Crawl a site and make an RSS feed

server-rss.js

var http = require('http');
var select = require('soupselect').select;
var htmlparser = require('htmlparser');

var ua = 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/535.2 (KHTML, like Gecko) Chrome/15.0.874.83 Safari/535.2';

var updateFromTheSite = function (callback) {
  var cookies = 'uid=0000; pass=0000000000000000';
  var siteName = 'The Site name';
  var url = 'thedomain.com';
  var client = http.createClient(80, url);
  var request = client.request('GET', '/browse.php', {'Host': url, 'Cookie': cookies, 'User-Agent': ua});
  
  request.on('response', function (response) {
    response.setEncoding('utf8');
    
    var body = '';
    response.on('data', function (chunk) {
      body += chunk;
    });
    
    response.on('end', function () {
      var handler = new htmlparser.DefaultHandler(function (error, dom) {
        if (error) {
          throw error;
        }
        
        var titles = select(dom, 'td[align="left"] a');
        var downloadUrl, title;
        var torrents = [];
        
        for (var i = 3; i < titles.length; i++) {
          if (titles[i].children[0].raw === 'b') {
            title = titles[i].children[0].children[0].raw;
            downloadUrl = 'http://' + url + '/' + titles[i+1].attribs.href;
            torrents.push({title: title, downloadUrl: downloadUrl});
            i++;
          }
        }
        
        if (callback !== null) {
          callback(torrents);
        }
      });
      
      var parser = new htmlparser.Parser(handler);
      parser.parseComplete(body);
    });
  });
  
  request.end();
};

http.createServer(function (req, res) {
  res.writeHead(200, {'Content-Type': 'application/xml'});
  var d = new Date();
  // example: 2011-10-13 19:31:30
  var pad = function (n) {
    return n < 10 ? '0'+n : n;
  }
  var added = d.getUTCFullYear() + '-' + pad(d.getUTCMonth()+1) + '-' + pad(d.getUTCDate()) + ' ' + pad(d.getHours()) + ':' +  pad(d.getMinutes()) + ':' + pad(d.getSeconds());
  
  var xml = '<?xml version="1.0" encoding="utf-8"?>';
  xml += '<rss version="0.91">';
  xml += '<channel><title>The Site</title><link>http://thedomain.com</link>';
  xml += '<description></description>';
  
  updateFromTheSite(function (torrents) {
    torrents.forEach(function (obj) {
      xml += '<item>';
      xml += '<title>' + obj.title + '</title>';
      xml += '<link>' + obj.downloadUrl + '</link>';
      xml += '<description>Added: ' + added + '</description>';
      xml += '</item>';
    });
    xml += '</channel>';
    xml += '</rss>';
    
    res.end(xml);
  });
}).listen(22221, '127.0.0.1');