Club Drupal

<?php

/**

 * @file

 * Contains FeedsSimplePieParser and related classes.

 */

/**

 * Adapter to present SimplePie_Enclosure as FeedsEnclosure object.

 */

class FeedsSimplePieEnclosure extends FeedsEnclosure {

  protected $simplepie_enclosure;

  private $_serialized_simplepie_enclosure;

  /**

   * Constructor requires SimplePie enclosure object.

   */

  function __construct(SimplePie_Enclosure $enclosure) {

    $this->simplepie_enclosure = $enclosure;

  }

  /**

   * Serialization helper.

   *

   * Handle the simplepie enclosure class seperately ourselves.

   */

  public function __sleep() {

    $this->_serialized_simplepie_enclosure = serialize($this->simplepie_enclosure);

    return array('_serialized_simplepie_enclosure');

  }

  /**

   * Unserialization helper.

   *

   * Ensure that the simplepie class definitions are loaded for the enclosure when unserializing.

   */

   public function __wakeup() {

     feeds_include_simplepie();

     $this->simplepie_enclosure = unserialize($this->_serialized_simplepie_enclosure);

  }

  /**

   * Override parent::getValue().

   */

  public function getValue() {

    return $this->simplepie_enclosure->get_link();

  }

  /**

   * Override parent::getMIMEType().

   */

  public function getMIMEType() {

    return $this->simplepie_enclosure->get_real_type();

  }

}

/**

 * Class definition for Common Syndication Parser.

 *

 * Parses RSS and Atom feeds.

 */

class FeedsSimplePieParser extends FeedsParser {

  /**

   * Implements FeedsParser::parse().

   */

  public function parse(FeedsSource $source, FeedsFetcherResult $fetcher_result) {

    feeds_include_simplepie();

    // Initialize SimplePie.

    $parser = new SimplePie();

    $parser->set_raw_data($fetcher_result->getRaw());

    $parser->set_stupidly_fast(TRUE);

    $parser->encode_instead_of_strip(FALSE);

    // @todo Is caching effective when we pass in raw data?

    $parser->enable_cache(TRUE);

    $parser->set_cache_location($this->cacheDirectory());

    $parser->init();

    // Construct the standard form of the parsed feed

    $result = new FeedsParserResult();

    $result->title = html_entity_decode(($title = $parser->get_title()) ? $title : $this->createTitle($parser->get_description()));

    $result->description = $parser->get_description();

    $result->link = html_entity_decode($parser->get_link());

    $items_num = $parser->get_item_quantity();

    for ($i = 0; $i < $items_num; $i++) {

      $item = array();

      $simplepie_item = $parser->get_item($i);

      $item['title'] = html_entity_decode(($title = $simplepie_item->get_title()) ? $title : $this->createTitle($simplepie_item->get_content()));

      $item['description'] = $simplepie_item->get_content();

      $item['url'] = html_entity_decode($simplepie_item->get_link());

      // Use UNIX time. If no date is defined, fall back to REQUEST_TIME.

      $item['timestamp'] = $simplepie_item->get_date("U");

      if (empty($item['timestamp'])) {

        $item['timestamp'] = REQUEST_TIME;

      }

      $item['guid'] = $simplepie_item->get_id();

      // Use URL as GUID if there is no GUID.

      if (empty($item['guid'])) {

        $item['guid'] = $item['url'];

      }

      $author = $simplepie_item->get_author();

      $item['author_name'] = isset($author->name) ? html_entity_decode($author->name) : '';

      $item['author_link'] = isset($author->link) ? $author->link : '';

      $item['author_email'] = isset($author->email) ? $author->email : '';

      // Enclosures

      $enclosures = $simplepie_item->get_enclosures();

      if (is_array($enclosures)) {

        foreach ($enclosures as $enclosure) {

          $item['enclosures'][] = new FeedsSimplePieEnclosure($enclosure);

        }

      }

      // Location

      $latitude = $simplepie_item->get_latitude();

      $longitude = $simplepie_item->get_longitude();

      if (!is_null($latitude) && !is_null($longitude)) {

        $item['location_latitude'][] = $latitude;

        $item['location_longitude'][] = $longitude;

      }

      // Extract tags related to the item

      $simplepie_tags = $simplepie_item->get_categories();

      $tags = array();

      $domains = array();

      if (count($simplepie_tags) > 0) {

        foreach ($simplepie_tags as $tag) {

          $tags[] = (string) $tag->term;

          $domain = (string) $tag->get_scheme();

          if (!empty($domain)) {

            if (!isset($domains[$domain])) {

              $domains[$domain] = array();

            }

            $domains[$domain][] = count($tags) - 1;

          }

        }

      }

      $item['domains'] = $domains;

      $item['tags'] = $tags;

      // Allow parsing to be extended.

      $this->parseExtensions($item, $simplepie_item);

      $item['raw'] = $simplepie_item->data;

      $result->items[] = $item;

    }

    // Release parser.

    unset($parser);

    return $result;

  }

  /**

   * Allow extension of FeedsSimplePie item parsing.

   */

  protected function parseExtensions(&$item, $simplepie_item) {}

  /**

   * Return mapping sources.

   */

  public function getMappingSources() {

    return array(

      'title' => array(

        'name' => t('Title'),

        'description' => t('Title of the feed item.'),

      ),

      'description' => array(

        'name' => t('Description'),

        'description' => t('Description of the feed item.'),

      ),

      'author_name' => array(

        'name' => t('Author name'),

        'description' => t('Name of the feed item\'s author.'),

      ),

      'author_link' => array(

        'name' => t('Author link'),

        'description' => t('Link to the feed item\'s author.'),

      ),

      'author_email' => array(

        'name' => t('Author email'),

        'description' => t('Email address of the feed item\'s author.'),

      ),

      'timestamp' => array(

        'name' => t('Published date'),

        'description' => t('Published date as UNIX time GMT of the feed item.'),

      ),

      'url' => array(

        'name' => t('Item URL (link)'),

        'description' => t('URL of the feed item.'),

      ),

      'guid' => array(

        'name' => t('Item GUID'),

        'description' => t('Global Unique Identifier of the feed item.'),

      ),

      'tags' => array(

        'name' => t('Categories'),

        'description' => t('An array of categories that have been assigned to the feed item.'),

      ),

      'domains' => array(

        'name' => t('Category domains'),

        'description' => t('Domains of the categories.'),

      ),

      'location_latitude' => array(

        'name' => t('Latitudes'),

        'description' => t('An array of latitudes assigned to the feed item.'),

      ),

      'location_longitude' => array(

        'name' => t('Longitudes'),

        'description' => t('An array of longitudes assigned to the feed item.'),

      ),

      'enclosures' => array(

        'name' => t('Enclosures'),

        'description' => t('An array of enclosures attached to the feed item.'),

      ),

    ) + parent::getMappingSources();

  }

  /**

   * Returns cache directory. Creates it if it doesn't exist.

   */

  protected function cacheDirectory() {

    $directory = 'public://simplepie';

    file_prepare_directory($dir, FILE_CREATE_DIRECTORY | FILE_MODIFY_PERMISSIONS);

    return $directory;

  }

  /**

   * Generate a title from a random text.

   */

  protected function createTitle($text = FALSE) {

    // Explode to words and use the first 3 words.

    $words = preg_split("/[\s,]+/", $text);

    $words = array_slice($words, 0, 3);

    return implode(' ', $words);

  }

}