<?php

if (!defined('STDIN')) {
  header('Content-type: text/plain;charset=UTF-8');
  return readfile(__FILE__);
}

/* This file converts GReader shared items to a Tumblr blog. It is written in PHP in
   case anyone wants to make a web tool out of it, but I couldn't be bothered, so
   this script will just run on your own machine using `php -f GReader2Tumblr.php`.

   There are two steps. If there is no file called 'GReader.xml' in the directory, the
   script will make for you to edit and tweak before uploading the posts
   to Tumblr. Your GReader ID is on the {Sharing settings}[*] page, under ‘Choose
   a custom URL’. Even if you have chosen a custom URL, the Atom API the script
   uses requires the number listed there.
   [*] <http://www.google.co.uk/reader/view/#friends-manager-page>

   The second step, if the file is present, once you have tweaked what will be
   uploaded to your satisfaction, will be to upload the posts to Tumblr. I don't
   try to manually detect video links; you'll have to go through the uploads file
   yourself and add those in manually.

For reference from <http://www.tumblr.com/docs/en/api#api_write>, the possible data you can send are:

    regular - Requires at least one:
        title
        body (HTML allowed)
    photo - Requires either source or data, but not both. If both are specified, source is used.
        source - The URL of the photo to copy. This must be a web-accessible URL, not a local file or intranet location.
        data - An image file. See File uploads below.
        caption (optional, HTML allowed)
        click-through-url (optional)
    quote
        quote
        source (optional, HTML allowed)
    link
        name (optional)
        url
        description (optional, HTML allowed)
    conversation
        title (optional)
        conversation
    video - Requires either embed or data, but not both.
        embed - Either the complete HTML code to embed the video, or the URL of a YouTube video page.
        data - A video file for a Vimeo upload. See File uploads below.
        title (optional) - Only applies to Vimeo uploads.
        caption (optional, HTML allowed)
    audio
        data - An audio file. Must be MP3 or AIFF format. See File uploads below.
        externally-hosted-url (optional, replaces data) - Create a post that uses this externally hosted audio-file URL instead of having Tumblr copy and host an uploaded file. Must be MP3 format. No size or duration limits are imposed on externally hosted files.
        caption (optional, HTML allowed)

 */

$gReader = 'averylongnumber';
$email   = 'me@example.com';
$password= 'ahotp4assphras3withpunctuat!onihope';

function encode($dat, $padd) {
  $rv = '';
  if (is_array($dat)) {
    $rv .= $padd."{";
    $first = '';
    foreach ($dat as $key=>$val) {
      $rv .= $first."\n".$padd.'  '.encode($key, $padd.'  ').' : '.encode($val, $padd.'  ');
      $first = ',';
    }
    $rv .= "\n".$padd."}\n";
  } else
    $rv .= json_encode($dat);
  return $rv;
}

$data = @file_get_contents('GReader.json');
if ($data === false) {
  echo "Did not find a file called GReader.json. Creating it for you to edit before it is uploaded.\n";

  $url = "http://www.google.com/reader/public/atom/user%2F$gReader%2Fstate%2Fcom.google%2Fbroadcast";
  $continuation = '';
  $data = FALSE;

  do {
    $curURL = $url.($continuation != '' ? "?c=$continuation" : '');
    $curl = curl_init($curURL);
    curl_setopt($curl, CURLOPT_RETURNTRANSFER, TRUE);
    $rv = curl_exec($curl);
    if ($rv === false || curl_errno($curl)) {
      echo "Failed to load $curURL. Aborting.\n";
      return;
    }
    curl_close($curl);
    $doc = DOMDocument::loadXML($rv);
    if ($doc === FALSE) {
      echo "Google gave us some malformed XML at $curURL. What!?! Aborting.\n";
      return;
    }
    if ($data === FALSE) $data = $doc;
    else {
      $newDoc = $data->importNode($doc->documentElement, true);
      while ($newDoc->hasChildNodes())
        $data->documentElement->appendChild($newDoc->firstChild);
    }
    preg_match('/continuation>([^>]*)<\//', $rv, $continuation); //Lazy; sorry.
    $continuation = @$continuation[1];
  } while ($continuation != '');

  $data = simplexml_import_dom($data);

  $dats = array();
  foreach ($data->entry as $ent) {
    $ent->registerXPathNamespace('atom', "http://www.w3.org/2005/Atom");
    $link = $ent->xpath('atom:link[@rel=\'alternate\']');
    $ent->registerXPathNamespace('gr', 'http://www.google.com/schemas/reader/atom/');
    $annot = $ent->xpath('gr:annotation');

    if ((string)$link[0]['href'] == '')
      $arr = array(
        'type' => 'regular',
        'date' => (string)$ent->published,
        'title' =>  (string)$ent->title->asXML(),
        'body' => preg_replace('/^(<?[^>]*>\s*)?<[^>]*>|<[^>]*>$/', '', $ent->content->asXML())
      );
    else
      $arr = array(
        'type' => 'link',
        'date' => (string)$ent->published,
        'url' => (string)$link[0]['href'],
        'name' => (string)$ent->title
      );
    if (count($annot) > 0 && $arr['type'] == 'link')
      $arr['description'] = preg_replace('/^(<?[^>]*>\s*)?<[^>]*>|<[^>]*>$/', '', $annot[0]->content->asXML());
    $dats[] = $arr;
  }

  $encoded = encode($dats, '');

  if (file_put_contents('GReader.json', $encoded) === false) {
    echo PHP_EOL.PHP_EOL.$encoded.
    "\nCould not save file. Please put the above data in a file 'GReader.json' and tweak it to contain what you want to upload.\n";
  } else {
    echo "Done. Check the file and then run the tool a second time.\n";
  }
  return;
}

$dats = json_decode($data, true);

echo "Uploading data from 'GReader.json'\n";
$rejects = array();
foreach($dats as $ent) {
  $opts = array_merge($ent, array(
      'email' => $email,
      'password' => $password,
      'generator' => 'GReader imports',
      'format' => 'html',
      'send-to-twitter' => 'no'));
  $curl = curl_init('http://www.tumblr.com/api/write');
  curl_setopt($curl, CURLOPT_POST, TRUE);
  curl_setopt($curl, CURLOPT_POSTFIELDS, $opts);
  curl_setopt($curl, CURLOPT_RETURNTRANSFER, TRUE);
  $rv = curl_exec($curl);
  if ($rv === false || curl_getinfo($curl, CURLINFO_HTTP_CODE) != '201') {
    $rejects[] = $opts;
  }
  curl_close($curl);
  echo '.';
  usleep(700);//Avoid rate limit.
}
echo "\n";

if (count($rejects) > 0) {
  echo "Done uploading. Some posts failed to upload.";
  $encoded = encode($rejects, '');
  if (file_put_contents('rejects.json', $encoded) === false) {
    echo "\n\n$encoded\n\nThe above data failed to be uploaded.\n";
  } else {
    echo " The rejected posts were saved in 'rejects.json'.\n";
  }
} else
  echo "All done!\n";


?>