Using PhantomJS to monitor Google Analytics - full code

get_ga_resources.js

// initialise various variables
var page = require('webpage').create(),
    system = require('system'),
    address;

// how long should we wait for the page to load before we exit
// in ms
var WAIT_TIME = 5000; 

// if the page hasn't loaded after this long, something is probably wrong.
// in ms
var MAX_EXECUTION_TIME = 15000; 

// output error messages
var DEBUG = true;

// a list of regular expressions of resources (urls) to log when we load them
var resources_to_log = [
    new RegExp('^http(?:s)?://(?:www|ssl)\.google-analytics\.com.*collect.*tid\=([UA0-9\-]+)'),
    new RegExp('^http(?:s)?://stats\.g\.doubleclick\.net.*collect.*tid\=([UA0-9\-]+)')
];

// check we have a url, if not exit
if (system.args.length === 1) {
    console.log('Usage: get_ga_resources.js http://www.yoururl.com');
    phantom.exit(1);
} else {
    // address is the url passed
    address = system.args[1];

    // create a function that is called every time a resource is requested
    // http://phantomjs.org/api/webpage/handler/on-resource-requested.html

    
    page.onResourceRequested = function (res, networkRequest) {
        // loop round all our regexs to see if this url matches any of them
        var length = resources_to_log.length;
        while(length--) {                
            if (resources_to_log[length].test(res.url)){		
		var matches = res.url.match(resources_to_log[length]);
		if (matches[1]){
		    console.log(matches[1]);
		}		
		// don't actually call GA.
		//networkRequest.abort();
            }
        }
    };


    // if debug is true, log errors, else ignore them
    page.onError = function(msg, trace){
        if (DEBUG) {
            console.log('ERROR: ' + msg)
            console.log(trace)
        }
    };

    // make a note of any errors so we can print them out
    page.onResourceError = function(resourceError) {
        page.reason = resourceError.errorString;
        page.reason_url = resourceError.url;
    };

    // now all we have to do is open the page, wait WAIT_TIME ms and exit
    try {
        page.open(address, function (status) {
            if (status !== 'success') {
                console.log("FAILED: to load " + system.args[1]);
                console.log(page.reason_url);
                console.log(page.reason);
                phantom.exit();
            } else {
                if (address != page.url){
                   console.log('Redirected: ' + page.url)
                }
                setTimeout(function () {
                    phantom.exit();
                }, WAIT_TIME);
            }
        });
    } finally {
        // if we are still running after MAX_EXECUTION_TIME ms exit
        setTimeout(function() {
            console.log("FAILED: Max execution time " + Math.round(MAX_EXECUTION_TIME) + " seconds exceeded");
            phantom.exit(1);
        }, MAX_EXECUTION_TIME);
    }
}


function onlyUnique(value, index, self) { 
    return self.indexOf(value) === index;
}