testing performance of parsing xml/html with nodejs

performance-parser.js

  //
 // this is about parsing-performance of xml/html-parsers
//  to be more precisely, i just want to look for specific nodes/attributes
//
// i am testing
//
// * (htmlparser)[https://github.com/tautologistics/node-htmlparser]
// * (html5)[https://github.com/aredridel/html5]
// * (sax)[https://github.com/isaacs/sax-js]
// * (jsdom)[https://github.com/tmpvar/jsdom] + sizzle/jquery
//
// the output i get is:
//
//     htmlparser done in 73 ms - memory: 2.09375 mb RSS - found 200 items
//     html5 done in 1727 ms - memory: 15.60546875 mb RSS - found 0 items (TODO)
//     sax done in 100 ms - memory: 0.0390625 mb RSS - found 200 items
//     jsdom/jquery done in 728 ms - memory: 0.1171875 mb RSS - found 200 items

var Seq     = require('seq')
  , Step    = require('step')
  , events  = require('events')
  , request = require('request')
  , saxLib  = require('sax')
  , html    = require('htmlparser')
  , html5   = require('html5')
  , jsdom   = require('jsdom')
  , uri     = 'http://twitter.com/statuses/user_timeline/18975861.rss'
  , lookfor = 'item'
  , n       = 10

//------------------------------------------------ request xml and test all

request({uri:uri}, function(err, res, body) {
  Seq()
    .seq(function(){testHtml(body, this)})
    .seq(function(){testHtml5(body, this)})
    .seq(function(){testSax(body, this)})
    .seq(function(){testJsdom(body, this)})
    .seq(function(){console.log('all done')}) 
})
  
//------------------------------------------------ tautologistic's htmlparser

function testHtml(body, cb) {
  var htmlHandler, htmlParser
    , t = Date.now()
    , m = process.memoryUsage().rss
    , done = 0
    , items = 0
  
  htmlHandler = new html.DefaultHandler(function(err, dom) {
    function walkDom(dom) {
      for (var i=0, len=dom.length; i<len; i++) {
        if (dom[i].type == 'tag' && dom[i].name == lookfor) items++
        if (dom[i].children && dom[i].children.length) 
          walkDom(dom[i].children)
      }
    }
    walkDom(dom)
    if (++done==n) {
      console.log
        ( 'htmlparser done in %s ms - memory: %s mb RSS - found %s items'
        , (Date.now())-t, (process.memoryUsage().rss-m)/1048576, items )
      cb()
    } else {
      htmlParser.parseComplete(body)
    }
  })
  htmlParser = new html.Parser(htmlHandler)
  htmlParser.parseComplete(body)
}

//------------------------------------------------ aredridel's html5

function testHtml5(body, cb) {
  var parser = new html5.Parser()
    , em = new events.EventEmitter()
    , t = Date.now()
    , m = process.memoryUsage().rss
    , done = 0
    , items = 0
    
  parser.on('done', function() { // i guess this is not done yet? altough its in the doc
    console.log('html5-done')
  })
  
  parser.parse(em)
  while(++done <= 10) {
    em.emit('data', body)
    em.emit('end')
  }
  console.log
    ( 'html5 done in %s ms - memory: %s mb RSS - found %s items (TODO)'
    , (Date.now())-t, (process.memoryUsage().rss-m)/1048576, items )
  cb()
}

//------------------------------------------------ isaac's sax-parser

function testSax(body, cb) {
  var sax = saxLib.parser(true)
    , t = Date.now()
    , m = process.memoryUsage().rss
    , done = 0
    , items = 0
    
  sax.onerror = function(err) {console.log(err)}
  sax.onopentag = function(node) {
    if (node.name == lookfor) items++
  } 
  sax.onend = function() {
    if (++done==n) {
      console.log
        ( 'sax done in %s ms - memory: %s mb RSS - found %s items'
        , (Date.now())-t, (process.memoryUsage().rss-m)/1048576, items )
      cb()
    } 
  }
  while (done<n) sax.write(body).close()
}

//------------------------------------------------ tmpvar's jsdom + sizzle/jquery

function testJsdom(body, cb) {
  var window = jsdom.jsdom().createWindow()
  jsdom.jQueryify( window
                 , 'http://code.jquery.com/jquery.min.js' 
                 , function() {
    var done = 0
      , t = Date.now()
      , m = process.memoryUsage().rss
      , items = 0
    
    while(++done <= n) {
      window.$('body').html(body)
      items += window.$(lookfor).length
    }
    console.log
      ( 'jsdom/jquery done in %s ms - memory: %s mb RSS - found %s items'
      , (Date.now())-t, (process.memoryUsage().rss-m)/1048576, items )
    cb()
  }) 
}



  

which version of jsdom are you using? I would try HEAD

with HEAD (6019785d) i get

sax done in 117 ms - memory: 1.96484375 mb RSS
htmlparser done in 47 ms - memory: 0.5703125 mb RSS
jsdom/jquery done in 924 ms - memory: 5.82421875 mb RSS   

good job! :D

anyway this test/code is very incomplete and may not be worth anything - just wanted to make a snapshot

cool, looks like I managed to chop off 1/3 of your execution time! I cant went to spend some real time optimizing.