Simple PhantomJS-based web crawler library

webcrawler.js

//PhantomJS http://phantomjs.org/ based web crawler Anton Ivanov [email protected] 2012
//UPDATE: This gist has been made into a Node.js module and now can be installed with "npm install js-crawler"
//the Node.js version does not use Phantom.JS, but the API available to the client is similar to the present gist

(function(host) {

    function Crawler() {
        this.visitedURLs = {};
    };
    
    Crawler.webpage = require('webpage');

    Crawler.prototype.crawl = function (url, depth, onSuccess, onFailure) {
        if (0 == depth || this.visitedURLs[url]) {
            return;
        };
        var self = this;
        var page = Crawler.webpage.create();

        page.open(url, function (status) {
            if ('fail' === status) { 
                onFailure({
                    url: url, 
                    status: status
                });
            } else {
                var documentHTML = page.evaluate(function () {
                    return document.body && document.body.innerHTML ? document.body.innerHTML : "";
                });
                self.crawlURLs(self.getAllURLs(page), depth - 1, onSuccess, onFailure);
                self.visitedURLs[url] = true;
                onSuccess({
                    url: url,
                    status: status,
                    content: documentHTML
                });
            };
        });
    };

    Crawler.prototype.getAllURLs = function(page) {
        return page.evaluate(function () {
            return Array.prototype.slice.call(document.querySelectorAll("a"), 0)
                .map(function (link) {
                    return link.getAttribute("href");
                });
        });
    };

    Crawler.prototype.crawlURLs = function(urls, depth, onSuccess, onFailure) {
        var self = this;
        urls.filter(function (url) {
            return Crawler.isTopLevelURL(url);
        }).forEach(function (url) {
            self.crawl(url, depth, onSuccess, onFailure);
        });
    };

    Crawler.isTopLevelURL = function(url) {
        return 0 == url.indexOf("http");
    };

    host.Crawler = Crawler;
})(phantom);

new phantom.Crawler().crawl("https://github.com/ariya/phantomjs/wiki/Quick-Start", 2, 
    function onSuccess(page) {
        console.log("Loaded page. URL = " + page.url + " content length = " + page.content.length + " status = " + page.status);
    }, 
    function onFailure(page) {
        console.log("Could not load page. URL = " +  page.url + " status = " + page.status);
    }
);

Is it supposed to be used with nodejs?

It is supposed to be used with PhantomJS, a headless WebKit browser http://phantomjs.org/

Great gist :)

Great gist, however if you increase the crawling depth it seems it will start fail with 'select: Invalid argument'.
Any idea on this issue?

Please, note that now a Node.JS module based on this gist is available https://www.npmjs.org/package/js-crawler. Please, file issues at https://github.com/antivanov/js-crawler/issues and you are welcome to contribute

The Node.js version does not use Phantom.JS but rather standard Node means for making HTTP requests with the Request module https://github.com/mikeal/request

hc0519, I suspect that this can be a JavaScript error on the page that is loaded into the Crawler, there are no variables or functions named 'select' in the source code of the crawler itself.

The JavaScript error then is not related to the crawling depth and occurs not in the Crawler itself but in a Phantom.JS browser instance when it loads a page with a JavaScript error. This error should not affect the work of the Crawler.

@hc0519 @antivanov 'select: Invalid argument' is error message from ssh.

I don't know why it appears, but it seems definitely bug of PhantomJs.

Any reason you switched away from PhantomJS?

how to use this code in phantomjs

Any reason you switched away from PhantomJS?

PhantomJS may be a bit heavier to use than just issuing HTTP requests over the network, however, maybe PhantomJS can be supported as an alternative underlying request making tool for js-crawler https://github.com/antivanov/js-crawler

how to use this code in phantomjs

You can run phantomjs webcrawler.js

when the crawler will stop? or it never end?

#sorry for my bad english

hi,
This program can render the js code which not in the html !
I found a lot of pages need to render the HTML that can loaded the effect infomation !

thx !