<?php
// $Id: googleanalytics.module,v 1.70.2.16 2011-02-23 23:57:47 hass Exp $

/*
 * @file
 * Drupal Module: GoogleAnalytics
 * Adds the required Javascript to the bottom of all your Drupal pages
 * to allow tracking by the Google Analytics statistics package.
 *
 * @author: Alexander Hass <http://drupal.org/user/85918>
 */

define('GA_TRACKFILES_EXTENSIONS', '7z|aac|arc|arj|asf|asx|avi|bin|csv|doc|exe|flv|gif|gz|gzip|hqx|jar|jpe?g|js|mp(2|3|4|e?g)|mov(ie)?|msi|msp|pdf|phps|png|ppt|qtm?|ra(m|r)?|sea|sit|tar|tgz|torrent|txt|wav|wma|wmv|wpd|xls|xml|z|zip');

/**
 * Implements hook_help().
 */
function googleanalytics_help($path, $arg) {
  switch ($path) {
    case 'admin/config/system/googleanalytics':
      return t('<a href="@ga_url">Google Analytics</a> is a free (registration required) website traffic and marketing effectiveness service.', array('@ga_url' => 'http://www.google.com/analytics/'));
  }
}

/**
 * Implements hook_theme().
 */
function googleanalytics_theme() {
  return array(
    'googleanalytics_admin_custom_var_table' => array(
      'render element' => 'form',
    ),
  );
}

/**
 * Implements hook_permission().
 */
function googleanalytics_permission() {
  return array(
    'administer google analytics' => array(
      'title' => t('Administer Google Analytics'),
      'description' => t('Perform maintenance tasks for Google Analytics.'),
    ),
    'opt-in or out of tracking' => array(
      'title' => t('Opt-in or out of tracking'),
      'description' => t('Allow users to decide if tracking code will be added to pages or not.'),
    ),
    'use PHP for tracking visibility' => array(
      'title' => t('Use PHP for tracking visibility'),
      'description' => t('Enter PHP code in the field for tracking visibility settings. %warning', array('%warning' => t('Warning: Give to trusted roles only; this permission has security implications.'))),
    ),
  );
}

/**
 * Implements hook_menu().
 */
function googleanalytics_menu() {
  $items['admin/config/system/googleanalytics'] = array(
    'title' => 'Google Analytics',
    'description' => 'Configure tracking behavior to get insights into your website traffic and marketing effectiveness.',
    'page callback' => 'drupal_get_form',
    'page arguments' => array('googleanalytics_admin_settings_form'),
    'access arguments' => array('administer google analytics'),
    'type' => MENU_NORMAL_ITEM,
    'file' => 'googleanalytics.admin.inc',
  );

  return $items;
}

/**
 * Implements hook_page_alter() to insert JavaScript to the appropriate scope/region of the page.
 */
function googleanalytics_page_alter(&$page) {
  global $user;

  $id = variable_get('googleanalytics_account', '');

  // Get page status code for visibility filtering.
  $status = drupal_get_http_header('Status');
  $trackable_status_codes = array(
    '403 Forbidden',
    '404 Not Found',
  );

  // 1. Check if the GA account number has a value.
  // 2. Track page views based on visibility value.
  // 3. Check if we should track the currently active user's role.
  // 4. Ignore pages visibility filter for 404 or 403 status codes.
  if (!empty($id) && (_googleanalytics_visibility_pages() || in_array($status, $trackable_status_codes)) && _googleanalytics_visibility_user($user)) {

    // We allow different scopes. Default to 'header' but allow user to override if they really need to.
    $scope = variable_get('googleanalytics_js_scope', 'header');

    if (variable_get('googleanalytics_trackadsense', FALSE)) {
      // Custom tracking. Prepend before all other JavaScript.
      drupal_add_js('window.google_analytics_uacct = ' . drupal_json_encode($id) . ';', array('type' => 'inline', 'group' => JS_LIBRARY-1));
    }

