<?php

/**
 * Implementation of hook_menu().
 */
function google_cse_adv_menu() {
  $items = array();
  $items['admin/settings/google_cse_adv'] = array(
    'title' => 'Google CSE Advanced',
    'page callback' => 'drupal_get_form',
    'page arguments' => array('google_cse_adv_settings_form'),
    'access arguments' => array('administer search'),
    'type' => MENU_NORMAL_ITEM
  );
  return $items;
}

/**
 * Form callback.
 */
function google_cse_adv_settings_form() {
  $form = array();
  $form['google_cse_adv_cx'] = array(
    '#type' => 'textfield',
    '#title' => t('CX key'),
    '#description' => t('A unique code that identifies a custom search engine.'),
    '#size' => 50,
    '#maxlength' => 100,
    '#default_value' => variable_get('google_cse_adv_cx', '')
  );
  $form['google_cse_adv_language'] = array(
    '#type' => 'select',
    '#title' => t('Language awareness'),
    '#description' => t('Search results optimalization based on a language.'),
    '#options' => array(
      'none' => t('None'),
      'active' => t('Active language')
    ),
    '#default_value' => variable_get('google_cse_adv_language', 'none')
  );
  $form['google_cse_adv_populate_search_block'] = array(
    '#type' => 'checkbox',
    '#title' => t('Populate search block'),
    '#description' => t('Populate search block form with a current term.'),
    '#default_value' => variable_get('google_cse_adv_populate_search_block', 0)
  );
  $form['google_cse_adv_proxy'] = array(
    '#type' => 'textfield',
    '#title' => t('Proxy server'),
    '#description' => t('If your Drupal site lies behind a proxy server, set it\'s address here, e.g. "http://1.2.3.4:8080"'),
    '#size' => 50,
    '#maxlength' => 100,
    '#default_value' => variable_get('google_cse_adv_proxy', '')
  );
  $form['google_cse_adv_results_per_page'] = array(
    '#type' => 'textfield',
    '#title' => t('Number of results per page'),
    '#size' => 3,
    '#default_value' => variable_get('google_cse_adv_results_per_page', 10),
  );
  return system_settings_form($form);
}

/**
 * Settings form validation.
 */
function google_cse_adv_settings_form_validate($form, &$form_state) {
  // Ensure that results_per_page is a positive integer.
  $field = 'google_cse_adv_results_per_page';
  $value =& $form_state['values'][$field];
  if (!(is_numeric($value) && ($value = (int)$value) && $value > 0)) {
    form_set_error($field, t("Number of results per page must be a positive integer."));
  }
}

/**
 * Implementation of hook_menu_alter().
 *
 * @see apachesolr_search_menu_alter()
 */
function google_cse_adv_menu_alter(&$menu) {
  if (isset($menu['search/google_cse_adv/%menu_tail'])) {
    $menu['search']['page callback'] = 'google_cse_adv_search_view';
    $menu['search/google_cse_adv/%menu_tail']['page callback'] = 'google_cse_adv_search_view';
  }
  if (isset($menu['search/node/%menu_tail'])) {
    // Hide the node search tab.
    $menu['search/node/%menu_tail']['type'] = MENU_CALLBACK;
    unset($menu['search/node/%menu_tail']['title callback'], $menu['search/node/%menu_tail']['title arguments']);
    $menu['search/node/%menu_tail']['title'] = 'Search';
  }
  if (isset($menu['search/google_cse_adv/%menu_tail'])) {
    // Alter the google_cse_adv search tab
    $menu['search/google_cse_adv/%menu_tail']['weight'] = -10;
    $menu['search/google_cse_adv/%menu_tail']['title'] = 'Content';
  }
}

/**
 * Implementation of hook_form_[form_id]_alter().
 *
 * @see apachesolr_search_form_search_theme_form_alter()
 */
function google_cse_adv_form_search_theme_form_alter(&$form, $form_state) {
  google_cse_adv_form_search_block_form_alter($form, $form_state);
}

/**
 * Implementation of hook_form_[form_id]_alter().
 *
 * @see apachesolr_search_form_search_block_form_alter()
 */
function google_cse_adv_form_search_block_form_alter(&$form, $form_state) {
  if (!isset($form['#submit'])) {
    $form['#submit'] = array('google_cse_adv_search_box_form_submit');
  }
  else {
    $key = array_search('search_box_form_submit', $form['#submit']);
    if ($key !== FALSE) {
      // Replace the search module's function.
      $form['#submit'][$key] = 'google_cse_adv_search_box_form_submit';
    }
  }

  // Populate search block with a current term
  if (variable_get('google_cse_adv_populate_search_block', 0)) {
    $keys = trim(search_get_keys());

