Discover missinglink/huntsman Open Source project

Install

npm install huntsman --save

Example Script

/** Crawl wikipedia and use jquery syntax to extract information from the page **/

var huntsman = require('huntsman');
var spider = huntsman.spider();

spider.extensions = [
  huntsman.extension( 'recurse' ), // load recurse extension & follow anchor links
  huntsman.extension( 'cheerio' ) // load cheerio extension
];

// follow pages which match this uri regex
spider.on( /http:\/\/en\.wikipedia\.org\/wiki\/\w+:\w+$/, function ( err, res ){

  // use jquery-style selectors & functions
  var $ = res.extension.cheerio;
  if( !$ ) return; // content is not html

  // extract information from page body
  var wikipedia = {
    uri: res.uri,
    heading: $('h1.firstHeading').text().trim(),
    body: $('div#mw-content-text p').text().trim()
  };

  console.log( wikipedia );

});

spider.queue.add( 'http://en.wikipedia.org/wiki/Huntsman_spider' );
spider.start();

Example Output

peter@edgy:/tmp$ node examples/html.js 
{
  "uri": "http://en.wikipedia.org/wiki/Wikipedia:Recent_additions",
  "heading": "Wikipedia:Recent additions",
  "body": "This is a selection of recently created new articles and greatly expanded former stub articles on Wikipedia that were featured on the Main Page as part of Did you know? You can submit new pages for consideration. (Archives are grouped by month of Main page appearance.)Tip: To find which archive contains the fact that appeared on Did You Know?, return to the article and click \"What links here\" to the left of the article. Then, in the dropdown menu provided for namespace, choose Wikipedia and click \"Go\". When you find \"Wikipedia:Recent additions\" and a number, click it and search for the article name.\n\nCurrent archive"
}

... etc

More examples are available in the /examples directory

How it works

Huntsman takes one or more 'seed' urls with the spider.queue.add() method.

Once the process is kicked off with spider.start(), it will take care of extracting links from the page and following only the pages we want.

To define which pages are crawled use the spider.on() function with a string or regular expression.

Each page will only be crawled once. If multiple regular expressions match the uri, they will all be called.

Page URLs which do not match an on condition will never be crawled

Configuration

The spider has default settings, you can override them by passing a settings object when you create a spider.

// use default settings
var huntsman = require('huntsman');
var spider = huntsman.spider();

// override default settings
var huntsman = require('huntsman');
var spider = huntsman.spider({
  throttle: 10, // maximum requests per second
  timeout: 5000 // maximum gap of inactivity before exiting (in milliseconds)
});

Crawling a site

How you configure your spider will vary from site to site, generally you will only be looking for for pages with a specific url format.

Scrape product information from amazon

In this example we can see that amazon product uris all seem to share the format '/gp/product/'.

After queueing the seed uri http://www.amazon.co.uk/ huntsman will follow all the product pages it finds recursively.

/** Example of scraping products from the amazon website **/

var huntsman = require('huntsman');
var spider = huntsman.spider();

spider.extensions = [
  huntsman.extension( 'recurse' ), // load recurse extension & follow anchor links
  huntsman.extension( 'cheerio' ) // load cheerio extension
];

// target only product uris
spider.on( '/gp/product/', function ( err, res ){

  if( !res.extension.cheerio ) return; // content is not html
  var $ = res.extension.cheerio;

  // extract product information
  var product = {
    uri: res.uri,
    heading: $('h1.parseasinTitle').text().trim(),
    image: $('img#main-image').attr('src'),
    description: $('#productDescription').text().trim().substr( 0, 50 )
  };

  console.log( product );

});

spider.queue.add( 'http://www.amazon.co.uk/' );
spider.start();

Find pets for sale on craigslist in london

More complex crawls may require you to specify hub pages to follow before you can get to the content you really want. You can add an on event without a callback & huntsman will still follow and extract links from it.

/** Example of scraping information about pets for sale on cragslist in london **/

var huntsman = require('huntsman');
var spider = huntsman.spider({
  throttle: 2
});

spider.extensions = [
  huntsman.extension( 'recurse' ), // load recurse extension & follow anchor links
  huntsman.extension( 'cheerio' ), // load cheerio extension
  huntsman.extension( 'stats' ) // load stats extension
];

// target only pet uris
spider.on( /\/pet\/(\w+)\.html$/, function ( err, res ){

  if( !res.extension.cheerio ) return; // content is not html
  var $ = res.extension.cheerio;

  // extract listing information
  var listing = {
    heading: $('h2.postingtitle').text().trim(),
    uri: res.uri,
    image: $('img#iwi').attr('src'),
    description: $('#postingbody').text().trim().substr( 0, 50 )
  };

  console.log( listing );

});

// hub pages
spider.on( /http:\/\/london\.craigslist\.co\.uk$/ );
spider.on( /\/pet$/ );

spider.queue.add( 'http://www.craigslist.org/about/sites' );
spider.start();

Extensions

Extensions have default settings, you can override them by passing an optional second argument when the extension is loaded.