    // Add link tracking.
    $link_settings = array();
    if ($track_outbound = variable_get('googleanalytics_trackoutbound', 1)) {
      $link_settings['trackOutbound'] = $track_outbound;
    }
    if ($track_mailto = variable_get('googleanalytics_trackmailto', 1)) {
      $link_settings['trackMailto'] = $track_mailto;
    }
    if (($track_download = variable_get('googleanalytics_trackfiles', 1)) && ($trackfiles_extensions = variable_get('googleanalytics_trackfiles_extensions', GA_TRACKFILES_EXTENSIONS))) {
      $link_settings['trackDownload'] = $track_download;
      $link_settings['trackDownloadExtensions'] = $trackfiles_extensions;
    }
    if ($track_outbound_as_pageview = variable_get('googleanalytics_trackoutboundaspageview', 0)) {
      $link_settings['trackOutboundAsPageview'] = $track_outbound_as_pageview;
    }
    if ($track_domain_mode = variable_get('googleanalytics_domain_mode', 0)) {
      $link_settings['trackDomainMode'] = $track_domain_mode;
    }
    if ($track_cross_domains = variable_get('googleanalytics_cross_domains', '')) {
      $domains = preg_split('/(\r\n?|\n)/', $track_cross_domains);
      $link_settings['trackCrossDomains'] = implode('|', $domains);
    }

    if (!empty($link_settings)) {
      drupal_add_js(array('googleanalytics' => $link_settings), 'setting');
      drupal_add_js(drupal_get_path('module', 'googleanalytics') . '/googleanalytics.js');
    }

    // Add User profile segmentation values.
    $profile_fields = variable_get('googleanalytics_segmentation', array());
    if (!empty($profile_fields)) {
      // Invoke all modules having a hook_user_load() implemented.
      // If the invoked modules extend the $user object, the additional
      // data can be tracked.
      module_invoke_all('user_load', array($user->uid => $user));

      $fields = array();
      foreach ($profile_fields as $field => $title) {
        $fields[$field] = is_array($user->$field) ? implode(',', $user->$field) : $user->$field;
      }

      // Only show segmentation variable if there are specified fields.
      $segmentation = '';
      if (count($fields) > 0) {
        $segmentation = '_gaq.push(["_setVar", ' . drupal_json_encode(implode(':', $fields)) . ']);';
      }
    }

    // Site search tracking support.
    $url_custom = '';
    if (module_exists('search') && variable_get('googleanalytics_site_search', FALSE) && arg(0) == 'search' && $keys = googleanalytics_search_get_keys()) {
      $url_custom = drupal_json_encode(rawurldecode(url('search/' . arg(1), array('query' => array('search' => $keys)))));
    }

    // If this node is a translation of another node, pass the original
    // node instead.
    if (module_exists('translation') && variable_get('googleanalytics_translation_set', 0)) {
      // Check we have a node object, it supports translation, and its
      // translated node ID (tnid) doesn't match its own node ID.
      $node = menu_get_object();
      if ($node && translation_supported_type($node->type) && isset($node->tnid) && ($node->tnid != $node->nid)) {
        $source_node = node_load($node->tnid);
        $languages = language_list();
        $url_custom = drupal_json_encode(url('node/' . $source_node->nid, array('language' => $languages[$source_node->language])));
      }
    }

    // Track access denied (403) and file not found (404) pages.
    if ($status == '403 Forbidden') {
      // See http://www.google.com/support/analytics/bin/answer.py?answer=86927
      $url_custom = '"/403.html?page=" + document.location.pathname + document.location.search + "&from=" + document.referrer';
    }
    elseif ($status == '404 Not Found') {
      $url_custom = '"/404.html?page=" + document.location.pathname + document.location.search + "&from=" + document.referrer';
    }

    // Add any custom code snippets if specified.
    $codesnippet_before = variable_get('googleanalytics_codesnippet_before', '');
    $codesnippet_after = variable_get('googleanalytics_codesnippet_after', '');

    // Add custom variables.
    $googleanalytics_custom_vars = variable_get('googleanalytics_custom_var', array());
    $custom_var = '';
    for ($i = 1; $i < 6; $i++) {
      $custom_var_name = !empty($googleanalytics_custom_vars['slots'][$i]['name']) ? $googleanalytics_custom_vars['slots'][$i]['name'] : '';
      if (!empty($custom_var_name)) {
        $custom_var_value = !empty($googleanalytics_custom_vars['slots'][$i]['value']) ? $googleanalytics_custom_vars['slots'][$i]['value'] : '';
        $custom_var_scope = !empty($googleanalytics_custom_vars['slots'][$i]['scope']) ? $googleanalytics_custom_vars['slots'][$i]['scope'] : 3;

        $node = menu_get_object();
        $custom_var_value = token_replace($custom_var_value, array('node' => $node), array('clear' => TRUE));