    // We have to check if we are on the search page - search_get_keys()
    // can return some string even if there is no search request
    if ($keys && arg(0) == 'search' && arg(1) == 'google_cse_adv') {
      // Classic block
      if (isset($form['search_block_form'])) {
        $form['search_block_form']['#default_value'] = $keys;
      }

      // Internal block
      if (isset($form['search_theme_form'])) {
        $form['search_theme_form']['#default_value'] = $keys;
      }
    }
  }
}

/**
 * Process a block search form submission.
 *
 * @see apachesolr_search_search_box_form_submit()
 */
function google_cse_adv_search_box_form_submit($form, &$form_state) {
  $form_id = $form['form_id']['#value'];
  $keys = $form_state['values'][$form_id];
  // Handle Apache webserver clean URL quirks.
  if (variable_get('clean_url', '0')) {
    $keys = str_replace('+', '%2B', $keys);
  }
  $form_state['redirect'] = 'search/google_cse_adv/'. trim($keys);
}

/**
 * Re-implementation of search_view().
 *
 * @see apachesolr_search_view()
 * @see search_view()
 */
function google_cse_adv_search_view($type = NULL) {
  // Search form submits with POST but redirects to GET.
  $results = '';

  if (!isset($_POST['form_id'])) {
    if (empty($type)) {
      // Note: search/X can not be a default tab because it would take on the
      // path of its parent (search). It would prevent remembering keywords when
      // switching tabs. This is why we drupal_goto to it from the parent instead.
      drupal_goto('search/google_cse_adv');
    }
    $keys = trim(search_get_keys());

    // Only perform search if there is non-whitespace search term:
    if ($keys) {
      // Log the search keys:
      $log = $keys;
      watchdog('search', '%keys (@type).', array('%keys' => $log, '@type' => t('Search')), WATCHDOG_NOTICE, l(t('results'), 'search/'. $type .'/'. $keys));

      // Collect the search results:
      $results = search_data($keys, $type);

      if ($results) {
        $results = theme('box', t('Search results'), $results);
      }
      else {
        $results = theme('box', t('Your search yielded no results'), search_help('search#noresults', drupal_help_arg()));
      }
    }
  }
  // Construct the search form.
  return drupal_get_form('search_form', NULL, $keys, $type) . $results;
}

/**
 * Implementation of hook_search().
 */
function google_cse_adv_search($op = 'search', $keys = NULL) {
  switch ($op) {
    case 'name':
      return t('Search');

    case 'search':
      $page = isset($_GET['page']) ? $_GET['page'] : 0;
      $response = google_cse_adv_execute($keys, $page);
      return google_cse_adv_response_results($response, $keys);
  }
}

/**
 * Sends query to Google's Custom Search Engine and returns response.
 * Can return cached response unless $reset param is set to TRUE (cached
 * upon one page load only) - useful for working with a block content.
 *
 * Returns a response string (XML).
 *
 * @param string $keys
 * @param int $page
 * @param bool $reset
 * @return string
 */
function google_cse_adv_execute($keys, $page, $reset = FALSE) {
  static $response = array(); // Cache
  if (isset($response[$keys]) && FALSE === $reset) {
    return $response[$keys];
  }

  // Number of results per page. 10 is the default for Google CSE.
  $rows = (int) variable_get('google_cse_adv_results_per_page', 10);
  // Language interface.
  $hl = google_cse_adv_param_hl();
  // Language filtering.
  $lr = google_cse_adv_param_lr();
  $url = 'http://www.google.com/cse'
		   . '?cx=' . urlencode(variable_get('google_cse_adv_cx', ''))
		   . '&client=google-csbe'
		   . '&output=xml_no_dtd'
       . ($hl ? '&hl=' . urlencode($hl) : '')
       . ($lr ? '&lr=' . urlencode($lr) : '')
       . '&q=' . urlencode($keys)
       .   (isset($_GET['more']) ? '+more:' . urlencode($_GET['more']) : '')
       . '&num=' . urlencode($rows)
       . '&start=' . urlencode($page * $rows);

  $request = new stdClass();
  $request->url = $url;

  module_invoke_all('google_cse_adv_execute', $request);

