Tries to find a good one-sentence summary from a wikipedia article.

gistfile1.aw

<?php
class WikiLayer {
/*
 * wikilayer::getSignificantSentence()
 * -----------------------------------
 * Grabs the first significant sentence from a wikipedia article
 */
  public function getSignificantSentence($id) {
    $url = 'http://en.wikipedia.org/w/index.php?curid=' . $id;
    $html = $this->get_data($url);

    // Article text is between "bodytext" comment tags
    if(preg_match('/<!-- bodytext -->(.+)<!-- \/bodytext -->/s', $html, $match)) {
      $article = $match[1];

      // Remove tables since they don't contain explanatory text
      $article = preg_replace('#<table[^>]*>.+?</table>#is', '', $article);

      // Remove thumbnail pictures
      $article = preg_replace('#<div class="thumb[^>]+>.+?</div>\s*</div>\s*</div>#s', '', $article);

      // Strip all remaining HTML tags to get plain text
      $article = strip_tags($article);

      // Split the text into sentences
      $sentences = preg_split('/(?<!(mr|ms|dr|mt|st))\.\s*/i', $article);

      // Look for the first sentence not in the format (* is * in *)
      $sentence = FALSE;
      foreach($sentences as $s) {
        $s = trim($s);
        if(preg_match('/.+ is .+ in .+/', $s) == FALSE) {
          $sentence = $s;
          break;
        }
      }
      return $sentence;
    } else {
      return false;
    }
  }

  protected function get_data($url) {
    $ch = curl_init();
    curl_setopt($ch, CURLOPT_URL, $url);
    curl_setopt($ch, CURLOPT_RETURNTRANSFER, TRUE);
    curl_setopt($ch, CURLOPT_USERAGENT, 'Geoloqi (Geo-coded Wikipedia Article Layer) http://geoloqi.com');
    $data = curl_exec($ch);
    return $data;
  }
}
?>