        // Suppress empty custom variables.
        if (!is_numeric($custom_var_value) && empty($custom_var_value)) {
          continue;
        }

        // The length of the string used for the 'name' and the length of the
        // string used for the 'value' must not exceed 64 bytes after url encoding.
        $name_length = drupal_strlen(rawurlencode($custom_var_name));
        $tmp_value = rawurlencode($custom_var_value);
        $value_length = drupal_strlen($tmp_value);
        if ($name_length + $value_length > 64) {
          // Trim value and remove fragments of url encoding.
          $tmp_value = rtrim(substr($tmp_value, 0, 63 - $name_length), '%0..9A..F');
          $custom_var_value = urldecode($tmp_value);
        }

        $custom_var_name = drupal_json_encode($custom_var_name);
        $custom_var_value = drupal_json_encode($custom_var_value);
        $custom_var .= "_gaq.push(['_setCustomVar', $i, $custom_var_name, $custom_var_value, $custom_var_scope]);";
      }
    }

    // Build tracker code.
    $script = 'var _gaq = _gaq || [];';
    $script .= '_gaq.push(["_setAccount", ' . drupal_json_encode($id) . ']);';
    if (variable_get('googleanalytics_tracker_anonymizeip', 0)) {
      // FIXME: The Google API is currently broken and "_gat._anonymizeIp" is only
      // a workaround until "_anonymizeIp" has been implemented/fixed.
      $script .= '_gaq.push(["_gat._anonymizeIp"]);';
    }

    // Domain tracking type.
    global $cookie_domain;
    $domain_mode = variable_get('googleanalytics_domain_mode', 0);

    // Per RFC 2109, cookie domains must contain at least one dot other than the
    // first. For hosts such as 'localhost' or IP Addresses we don't set a cookie domain.
    if ($domain_mode == 1 && count(explode('.', $cookie_domain)) > 2 && !is_numeric(str_replace('.', '', $cookie_domain))) {
      $script .= '_gaq.push(["_setDomainName", ' . drupal_json_encode($cookie_domain) . ']);';
    }
    elseif ($domain_mode == 2) {
      $script .= '_gaq.push(["_setDomainName", "none"]);';
      $script .= '_gaq.push(["_setAllowLinker", true]);';
    }

    if (!empty($segmentation)) {
      $script .= $segmentation;
    }
    if (!empty($codesnippet_before)) {
      $script .= $codesnippet_before;
    }
    if (!empty($custom_var)) {
      $script .= $custom_var;
    }
    if (empty($url_custom)) {
      $script .= '_gaq.push(["_trackPageview"]);';
    }
    else {
      $script .= '_gaq.push(["_trackPageview", ' . $url_custom . ']);';
    }
    if (!empty($codesnippet_after)) {
      $script .= $codesnippet_after;
    }

    $script .= '(function() {';
    $script .= 'var ga = document.createElement("script");';
    $script .= 'ga.type = "text/javascript";';
    $script .= 'ga.async = true;';

    // Should a local cached copy of ga.js be used?
    if (variable_get('googleanalytics_cache', 0) && $url = _googleanalytics_cache('http://www.google-analytics.com/ga.js')) {
      // A dummy query-string is added to filenames, to gain control over
      // browser-caching. The string changes on every update or full cache
      // flush, forcing browsers to load a new copy of the files, as the
      // URL changed.
      $query_string = '?' . variable_get('css_js_query_string', '0');

      $script .= 'ga.src = "' . $url . $query_string . '";';
    }
    else {
      $script .= 'ga.src = ("https:" == document.location.protocol ? "https://ssl" : "http://www") + ".google-analytics.com/ga.js";';
    }
    $script .= 'var s = document.getElementsByTagName("script")[0];';
    $script .= 's.parentNode.insertBefore(ga, s);';
    $script .= '})();';

    drupal_add_js($script, array('scope' => $scope, 'type' => 'inline'));
  }
}

/**
 * Implements hook_form_FORM_ID_alter().
 *
 * Allow users to decide if tracking code will be added to pages or not.
 */
function googleanalytics_form_user_profile_form_alter(&$form, &$form_state) {
  $account = $form['#user'];
  $category = $form['#user_category'];

  if ($category == 'account' && user_access('opt-in or out of tracking') && ($custom = variable_get('googleanalytics_custom', 0)) != 0 && _googleanalytics_visibility_roles($account)) {
    $form['googleanalytics'] = array(
      '#type' => 'fieldset',
      '#title' => t('Google Analytics configuration'),
      '#weight' => 3,
      '#collapsible' => TRUE,
      '#tree' => TRUE
    );

    switch ($custom) {
      case 1:
        $description = t('Users are tracked by default, but you are able to opt out.');
        break;

      case 2:
        $description = t('Users are <em>not</em> tracked by default, but you are able to opt in.');
        break;
    }