  $curl = curl_init();
  curl_setopt($curl, CURLOPT_URL, $request->url);
  curl_setopt($curl, CURLOPT_RETURNTRANSFER, 1); // Return into a variable
  if (($proxy = variable_get('google_cse_adv_proxy', '')) > '') {
    curl_setopt($curl, CURLOPT_PROXY, $proxy);
  }
  $response[$keys] = curl_exec($curl);
  curl_close($curl);

  watchdog('search', 'Google CSE request: ' . $request->url, array(), WATCHDOG_DEBUG);

  return $response[$keys];
}

function google_cse_adv_response_results($response, $keys) {
  $xml = simplexml_load_string($response);
  $results = array();

  if (isset($xml->RES->R)) {

    // Cap the result total if necessary.
    // Google will not return more than 1000 results, but RES->M may
    // be higher than this, which messes up our paging. Retain a copy
    // of the original so that themers can still display it.
    $max_results = variable_get('google_cse_adv_maximum_results', 1000);
    $total = (int) $xml->RES->M;
    $xml->RES->M_ORIGINAL = $total;
    if ($total > $max_results) {
      $xml->RES->M = $total = $max_results;
    }

    foreach($xml->RES->R as $result) {
      // Markup is escaped in title - we have to strip it (and strip_tags is not enough protection for XSS)
      $title = strip_tags((string)$result->T);
      $title = function_exists('htmlspecialchars_decode') ? htmlspecialchars_decode($title, ENT_QUOTES) : $title;
      $snippet = check_plain(strip_tags((string)$result->S));
      $snippet = function_exists('htmlspecialchars_decode') ? htmlspecialchars_decode($snippet, ENT_QUOTES) : $snippet;

      $results[] = array(
        'link' => (string)$result->U,
        'title' => $title,
        'snippet' => $snippet,
        'keys' => check_plain($keys),
      );
    }

    // No pager query was executed - we have to set the pager manually
    $limit = variable_get('google_cse_adv_results_per_page', 10);
    _google_cse_adv_pager_init($total, $limit);
  }
  return $results;
}

function google_cse_adv_response_filters($response) {
  $xml = simplexml_load_string($response);
  $results = array();

  if (isset($xml->Context->Facet)) {
    $filters = array();
    // Find "more:" (filter) part
    // We will take it from result rather then from URL because the request
    // can be changed during its generation. Also, "Q" property contains
    // an unescaped string so there is " more:" rather than "+more:".
    $matches = array();
    preg_match('/ more:([^ ]+)/', (string)$xml->Q, $matches);
    $more = isset($matches[1]) ? $matches[1] : '';

    foreach ($xml->Context->Facet as $val) {
      $label = (string)$val->FacetItem->label;
      $text  = t((string)$val->FacetItem->anchor_text);

      if ($label == $more) {
        $active = TRUE;
      } else {
        $active = FALSE;
      }
      $filters[] = array(
        'label' => $label,
        'text' => $text,
        'active' => $active
      );
    }
    $data = new stdClass();
    $data->filters = $filters;
    $data->response = $response;

    module_invoke_all('google_cse_adv_filters', $data);
    
    $filters = $data->filters;

    foreach ($filters as $filter) {
      $results[] = google_cse_adv_filter_l(
        '<span>' . t(check_plain($filter['text'])) . '</span>',
        $_GET['q'],
        array(
          'attributes' => array('class' => 'filter-link filter-link-' . check_plain($filter['label']), 'title' => check_plain($filter['label'])),
          'query' => array('more' => $filter['label']),
          'active' => $filter['active'],
          'html' => TRUE
        )
      );
    }

  }
  
  if (count($results)) {
    return theme('item_list', $results);
  }
  return '';
}

function google_cse_adv_response_suggestion($response) {
  $xml = simplexml_load_string($response);

  if (isset($xml->Spelling->Suggestion)) {
    $suggestions = array();

    foreach ($xml->Spelling->Suggestion as $suggestion) {
      $query = array(); // l() argument
      $q     = (string)$suggestion['q'];
      $term  = preg_replace('/ more:.*/i', '', $q);

      if (strpos($q, ' more:')) {
        $query['more'] = preg_replace('/.* more:/i', '', $q);
      }
      $url   = 'search/google_cse_adv/' . $term;
      $link  = l($term, $url, array('query' => $query));
      $suggestions[] = $link;
    }
    return theme('google_cse_adv_suggestion', $suggestions);
  }
  return '';
}

/**
 * Implementation of hook_block().
 */
function google_cse_adv_block($op = 'list', $delta = 0, $edit = array()) {
  switch ($op) {
    case 'list';
      return array(
        0 => array(
          'info' => t('Google CSE Advanced filters')
        ),
        1 => array(
          'info' => t('Google CSE Advanced suggestion')
        )
      );

    case 'view';
      switch ($delta) {
        case 0:
          return google_cse_adv_block_filters();

        case 1:
          return google_cse_adv_block_suggestion();
      }
  }
}

function google_cse_adv_block_filters() {
  $block = array(
    'subject' => t('Filters'),
    'content' => ''
  );
  $keys = search_get_keys();

  // Execute only if there is something to be found (keys)
  // Note: search_get_keys() can return result even if there
  // was no search request so we have to check if we are on
  // the search page.
  if (user_access('search content') && $keys
      && arg(0) == 'search' && arg(1) == 'google_cse_adv') {
    $page = isset($_GET['page']) ? $_GET['page'] : 0;
    $response = google_cse_adv_execute($keys, $page);
    $block['content'] = google_cse_adv_response_filters($response);
  }
  return $block;
}

function google_cse_adv_block_suggestion() {
  $block = array(
    'subject' => t('Spelling Suggestion'),
    'content' => ''
  );
  $keys = search_get_keys();

