spam_filter_surbl.module in Spam 6

Surbl filter plug in for the spam module. Copyright(c) 2007-2008 Jeremy Andrews <[e-mail]>.

File

filters/spam_filter_surbl/spam_filter_surbl.module

<?php

/**
 * @file
 * Surbl filter plug in for the spam module.
 * Copyright(c) 2007-2008
 *  Jeremy Andrews <[e-mail]>.
 *
 */

/**
 * http://www.surbl.org/lists.html#multi bitmap
 */
define('SPAM_FILTER_SURBL_SC', 2);
define('SPAM_FILTER_SURBL_WS', 4);
define('SPAM_FILTER_SURBL_PH', 8);
define('SPAM_FILTER_SURBL_OB', 16);
define('SPAM_FILTER_SURBL_AB', 32);
define('SPAM_FILTER_SURBL_JP', 64);

/**
 * Spam hook_spamapi implementation.
 */
function spam_filter_surbl_spamapi($op, $type = NULL, $content = array(), $fields = array(), $extra = NULL) {
  switch ($op) {
    case 'filter':
      if (!module_invoke('spam', 'filter_enabled', 'spam_filter_surbl', $type, $content, $fields, $extra)) {
        return;
      }
      return spam_filter_surbl_spam_filter($content, $type, $fields, $extra);
    case 'filter_module':
      return 'spam_filter_surbl';
    case 'filter_info':
      return array(
        'name' => t('Surbl filter'),
        'module' => t('spam_filter_surbl'),
        'description' => t('A spam url filter.'),
        'help' => t('Look up URLs in SURBL to determine if is spam.'),
      );
      break;
    case 'filter_install':
      return array(
        'status' => SPAM_FILTER_ENABLED,
        'gain' => 250,
        'weight' => -7,
      );
  }
}

/**
 * Extract URLs from content.
 */
function _spam_filter_surbl_url_extract($content, $type, $fields, $extra = array()) {
  static $urls = array();
  $id = spam_invoke_module($type, 'content_id', $content, $extra);
  if (is_object($content)) {
    $content = (array) $content;
  }
  if (!isset($urls["{$type}-{$id}"])) {
    $string = '';
    foreach ($fields['main'] as $field) {
      $string .= $content["{$field}"] . ' ';
    }
    if (isset($fields['other']) && is_array($fields['other'])) {
      foreach ($fields['other'] as $field) {
        $string .= $content["{$field}"] . ' ';
      }
    }

    // TODO: Improve this matching.  We don't actually extract mailto: urls.
    $URI = "(http://|https://|ftp://|mailto:)";

    // Find all urls in content.
    preg_match_all("!(<p>|[ \n\r\t\\(]*)({$URI}([a-zA-Z0-9@:%_~#?&=.,/;-]*[a-zA-Z0-9@:%_~#&=/;-]))([.,?]?)(?=(</p>|[ \n\r\t\\)]*))!i", $string, $matches);
    $u = array();
    foreach ($matches[2] as $url) {
      $url = preg_replace("'{$URI}'", '', $url);

      // get full domain (ie www.sample.com)
      preg_match("/^()?([^\\/\"\\']+)/i", $url, $domain);

      // get root domain (ie sample.com)
      preg_match("/[^\\.\\/]+\\.[^\\.\\/]+\$/", $domain[2], $root);
      $u[md5($root[0])] = htmlspecialchars(drupal_strtolower($root[0]));
    }
    $urls["{$type}-{$id}"] = $u;
  }
  return $urls["{$type}-{$id}"];
}

/**
 * Search for known spam urls in content.
 */
function spam_filter_surbl_spam_filter($content, $type, $fields, $extra = array(), $filter_test = FALSE) {
  $action = array();
  $id = spam_invoke_module($type, 'content_id', $content, $extra);
  $spam = FALSE;
  $urls = _spam_filter_surbl_url_extract($content, $type, $fields, $extra);
  if (is_array($urls) && !empty($urls)) {
    foreach ($urls as $url) {
      $lookup = "{$url}.multi.surbl.org";
      $ip = gethostbyname($lookup);
      if ($ip != $lookup) {

        // this domain was in a SURBL, process accordingly
        preg_match("/[^\\.\\/]+\$/", $ip, $code);
        if ($code[0] & SPAM_FILTER_SURBL_SC) {
          spam_log(SPAM_IMPORTANT, 'spam_filter_surbl_spam_filter', t('found spam url(@url) @surbl', array(
            '@url' => $url,
            '@surbl' => t('SpamCop message-body URI domains'),
          )), $type, $id);
        }
        if ($code[0] & SPAM_FILTER_SURBL_WS) {
          spam_log(SPAM_IMPORTANT, 'spam_filter_surbl_spam_filter', t('found spam url(@url) @surbl', array(
            '@url' => $url,
            '@surbl' => t('sa-blacklist domains'),
          )), $type, $id);
        }
        if ($code[0] & SPAM_FILTER_SURBL_PH) {
          spam_log(SPAM_IMPORTANT, 'spam_filter_surbl_spam_filter', t('found spam url(@url) @surbl', array(
            '@url' => $url,
            '@surbl' => t('Phishing data source'),
          )), $type, $id);
        }
        if ($code[0] & SPAM_FILTER_SURBL_OB) {
          spam_log(SPAM_IMPORTANT, 'spam_filter_surbl_spam_filter', t('found spam url(@url) @surbl', array(
            '@url' => $url,
            '@surbl' => t('Outblaze URI blacklist'),
          )), $type, $id);
        }
        if ($code[0] & SPAM_FILTER_SURBL_AB) {
          spam_log(SPAM_IMPORTANT, 'spam_filter_surbl_spam_filter', t('found spam url(@url) @surbl', array(
            '@url' => $url,
            '@surbl' => t('AbuseButler spamvertised sites'),
          )), $type, $id);
        }
        if ($code[0] & SPAM_FILTER_SURBL_JP) {
          spam_log(SPAM_IMPORTANT, 'spam_filter_surbl_spam_filter', t('found spam url(@url) @surbl', array(
            '@url' => $url,
            '@surbl' => t('jwSpamSpy + Prolocation data source'),
          )), $type, $id);
        }
        $action['spam_filter_surbl'][] = array(
          'url' => $url,
          'probability' => 99,
        );
        $spam = TRUE;
      }
      else {
        spam_log(SPAM_DEBUG, 'spam_filter_surbl_spam_filter', t('not spam url(@url)', array(
          '@url' => $url,
        )), $type, $id);
      }
    }
  }
  if ($spam) {
    $action['total'] = 99;
  }
  else {
    $action['total'] = 0;
  }
  return $action;
}

Functions

Name	Description
spam_filter_surbl_spamapi	Spam hook_spamapi implementation.
spam_filter_surbl_spam_filter	Search for known spam urls in content.
_spam_filter_surbl_url_extract	Extract URLs from content.

Constants

Name	Description
SPAM_FILTER_SURBL_AB
SPAM_FILTER_SURBL_JP
SPAM_FILTER_SURBL_OB
SPAM_FILTER_SURBL_PH
SPAM_FILTER_SURBL_SC	http://www.surbl.org/lists.html#multi bitmap
SPAM_FILTER_SURBL_WS