Multiple RSS feeds reading applications

myrss_feedreader.php

<?php
/*
 * 
 * Copyright (C) 2010 - Arief M Utama <[email protected]>
 * All Rights Reserved.
 *
 * Feeds Reader - Multiple feeds reading 
 *
 * Version: 0.9
 * Requirement: PHP >= 5.2
 */

$outfile = 'myreader.xml';

// simple caching implementation
$refresh_time = 900; // 15 mins
if ($_GET['realtimemode'] != '1') {
  if (file_exists($outfile) && ( (time() - filemtime($outfile)) <= $refresh_time)) {
    echo file_get_contents($outfile);
    exit(null);
  } 
}

$sources = array(
  'USGS EarthQuake' => 'http://earthquake.usgs.gov/earthquakes/catalogs/eqs7day-M5.xml',
  'TEMPO' => 'http://rss.tempointeraktif.com/index.xml',
  'TEMPO Fokus' => 'http://rss.tempointeraktif.com/fokus.xml',
  'detikcom' => 'http://rss.detik.com',
  'MetroTV' => 'http://www.metrotvnews.com/index.php/metromain/rss/nusantara',
  'MetroTV' => 'http://www.metrotvnews.com/index.php/metromain/rss/internasional',
  'TVONE' => 'http://www.tvone.co.id/rss/news/',
);

// 
// Some RSS feeds does not give correct timezone information
// they used GMT when they mean +0700 (or Asia/Jakarta)
// Sometimes they just dont provide timezone info at all
// this makes our ordering failed or the time convert shows future time
// hence, I use this structure
//
$date_format = array( // 0 == use original pubDate timezone, 1 force JAVT on the pubDate even if it stated otherwise
  'USGS EarthQuake' => 0, 
  'TEMPO' => 0,
  'TEMPO Fokus' => 0,
  'detikcom' => 0,
  'MetroTV' => 1,
  'MetroTV' => 1,
  'TVONE' => 0,
);

// in case our php config does not set correctly
date_default_timezone_set('Asia/Jakarta');

// this one for later use (see below) 
$javt = new DateTimeZone('Asia/Jakarta');

$feeds = array();

foreach ($sources as $t => $s) {

  $doc = new DOMDocument();

  $doc->load($s);

  foreach ($doc->getElementsByTagName('item') as $node) { 
    $dt = trim($node->getElementsByTagName('pubDate')->item(0)->nodeValue); 

    if ($date_format[$t]) {
      // shall handle GMT==JAVT case nicely
      $dt = substr($dt, 0, 25 /* 25== date string length without timezone info */ );
    }

    // and we dont need to worry bout empty timezones
    // php's new DateTime class take care of it
    $date_object = new DateTime($dt);

    // date_default_timezone_set above was used on cases where timezone info is empty
    // this one function, convert the timezone, make sure we're in the right zone
    $date_object->setTimezone($javt);
    
    $converted_date = $date_object->format('D, d M Y H:i:s O'); // uses standard format on our result

    $item = array(
      'title' => '['. $t . '] ' . trim($node->getElementsByTagName('title')->item(0)->nodeValue),
      'desc' => trim($node->getElementsByTagName('description')->item(0)->nodeValue),
      'link' => trim($node->getElementsByTagName('link')->item(0)->nodeValue),
      'date' => $converted_date,
      'timestamp' => strtotime($converted_date),
    );

    array_push($feeds, $item);

    unset($date_object);
  }

  unset($doc);
}

function timecmp($a, $b)
{
  if ($a['timestamp'] == $b['timestamp']) return 1;
 
  return ($a['timestamp'] < $b['timestamp']) ? 1 : -1;
}

usort($feeds, 'timecmp');

$n = count($feeds);

ob_start();
echo '<?xml version="1.0" encoding="UTF-8"?>' . "\n";
echo '  <rss version="2.0">' . "\n";
echo '    <channel>' . "\n";
echo '      <title>amu@reader</title>' . "\n";
echo '      <link>http://www.arief-mulya.com</link>' . "\n";
echo '      <description>Arief Mulya Utama Reading Feeds</description>' . "\n";

$i = 0;
foreach ($feeds as $item) {
  $i++;
echo '      <item>' . "\n";
echo '        <title><![CDATA[' . ($item['title']) . ' - ' .$i. ' of ' .$n. ']]></title>' . "\n";
echo '        <link><![CDATA[' . ($item['link']) . ']]></link>' . "\n";
echo '        <pubDate><![CDATA[' . ($item['date']) . ']]></pubDate>' . "\n";
echo '        <description><![CDATA[' . ($item['desc']) . ']]></description>' . "\n";
echo '      </item>' . "\n";
}

echo '    </channel>' . "\n";
echo '  </rss>' . "\n";

$out = ob_get_contents();
ob_end_clean();

// cache the output
file_put_contents($outfile, $out);
echo $out;