    // Disable tracking for visitors who have opted out from tracking via DNT (Do-Not-Track) header.
    $disabled = FALSE;
    if (variable_get('googleanalytics_privacy_donottrack', 1) && !empty($_SERVER['HTTP_DNT'])) {
      $disabled = TRUE;

      // Override settings value.
      $account->data['googleanalytics']['custom'] = FALSE;

      $description .= '<span class="admin-disabled">';
      $description .= ' ' . t('You have opted out from tracking via browser privacy settings.');
      $description .= '</span>';
    }

    $form['googleanalytics']['custom'] = array(
      '#type' => 'checkbox',
      '#title' => t('Enable user tracking'),
      '#description' => $description,
      '#default_value' => isset($account->data['googleanalytics']['custom']) ? $account->data['googleanalytics']['custom'] : ($custom == 1),
      '#disabled' => $disabled,
    );

    return $form;
  }
}

/**
 * Implements hook_user_presave().
 */
function googleanalytics_user_presave(&$edit, $account, $category) {
  if (isset($edit['googleanalytics']['custom'])) {
    $edit['data']['googleanalytics']['custom'] = $edit['googleanalytics']['custom'];
  }
}

/**
 * Implements hook_cron().
 */
function googleanalytics_cron() {
  // Regenerate the tracking code file every day.
  if (REQUEST_TIME - variable_get('googleanalytics_last_cache', 0) >= 86400 && variable_get('googleanalytics_cache', 0)) {
    _googleanalytics_cache('http://www.google-analytics.com/ga.js', TRUE);
    variable_set('googleanalytics_last_cache', REQUEST_TIME);
  }
}

/**
 * Download/Synchronize/Cache tracking code file locally.
 *
 * @param $location
 *   The full URL to the external javascript file.
 * @param $sync_cached_file
 *   Synchronize tracking code and update if remote file have changed.
 * @return mixed
 *   The path to the local javascript file on success, boolean FALSE on failure.
 */
function _googleanalytics_cache($location, $sync_cached_file = FALSE) {
  $path = 'public://googleanalytics';
  $file_destination = $path . '/' . basename($location);

  if (!file_exists($file_destination) || $sync_cached_file) {
    // Download the latest tracking code.
    $result = drupal_http_request($location);

    if ($result->code == 200) {
      if (file_exists($file_destination)) {
        // Synchronize tracking code and and replace local file if outdated.
        $data_hash_local = drupal_hash_base64(file_get_contents($file_destination));
        $data_hash_remote = drupal_hash_base64($result->data);
        // Check that the files directory is writable.
        if ($data_hash_local != $data_hash_remote && file_prepare_directory($path)) {
          // Save updated tracking code file to disk.
          file_unmanaged_save_data($result->data, $file_destination, FILE_EXISTS_REPLACE);
          watchdog('googleanalytics', 'Locally cached tracking code file has been updated.', array(), WATCHDOG_INFO);

          // Change query-strings on css/js files to enforce reload for all users.
          _drupal_flush_css_js();
        }
      }
      else {
        // Check that the files directory is writable.
        if (file_prepare_directory($path, FILE_CREATE_DIRECTORY)) {
          // There is no need to flush JS here as core refreshes JS caches
          // automatically, if new files are added.
          file_unmanaged_save_data($result->data, $file_destination, FILE_EXISTS_REPLACE);
          watchdog('googleanalytics', 'Locally cached tracking code file has been saved.', array(), WATCHDOG_INFO);

          // Return the local JS file path.
          return file_create_url($file_destination);
        }
      }
    }
  }
  else {
    // Return the local JS file path.
    return file_create_url($file_destination);
  }
}

/**
 * Delete cached files and directory.
 */
function googleanalytics_clear_js_cache() {
  $path = 'public://googleanalytics';
  if (file_prepare_directory($path)) {
    file_scan_directory($path, '/.*/', array('callback' => 'file_unmanaged_delete'));
    drupal_rmdir($path);

