[Summary view] [Print] [Text view]

   1  <?php
   2  /**
   3   * RSS Importer
   4   *
   5   * @package WordPress
   6   * @subpackage Importer
   7   */
   8  
   9  /**
  10   * RSS Importer
  11   *
  12   * Will process a RSS feed for importing posts into WordPress. This is a very
  13   * limited importer and should only be used as the last resort, when no other
  14   * importer is available.
  15   *
  16   * @since unknown
  17   */
  18  class RSS_Import {
  19  
  20      var $posts = array ();
  21      var $file;
  22  
  23  	function header() {
  24          echo '<div class="wrap">';
  25          screen_icon();
  26          echo '<h2>'.__('Import RSS').'</h2>';
  27      }
  28  
  29  	function footer() {
  30          echo '</div>';
  31      }
  32  
  33  	function unhtmlentities($string) { // From php.net for < 4.3 compat
  34          $trans_tbl = get_html_translation_table(HTML_ENTITIES);
  35          $trans_tbl = array_flip($trans_tbl);
  36          return strtr($string, $trans_tbl);
  37      }
  38  
  39  	function greet() {
  40          echo '<div class="narrow">';
  41          echo '<p>'.__('Howdy! This importer allows you to extract posts from an RSS 2.0 file into your blog. This is useful if you want to import your posts from a system that is not handled by a custom import tool. Pick an RSS file to upload and click Import.').'</p>';
  42          wp_import_upload_form("admin.php?import=rss&amp;step=1");
  43          echo '</div>';
  44      }
  45  
  46  	function get_posts() {
  47          global $wpdb;
  48  
  49          set_magic_quotes_runtime(0);
  50          $datalines = file($this->file); // Read the file into an array
  51          $importdata = implode('', $datalines); // squish it
  52          $importdata = str_replace(array ("\r\n", "\r"), "\n", $importdata);
  53  
  54          preg_match_all('|<item>(.*?)</item>|is', $importdata, $this->posts);
  55          $this->posts = $this->posts[1];
  56          $index = 0;
  57          foreach ($this->posts as $post) {
  58              preg_match('|<title>(.*?)</title>|is', $post, $post_title);
  59              $post_title = str_replace(array('<![CDATA[', ']]>'), '', $wpdb->escape( trim($post_title[1]) ));
  60  
  61              preg_match('|<pubdate>(.*?)</pubdate>|is', $post, $post_date_gmt);
  62  
  63              if ($post_date_gmt) {
  64                  $post_date_gmt = strtotime($post_date_gmt[1]);
  65              } else {
  66                  // if we don't already have something from pubDate
  67                  preg_match('|<dc:date>(.*?)</dc:date>|is', $post, $post_date_gmt);
  68                  $post_date_gmt = preg_replace('|([-+])([0-9]+):([0-9]+)$|', '\1\2\3', $post_date_gmt[1]);
  69                  $post_date_gmt = str_replace('T', ' ', $post_date_gmt);
  70                  $post_date_gmt = strtotime($post_date_gmt);
  71              }
  72  
  73              $post_date_gmt = gmdate('Y-m-d H:i:s', $post_date_gmt);
  74              $post_date = get_date_from_gmt( $post_date_gmt );
  75  
  76              preg_match_all('|<category>(.*?)</category>|is', $post, $categories);
  77              $categories = $categories[1];
  78  
  79              if (!$categories) {
  80                  preg_match_all('|<dc:subject>(.*?)</dc:subject>|is', $post, $categories);
  81                  $categories = $categories[1];
  82              }
  83  
  84              $cat_index = 0;
  85              foreach ($categories as $category) {
  86                  $categories[$cat_index] = $wpdb->escape($this->unhtmlentities($category));
  87                  $cat_index++;
  88              }
  89  
  90              preg_match('|<guid.*?>(.*?)</guid>|is', $post, $guid);
  91              if ($guid)
  92                  $guid = $wpdb->escape(trim($guid[1]));
  93              else
  94                  $guid = '';
  95  
  96              preg_match('|<content:encoded>(.*?)</content:encoded>|is', $post, $post_content);
  97              $post_content = str_replace(array ('<![CDATA[', ']]>'), '', $wpdb->escape(trim($post_content[1])));
  98  
  99              if (!$post_content) {
 100                  // This is for feeds that put content in description
 101                  preg_match('|<description>(.*?)</description>|is', $post, $post_content);
 102                  $post_content = $wpdb->escape($this->unhtmlentities(trim($post_content[1])));
 103              }
 104  
 105              // Clean up content
 106              $post_content = preg_replace_callback('|<(/?[A-Z]+)|', create_function('$match', 'return "<" . strtolower($match[1]);'), $post_content);
 107              $post_content = str_replace('<br>', '<br />', $post_content);
 108              $post_content = str_replace('<hr>', '<hr />', $post_content);
 109  
 110              $post_author = 1;
 111              $post_status = 'publish';
 112              $this->posts[$index] = compact('post_author', 'post_date', 'post_date_gmt', 'post_content', 'post_title', 'post_status', 'guid', 'categories');
 113              $index++;
 114          }
 115      }
 116  
 117  	function import_posts() {
 118          echo '<ol>';
 119  
 120          foreach ($this->posts as $post) {
 121              echo "<li>".__('Importing post...');
 122  
 123              extract($post);
 124  
 125              if ($post_id = post_exists($post_title, $post_content, $post_date)) {
 126                  _e('Post already imported');
 127              } else {
 128                  $post_id = wp_insert_post($post);
 129                  if ( is_wp_error( $post_id ) )
 130                      return $post_id;
 131                  if (!$post_id) {
 132                      _e("Couldn't get post ID");
 133                      return;
 134                  }
 135  
 136                  if (0 != count($categories))
 137                      wp_create_categories($categories, $post_id);
 138                  _e('Done !');
 139              }
 140              echo '</li>';
 141          }
 142  
 143          echo '</ol>';
 144  
 145      }
 146  
 147  	function import() {
 148          $file = wp_import_handle_upload();
 149          if ( isset($file['error']) ) {
 150              echo $file['error'];
 151              return;
 152          }
 153  
 154          $this->file = $file['file'];
 155          $this->get_posts();
 156          $result = $this->import_posts();
 157          if ( is_wp_error( $result ) )
 158              return $result;
 159          wp_import_cleanup($file['id']);
 160          do_action('import_done', 'rss');
 161  
 162          echo '<h3>';
 163          printf(__('All done. <a href="%s">Have fun!</a>'), get_option('home'));
 164          echo '</h3>';
 165      }
 166  
 167  	function dispatch() {
 168          if (empty ($_GET['step']))
 169              $step = 0;
 170          else
 171              $step = (int) $_GET['step'];
 172  
 173          $this->header();
 174  
 175          switch ($step) {
 176              case 0 :
 177                  $this->greet();
 178                  break;
 179              case 1 :
 180                  check_admin_referer('import-upload');
 181                  $result = $this->import();
 182                  if ( is_wp_error( $result ) )
 183                      echo $result->get_error_message();
 184                  break;
 185          }
 186  
 187          $this->footer();
 188      }
 189  
 190  	function RSS_Import() {
 191          // Nothing.
 192      }
 193  }
 194  
 195  $rss_import = new RSS_Import();
 196  
 197  register_importer('rss', __('RSS'), __('Import posts from an RSS feed.'), array ($rss_import, 'dispatch'));
 198  ?>