// loading an extension
spider.extensions = [
  huntsman.extension( 'extension_name', options )
];

recurse

This extension extracts links from html pages and then adds them to the queue.

The default patterns only target anchor tags which use the http protocol, you can change any of the default patterns by declaring them when the extension is loaded.

// default patterns
huntsman.extension( 'recurse', {
  pattern: {
    search: /a([^>]+)href\s?=\s?['"]([^"'#]+)/gi,
    refine: /['"]([^"'#]+)$/,
    filter: /^https?:\/\//
  }
})

search must be a global regexp and is used to target the links we want to extract.
refine is a regexp used to extract the bits we want from the search regex matches.
filter is a regexp that must match or links are discarded.

// extract both anchor tags and script tags
huntsman.extension( 'recurse', {
  pattern: {
    search: /(a([^>]+)href|script([^>]+)src)\s?=\s?['"]([^"'#]+)/gi, // <a> or <script>
  }
})

// ignore query segment of uris (exclude everything from '?' onwards)
huntsman.extension( 'recurse', {
  pattern: {
    search: /a([^>]+)href\s?=\s?['"]([^"'#\?]+)/gi // charlist for end of uri [^"'#\?]
  }
})

// avoid some file extensions
huntsman.extension( 'recurse', {
  pattern: {
    filter: /^https?:\/\/.*(?!\.(pdf|png|jpg|gif|zip))....$/i, // regex lookahead
  }
})

// avoid all uris with three letter file extensions
huntsman.extension( 'recurse', {
  pattern: {
    filter: /^https?:\/\/.*(?!\.\w{3})....$/, // exclude three letter file extensions
  }
})

// stay on one domain
huntsman.extension( 'recurse', {
  pattern: {
    filter: /^https?:\/\/www\.example\.com/i, // uris must be prefixed with this domain
  }
})

By default recurse converts relative urls to absolute urls and strips fragment identifiers and trailing slashes.

If you need even more control you can override the resolver & normaliser functions to modify these behaviours.

cheerio

This extension parses html and provides jquery-style selectors & functions.

// default settings
huntsman.extension( 'cheerio', { lowerCaseTags: true } )

The res.extension.cheerio function is available in your on callbacks when the response body is HTML.

spider.on( 'example.com', function ( err, res ){

  // use jquery-style selectors & functions
  var $ = res.extension.cheerio;
  if( !$ ) return; // content is not html

  console.log( res.uri, $('h1').text().trim() );

});

cheerio reference: https://github.com/MatthewMueller/cheerio

json

This extension parses the response body with JSON.parse().

// enable json
huntsman.extension( 'json' )

The res.extension.json function is available in your on callbacks when the response body is json.

spider.on( 'example.com', function ( err, res ){

  var json = res.extension.json;
  if( !json ) return; // content is not json

  console.log( res.uri, json );

});

links

This extension extracts links from html pages and returns the result.

It exposes the same functionality that the recurse extension uses to extract links.

// enable extension
huntsman.extension( 'links' )

The res.extension.links function is available in your on callbacks when the response body is a string.

spider.on( 'example.com', function ( err, res ){

  if( !res.extension.links ) return; // content is not a string

  // extract all image tags from body
  var images = res.extension.links({
    pattern: {
      search: /(img([^>]+)src)\s?=\s?['"]([^"'#]+)/gi, // extract img tags
      filter: /\.jpg|\.gif|\.png/i // filter file types
    }
  });

  console.log( images );

});

stats

This extension displays statistics about pages crawled, error counts etc.

// default settings
huntsman.extension( 'stats', { tail: false } )

Custom queues and response storage adapters

I'm currently working on being able to persist the job queue via something like redis and potentially caching http responses in mongo with a TTL.

If you live life on the wild side, these adapters can be configured when you create a spider.

Pull requests welcome.

Build Status

License

(The MIT License)

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the 'Software'), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

missinglink/huntsman

missinglink

Reviews

Repository Details