    // Change query-strings on css/js files to enforce reload for all users.
    _drupal_flush_css_js();

    watchdog('googleanalytics', 'Local cache has been purged.', array(), WATCHDOG_INFO);
  }
}

/**
 * Helper function for grabbing search keys. Function is missing in D7.
 *
 * http://api.drupal.org/api/function/search_get_keys/6
 */
function googleanalytics_search_get_keys() {
  static $return;
  if (!isset($return)) {
    // Extract keys as remainder of path
    // Note: support old GET format of searches for existing links.
    $path = explode('/', $_GET['q'], 3);
    $keys = empty($_REQUEST['keys']) ? '' : $_REQUEST['keys'];
    $return = count($path) == 3 ? $path[2] : $keys;
  }
  return $return;
}

/**
 * Tracking visibility check for an user object.
 *
 * @param $account
 *   A user object containing an array of roles to check.
 * @return boolean
 *   A decision on if the current user is being tracked by Google Analytics.
 */
function _googleanalytics_visibility_user($account) {

  $enabled = FALSE;

  // Disable tracking if caching is disabled or a visitors is logged in and
  // have opted out from tracking via DNT (Do-Not-Track) header.
  $dnt = (($account->uid || variable_get('cache', 0) == 0) && variable_get('googleanalytics_privacy_donottrack', 1) && !empty($_SERVER['HTTP_DNT'])) ? TRUE : FALSE;

  // Is current user a member of a role that should be tracked?
  if (!$dnt && _googleanalytics_visibility_roles($account)) {

    // Use the user's block visibility setting, if necessary.
    if (($custom = variable_get('googleanalytics_custom', 0)) != 0) {
      if ($account->uid && isset($account->data['googleanalytics']['custom'])) {
        $enabled = $account->data['googleanalytics']['custom'];
      }
      else {
        $enabled = ($custom == 1);
      }
    }
    else {
      $enabled = TRUE;
    }

  }

  return $enabled;
}

/**
 * Based on visibility setting this function returns TRUE if GA code should
 * be added for the current role and otherwise FALSE.
 */
function _googleanalytics_visibility_roles($account) {

  $visibility = variable_get('googleanalytics_visibility_roles', 0);
  $enabled = $visibility;
  $roles = variable_get('googleanalytics_roles', array());

  if (array_sum($roles) > 0) {
    // One or more roles are selected.
    foreach (array_keys($account->roles) as $rid) {
      // Is the current user a member of one of these roles?
      if (isset($roles[$rid]) && $rid == $roles[$rid]) {
        // Current user is a member of a role that should be tracked/excluded from tracking.
        $enabled = !$visibility;
        break;
      }
    }
  }
  else {
    // No role is selected for tracking, therefore all roles should be tracked.
    $enabled = TRUE;
  }

  return $enabled;
}

/**
 * Based on visibility setting this function returns TRUE if GA code should
 * be added to the current page and otherwise FALSE.
 */
function _googleanalytics_visibility_pages() {
  static $page_match;

  // Cache visibility result if function is called more than once.
  if (!isset($page_match)) {

    $visibility = variable_get('googleanalytics_visibility_pages', 0);
    $setting_pages = variable_get('googleanalytics_pages', '');

    // Match path if necessary.
    if (!empty($setting_pages)) {
      // Convert path to lowercase. This allows comparison of the same path
      // with different case. Ex: /Page, /page, /PAGE.
      $pages = drupal_strtolower($setting_pages);
      if ($visibility < 2) {
        // Convert the Drupal path to lowercase
        $path = drupal_strtolower(drupal_get_path_alias($_GET['q']));
        // Compare the lowercase internal and lowercase path alias (if any).
        $page_match = drupal_match_path($path, $pages);
        if ($path != $_GET['q']) {
          $page_match = $page_match || drupal_match_path($_GET['q'], $pages);
        }
        // When $visibility has a value of 0, the tracking code is displayed on
        // all pages except those listed in $pages. When set to 1, it
        // is displayed only on those pages listed in $pages.
        $page_match = !($visibility xor $page_match);
      }
      elseif (module_exists('php')) {
        $page_match = php_eval($setting_pages);
      }
      else {
        $page_match = FALSE;
      }
    }
    else {
      $page_match = TRUE;
    }

  }
  return $page_match;
}