function find_content($xpath, $query)
{
$results = $xpath->evaluate($query);
$result = $results->item(0);
return dom_to_html($result);
}
function process_bbs_appgame_url($req_url)
{
preg_match('#^http://bbs\\.appgame\\.com/thread-[\\d]+-[\\d]+-[\\d]+\\.html$#us', $req_url, $matches);
if ($matches == null) {
return false;
}
$user_agent = 'User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_9_0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/31.0.1650.57 Safari/537.36';
$html = curl_get_content($req_url, $user_agent);
if (empty($html)) {
return false;
}
$regex_match = "#<div id=\"post_(\\d+)\">#s";
if (!preg_match($regex_match, $html, $match)) {
return false;
}
$pid = $match[1];
preg_match('#发表于[^<]*?<span title="([^"]*?)">.*?</span>#s', $html, $match);
$time_str = $match[1];
preg_match("#<title>([^<]*?)</title>#s", $html, $match);
$title = $match[1];
$title = preg_replace("#_[^_]+_任玩堂.*\$#u", '', $title);
$html = mb_convert_encoding($html, 'HTML-ENTITIES', mb_detect_encoding($html));
$saw = new nokogiri($html);
$target = $saw->get('td#postmessage_' . $pid);
$dom = $target->getDom();
$node = $dom->firstChild->childNodes->item(0);
$content = strip_tags(dom_to_html($node));
$content = preg_replace("#[\\s]+#us", '', $content);
preg_match('#showauthor\\(this, \'userinfo' . $pid . '\'.*?<img .*?src="([^"]+?)"#s', $html, $match);
$user_pic = $match[1];
$user_img = get_redirect_url($user_pic);
$res = array();
$res['onebox'] = 'appgame-bbs';
$res['provider_name'] = '任玩堂论坛';
$res['provider_url'] = 'http://bbs.appgame.com/';
$res['favicon_url'] = 'http://www.appgame.com/favicon.ico';
$res['ori_url'] = $req_url;
$res['title'] = $title;
$res['image'] = $user_img;
$res['ID'] = intval($pid);
$res['description'] = trim($content);
$res['update_time'] = format_time($time_str);
$res['create_time'] = $res['update_time'];
return $res;
}