStream.write('haay')
inStream.write('wuurl') // simulate disconnect
setTimeout(() => inTransport.destroy())
}) test('basic - obj', (t) => {
t.plan() const {
inTransport, outTransport,
inMux, outMux,
inStream, outStream,
} = basicTestSetup() bufferToEnd(outStream, (err, results) => {
t.error(err, 'should not error')
t.deepEqual(results, [{ message: 'haay' }, { message: 'wuurl' }], 'results should match')
t.end()
}) // pass in messages
inStream.write({ message: 'haay' })
inStream.write({ message: 'wuurl' }) // simulate disconnect
setTimeout(() => inTransport.destroy())
}) test('roundtrip', (t) => {
t.plan() const {
inTransport, outTransport,
inMux, outMux,
inStream, outStream,
} = basicTestSetup() const doubler = new Transform({
objectMode: true,
transform (chunk, end, callback) {//对流内数据进行*2计算
// console.log('doubler!', chunk)
const result = chunk *
callback(null, result)
}
}) pump(//即将从outStream处得到的数据进行*2处理后再传回outStream
outStream,
doubler,
outStream
) bufferToEnd(inStream, (err, results) => {
t.error(err, 'should not error')
t.deepEqual(results, [, ], 'results should match')
t.end()
}) // pass in messages
inStream.write()
inStream.write() // simulate disconnect
setTimeout(() => outTransport.destroy(), )
}) // util function basicTestSetup() { // setup multiplex and Transport
const inMux = new ObjMultiplex()//定义了两个parent流
const outMux = new ObjMultiplex()
const inTransport = new PassThrough({ objectMode: true })
const outTransport = new PassThrough({ objectMode: true }) // setup substreams
const inStream = inMux.createStream('hello')//分别在两个parent流中各自定义一个name为hello的子流
const outStream = outMux.createStream('hello') pump(//形成一个pipe流
inMux,
inTransport,
outMux,
outTransport,
inMux
) return {
inTransport, outTransport,
inMux, outMux,
inStream, outStream,
} } function bufferToEnd(stream, callback) {
const results = []
endOfStream(stream, (err) => callback(err, results))//定义了流结束后的回调
stream.on('data', (chunk) => results.push(chunk))//并监听data事件,用于将数据压入流
}

kumavis/obj-multiplex的更多相关文章

  1. obj.style.z-index的正确写法

    obj.style.z-index的正确写法 今天发现obj.style.z-index在js里面报错,后来才知道在js里应该把含"-"的字符写成驼峰式,例如obj.style.z ...

  2. CSharpGL(9)解析OBJ文件并用CSharpGL渲染

    CSharpGL(9)解析OBJ文件并用CSharpGL渲染 2016-08-13 由于CSharpGL一直在更新,现在这个教程已经不适用最新的代码了.CSharpGL源码中包含10多个独立的Demo ...

  3. BeanUtils.populate(obj, map);

    public static void populate(Object bean, Map<String, ? extends Object> properties) throws Ille ...

  4. VC++ : error LNK2005: ... already defined in *.obj

    今天写代码遇到了这么一个链接错误:"已经在*.obj中定义". error LNK2005: "void __cdecl ReplaceWstringVar(class ...

  5. 111. for(元素变量x:遍历对象obj)

    package com.chongrui.test;/* * for(元素变量x:遍历对象obj){ * 引用X的java语句 *  * } *  *  * */public class test { ...

  6. 让Git忽略所有obj和bin目录的同步

    DotNet的项目里,编译出来的二进制文件默认都是放在每个项目对应的bin和obj目录下,有时候开发人员会不小心把这些目录的文件错误的提交到Git服务器.Git里的忽略文件功能可以通过一个.gitig ...

  7. make: *** [out/host/linux-x86/obj/EXECUTABLES/aidl_intermediates/aidl] 错误 1,make: *** [out/host/linux-x86/obj/lib/libESR_Portable.so] 错误 1

    错误3: g++: g++: selected multilib '32' not installed selected multilib '32' not installed make: *** [ ...

  8. obj转json

    .js模型(JSON)的获取 方法一: 1.安装Python插件.安装完后配置环境变量,path中添加Python路径. 2.找到three.js\utils\converters\obj\conve ...

  9. three.js加载obj模型

    <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xht ...

随机推荐

  1. XML部分

    XML文档定义有几种形式?它们之间有何本质区别?解析XML文档有哪几种方式? 两种形式:DTD以及schema: 本质区别:schema本身是xml的,可以被XML解析器解析(这也是从DTD上发展sc ...

  2. MYSQL中SHOW的使用整理收藏

    好记性不如乱笔头吧....下面收藏整理了mysql中show 的使用技巧....有需要的博友可以看看哈 a. show tables或show tables from database_name; / ...

  3. webstrom vue项目让局域网访问

    vue项目package.json "dev": "webpack-dev-server --inline --progress --config build/webpa ...

  4. NIO学习笔记四 :SocketChannel 和 ServerSocketChannel

    Java NIO中的SocketChannel是一个连接到TCP网络套接字的通道.可以通过以下2种方式创建SocketChannel: 打开一个SocketChannel并连接到互联网上的某台服务器. ...

  5. HDU5887(SummerTrainingDay01-D)

    Herbs Gathering Time Limit: 3000/1000 MS (Java/Others)    Memory Limit: 32768/32768 K (Java/Others)T ...

  6. blfs(systemv版本)学习笔记-编译安装openssh软件包

    我的邮箱地址:zytrenren@163.com欢迎大家交流学习纠错! openssh项目地址:http://www.linuxfromscratch.org/blfs/view/8.3/postlf ...

  7. HTML5本地存储localStorage与sessionStorage详解

    前言 在最近的项目中用到了html5的本地存储,下面总结一下. 1.html5几种存储形式 本地存储(localStorage && sessionStorage) 离线缓存(appl ...

  8. 【读书笔记】iOS-设计模式

    一个可复用的解决方案,用于处理特定场景下的常见问题.一种设计模式并不是一个可以直接转化为代码的已完工设计.它是对于如何解决问题的一种描述或者模板,可以用在许多不同的场合. 参考资料:<iOS W ...

  9. sqoop简单配置与使用

    sqoop(sql-to-hadoop) Apache Sqoop是用来实现结构型数据(如关系数据库)和Hadoop之间进行数据迁移的工具.它充分利用了MapReduce的并行特点以批处理的方式加快数 ...

  10. angular升级指南

    废话少说 上链接 angular 升级指南