Find URLs inside SQL Dumps

findurls.js

/**
 * [~] npm install chalk lodash
 * [~] node thisfile.js yourfile.sql
 */

var _ = require('lodash');
var chalk = require('chalk');
var fs = require('fs');
var path = process.argv.length > 1 ? process.argv[process.argv.length - 1] : null;

if (!path){
	console.log(chalk.red('no arguments found'));
	process.exit();
}

console.info(chalk.blue('SCANNING ' + path));

var regex = /["'](((http|https|ftp):\/\/([\w-\d]+\.)+[\w-\d]+){0,1}(\/[\w~,;\-\.\/?%&+#=]*))["']/gmi;
var contents = fs.readFileSync(path, 'utf8');
var matches = contents.match(regex);

var urls = _.map(matches, function(url){
	return url.replace(/^['"]([^'"]+)['"]$/, '$1');
});

var uniqueUrls = _.uniq(urls);

console.info(chalk.blue('FOUND ' + urls.length + ' TOTAL URLs'));
console.info(chalk.blue('FOUND ' + uniqueUrls.length + ' UNIQUE URLs'));

for (var i in uniqueUrls){
	console.log(chalk.white(uniqueUrls[i]));
}