Use node.js and jquery (cheerio) to recursively search files for external links

findexternallinks.js

#!/usr/bin/env node
var cheerio = require('cheerio')

var recursive = require('recursive-readdir');

var fs = require('fs');

recursive('.', function (err, files) {
  // Files is an array of filename
  for(f = 0; f < files.length; f++)
  {
  	if(files[f].indexOf('.php') != -1)
  	{
			  var data = fs.readFileSync(files[f]).toString();
			  var $ = cheerio.load(data);
  			var links = $('a');
  			if(links.length)
  			{
  				console.log("===" + files[f]);
  				for(i = 0; i < links.length; i++)
				  {
				  	if(typeof(links[i].attribs.href) != 'undefined' && links[i].attribs.href.indexOf('http') != -1)
				  	{
				  		console.log(links[i].attribs.href);
				  	}
				  }
  			}
		
		}
  }
});