[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 _normalize_tag( $matches ) {
  47          return '<' . strtolower( $matches[1] );
  48      }
  49  
  50  	function get_posts() {
  51          global $wpdb;
  52  
  53          set_magic_quotes_runtime(0);
  54          $datalines = file($this->file); // Read the file into an array
  55          $importdata = implode('', $datalines); // squish it
  56          $importdata = str_replace(array ("\r\n", "\r"), "\n", $importdata);
  57  
  58          preg_match_all('|<item>(.*?)</item>|is', $importdata, $this->posts);
  59          $this->posts = $this->posts[1];
  60          $index = 0;
  61          foreach ($this->posts as $post) {
  62              preg_match('|<title>(.*?)</title>|is', $post, $post_title);
  63              $post_title = str_replace(array('<![CDATA[', ']]>'), '', $wpdb->escape( trim($post_title[1]) ));
  64  
  65              preg_match('|<pubdate>(.*?)</pubdate>|is', $post, $post_date_gmt);
  66  
  67              if ($post_date_gmt) {
  68                  $post_date_gmt = strtotime($post_date_gmt[1]);
  69              } else {
  70                  // if we don't already have something from pubDate
  71                  preg_match('|<dc:date>(.*?)</dc:date>|is', $post, $post_date_gmt);
  72                  $post_date_gmt = preg_replace('|([-+])([0-9]+):([0-9]+)$|', '\1\2\3', $post_date_gmt[1]);
  73                  $post_date_gmt = str_replace('T', ' ', $post_date_gmt);
  74                  $post_date_gmt = strtotime($post_date_gmt);
  75              }
  76  
  77              $post_date_gmt = gmdate('Y-m-d H:i:s', $post_date_gmt);
  78              $post_date = get_date_from_gmt( $post_date_gmt );
  79  
  80              preg_match_all('|<category>(.*?)</category>|is', $post, $categories);
  81              $categories = $categories[1];
  82  
  83              if (!$categories) {
  84                  preg_match_all('|<dc:subject>(.*?)</dc:subject>|is', $post, $categories);
  85                  $categories = $categories[1];
  86              }
  87  
  88              $cat_index = 0;
  89              foreach ($categories as $category) {
  90                  $categories[$cat_index] = $wpdb->escape($this->unhtmlentities($category));
  91                  $cat_index++;
  92              }
  93  
  94              preg_match('|<guid.*?>(.*?)</guid>|is', $post, $guid);
  95              if ($guid)
  96                  $guid = $wpdb->escape(trim($guid[1]));
  97              else
  98                  $guid = '';
  99  
 100              preg_match('|<content:encoded>(.*?)</content:encoded>|is', $post, $post_content);
 101              $post_content = str_replace(array ('<![CDATA[', ']]>'), '', $wpdb->escape(trim($post_content[1])));
 102  
 103              if (!$post_content) {
 104                  // This is for feeds that put content in description
 105                  preg_match('|<description>(.*?)</description>|is', $post, $post_content);
 106                  $post_content = $wpdb->escape($this->unhtmlentities(trim($post_content[1])));
 107              }
 108  
 109              // Clean up content
 110              $post_content = preg_replace_callback('|<(/?[A-Z]+)|', array( &$this, '_normalize_tag' ), $post_content);
 111              $post_content = str_replace('<br>', '<br />', $post_content);
 112              $post_content = str_replace('<hr>', '<hr />', $post_content);
 113  
 114              $post_author = 1;
 115              $post_status = 'publish';
 116              $this->posts[$index] = compact('post_author', 'post_date', 'post_date_gmt', 'post_content', 'post_title', 'post_status', 'guid', 'categories');
 117              $index++;
 118          }
 119      }
 120  
 121  	function import_posts() {
 122          echo '<ol>';
 123  
 124          foreach ($this->posts as $post) {
 125              echo "<li>".__('Importing post...');
 126  
 127              extract($post);
 128  
 129              if ($post_id = post_exists($post_title, $post_content, $post_date)) {
 130                  _e('Post already imported');
 131              } else {
 132                  $post_id = wp_insert_post($post);
 133                  if ( is_wp_error( $post_id ) )
 134                      return $post_id;
 135                  if (!$post_id) {
 136                      _e('Couldn&#8217;t get post ID');
 137                      return;
 138                  }
 139  
 140                  if (0 != count($categories))
 141                      wp_create_categories($categories, $post_id);
 142                  _e('Done !');
 143              }
 144              echo '</li>';
 145          }
 146  
 147          echo '</ol>';
 148  
 149      }
 150  
 151  	function import() {
 152          $file = wp_import_handle_upload();
 153          if ( isset($file['error']) ) {
 154              echo $file['error'];
 155              return;
 156          }
 157  
 158          $this->file = $file['file'];
 159          $this->get_posts();
 160          $result = $this->import_posts();
 161          if ( is_wp_error( $result ) )
 162              return $result;
 163          wp_import_cleanup($file['id']);
 164          do_action('import_done', 'rss');
 165  
 166          echo '<h3>';
 167          printf(__('All done. <a href="%s">Have fun!</a>'), get_option('home'));
 168          echo '</h3>';
 169      }
 170  
 171  	function dispatch() {
 172          if (empty ($_GET['step']))
 173              $step = 0;
 174          else
 175              $step = (int) $_GET['step'];
 176  
 177          $this->header();
 178  
 179          switch ($step) {
 180              case 0 :
 181                  $this->greet();
 182                  break;
 183              case 1 :
 184                  check_admin_referer('import-upload');
 185                  $result = $this->import();
 186                  if ( is_wp_error( $result ) )
 187                      echo $result->get_error_message();
 188                  break;
 189          }
 190  
 191          $this->footer();
 192      }
 193  
 194  	function RSS_Import() {
 195          // Nothing.
 196      }
 197  }
 198  
 199  $rss_import = new RSS_Import();
 200  
 201  register_importer('rss', __('RSS'), __('Import posts from an RSS feed.'), array ($rss_import, 'dispatch'));
 202  ?>