  // Execute only if there is something to be found (keys)
  // Note: search_get_keys() can return result even if there
  // was no search request so we have to check if we are on
  // the search page.
  if (user_access('search content') && $keys
      && arg(0) == 'search' && arg(1) == 'google_cse_adv') {
    $page = isset($_GET['page']) ? $_GET['page'] : 0;
    $response = google_cse_adv_execute($keys, $page);
    $block['content'] = google_cse_adv_response_suggestion($response);
  }
  return $block;
}

/**
 * Returns "hl" language param for search request.
 *
 * @return string
 */
function google_cse_adv_param_hl() {
  switch (variable_get('google_cse_adv_language', '')) {
    case 'active':
      global $language;
      return $language->language;

    default:
      return '';
  }
}

/**
 * Returns "lr" language param for search request.
 *
 * According to Google documentation, lr parameter text is 'lang_LANGPREFIX'.
 * We have to trust Drupal users that they don't use custom language prefixes
 * for their languages.
 * TODO: Support custom language prefixes?
 *
 * @see http://www.google.com/cse/docs/resultsxml.html#languageCollections
 *
 * @return string
 */
function google_cse_adv_param_lr() {
  switch (variable_get('google_cse_adv_language', '')) {
    case 'active':
      global $language;
      return 'lang_' . $language->language;

    default:
      return '';
  }
}

/**
 * A replacement for l()
 *  - doesn't add the default 'active' class because we have to check
 *    even the filter which may not be a part of 'q'
 *
 * @see l()
 */
function google_cse_adv_filter_l($text, $path, $options = array()) {
  global $language;

  // Merge in defaults.
  $options += array(
      'attributes' => array(),
      'html' => FALSE,
    );

  // Adds active
  if (!empty($options['active'])) {
    if (isset($options['attributes']['class'])) {
      $options['attributes']['class'] .= ' active';
    }
    else {
      $options['attributes']['class'] = 'active';
    }
  }

  // Remove all HTML and PHP tags from a tooltip. For best performance, we act only
  // if a quick strpos() pre-check gave a suspicion (because strip_tags() is expensive).
  if (isset($options['attributes']['title']) && strpos($options['attributes']['title'], '<') !== FALSE) {
    $options['attributes']['title'] = strip_tags($options['attributes']['title']);
  }

  return '<a href="'. check_url(url($path, $options)) .'"'. drupal_attributes($options['attributes']) .'>'. ($options['html'] ? $text : check_plain($text)) .'</a>';
}

/**
 * Implementation of hook_form_[form_id]_alter().
 *
 * This adds the 0 option to the search admin form.
 *
 * @see apachesolr_search_form_search_admin_settings_alter()
 */
function google_cse_adv_form_search_admin_settings_alter(&$form, $form_state) {
  $form['indexing_throttle']['search_cron_limit']['#options']['0'] = '0';
  ksort($form['indexing_throttle']['search_cron_limit']['#options']);
}

/**
 * Initialize a pager for theme('pager') without running an SQL query.
 *
 * @see apachesolr_pager_init()
 * @see pager_query()
 */
function _google_cse_adv_pager_init($total, $limit = 10, $element = 0) {
  global $pager_page_array, $pager_total, $pager_total_items;
  $page = isset($_GET['page']) ? $_GET['page'] : '';

  // Convert comma-separated $page to an array, used by other functions.
  $pager_page_array = explode(',', $page);

  // We calculate the total of pages as ceil(items / limit).
  $pager_total_items[$element] = $total;
  $pager_total[$element] = ceil($pager_total_items[$element] / $limit);
  $pager_page_array[$element] = max(0, min((int)$pager_page_array[$element], ((int)$pager_total[$element]) - 1));
  return $pager_page_array[$element];
}

/**
 * Implementation of hook_theme().
 */
function google_cse_adv_theme($existing, $type, $theme, $path) {
  return array(
    'google_cse_adv_suggestion' => array(
      'arguments' => array('suggestions' => NULL)
    ),
  );
}

/**
 * @param array $suggestions
 *   Suggestion links.
 * @return string
 */
function theme_google_cse_adv_suggestion($suggestions = NULL) {
  if (empty($suggestions)) {
    return '';
  }
  $items = array();

  foreach ($suggestions as $suggestion) {
    $items[] = $suggestion;
  }
  $output .= '<div class="mean">' . t('Did you mean:') . '</div>';
  $output .= theme('item_list', $items);
  return $output;
}