webmaster@1: <?php
webmaster@1: // $Id: search.module,v 1.250.2.1 2008/02/07 16:42:03 goba Exp $
webmaster@1: 
webmaster@1: /**
webmaster@1:  * @file
webmaster@1:  * Enables site-wide keyword searching.
webmaster@1:  */
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Matches Unicode character classes to exclude from the search index.
webmaster@1:  *
webmaster@1:  * See: http://www.unicode.org/Public/UNIDATA/UCD.html#General_Category_Values
webmaster@1:  *
webmaster@1:  * The index only contains the following character classes:
webmaster@1:  * Lu     Letter, Uppercase
webmaster@1:  * Ll     Letter, Lowercase
webmaster@1:  * Lt     Letter, Titlecase
webmaster@1:  * Lo     Letter, Other
webmaster@1:  * Nd     Number, Decimal Digit
webmaster@1:  * No     Number, Other
webmaster@1:  */
webmaster@1: define('PREG_CLASS_SEARCH_EXCLUDE',
webmaster@1: '\x{0}-\x{2f}\x{3a}-\x{40}\x{5b}-\x{60}\x{7b}-\x{bf}\x{d7}\x{f7}\x{2b0}-'.
webmaster@1: '\x{385}\x{387}\x{3f6}\x{482}-\x{489}\x{559}-\x{55f}\x{589}-\x{5c7}\x{5f3}-'.
webmaster@1: '\x{61f}\x{640}\x{64b}-\x{65e}\x{66a}-\x{66d}\x{670}\x{6d4}\x{6d6}-\x{6ed}'.
webmaster@1: '\x{6fd}\x{6fe}\x{700}-\x{70f}\x{711}\x{730}-\x{74a}\x{7a6}-\x{7b0}\x{901}-'.
webmaster@1: '\x{903}\x{93c}\x{93e}-\x{94d}\x{951}-\x{954}\x{962}-\x{965}\x{970}\x{981}-'.
webmaster@1: '\x{983}\x{9bc}\x{9be}-\x{9cd}\x{9d7}\x{9e2}\x{9e3}\x{9f2}-\x{a03}\x{a3c}-'.
webmaster@1: '\x{a4d}\x{a70}\x{a71}\x{a81}-\x{a83}\x{abc}\x{abe}-\x{acd}\x{ae2}\x{ae3}'.
webmaster@1: '\x{af1}-\x{b03}\x{b3c}\x{b3e}-\x{b57}\x{b70}\x{b82}\x{bbe}-\x{bd7}\x{bf0}-'.
webmaster@1: '\x{c03}\x{c3e}-\x{c56}\x{c82}\x{c83}\x{cbc}\x{cbe}-\x{cd6}\x{d02}\x{d03}'.
webmaster@1: '\x{d3e}-\x{d57}\x{d82}\x{d83}\x{dca}-\x{df4}\x{e31}\x{e34}-\x{e3f}\x{e46}-'.
webmaster@1: '\x{e4f}\x{e5a}\x{e5b}\x{eb1}\x{eb4}-\x{ebc}\x{ec6}-\x{ecd}\x{f01}-\x{f1f}'.
webmaster@1: '\x{f2a}-\x{f3f}\x{f71}-\x{f87}\x{f90}-\x{fd1}\x{102c}-\x{1039}\x{104a}-'.
webmaster@1: '\x{104f}\x{1056}-\x{1059}\x{10fb}\x{10fc}\x{135f}-\x{137c}\x{1390}-\x{1399}'.
webmaster@1: '\x{166d}\x{166e}\x{1680}\x{169b}\x{169c}\x{16eb}-\x{16f0}\x{1712}-\x{1714}'.
webmaster@1: '\x{1732}-\x{1736}\x{1752}\x{1753}\x{1772}\x{1773}\x{17b4}-\x{17db}\x{17dd}'.
webmaster@1: '\x{17f0}-\x{180e}\x{1843}\x{18a9}\x{1920}-\x{1945}\x{19b0}-\x{19c0}\x{19c8}'.
webmaster@1: '\x{19c9}\x{19de}-\x{19ff}\x{1a17}-\x{1a1f}\x{1d2c}-\x{1d61}\x{1d78}\x{1d9b}-'.
webmaster@1: '\x{1dc3}\x{1fbd}\x{1fbf}-\x{1fc1}\x{1fcd}-\x{1fcf}\x{1fdd}-\x{1fdf}\x{1fed}-'.
webmaster@1: '\x{1fef}\x{1ffd}-\x{2070}\x{2074}-\x{207e}\x{2080}-\x{2101}\x{2103}-\x{2106}'.
webmaster@1: '\x{2108}\x{2109}\x{2114}\x{2116}-\x{2118}\x{211e}-\x{2123}\x{2125}\x{2127}'.
webmaster@1: '\x{2129}\x{212e}\x{2132}\x{213a}\x{213b}\x{2140}-\x{2144}\x{214a}-\x{2b13}'.
webmaster@1: '\x{2ce5}-\x{2cff}\x{2d6f}\x{2e00}-\x{3005}\x{3007}-\x{303b}\x{303d}-\x{303f}'.
webmaster@1: '\x{3099}-\x{309e}\x{30a0}\x{30fb}-\x{30fe}\x{3190}-\x{319f}\x{31c0}-\x{31cf}'.
webmaster@1: '\x{3200}-\x{33ff}\x{4dc0}-\x{4dff}\x{a015}\x{a490}-\x{a716}\x{a802}\x{a806}'.
webmaster@1: '\x{a80b}\x{a823}-\x{a82b}\x{d800}-\x{f8ff}\x{fb1e}\x{fb29}\x{fd3e}\x{fd3f}'.
webmaster@1: '\x{fdfc}-\x{fe6b}\x{feff}-\x{ff0f}\x{ff1a}-\x{ff20}\x{ff3b}-\x{ff40}\x{ff5b}-'.
webmaster@1: '\x{ff65}\x{ff70}\x{ff9e}\x{ff9f}\x{ffe0}-\x{fffd}');
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Matches all 'N' Unicode character classes (numbers)
webmaster@1:  */
webmaster@1: define('PREG_CLASS_NUMBERS',
webmaster@1: '\x{30}-\x{39}\x{b2}\x{b3}\x{b9}\x{bc}-\x{be}\x{660}-\x{669}\x{6f0}-\x{6f9}'.
webmaster@1: '\x{966}-\x{96f}\x{9e6}-\x{9ef}\x{9f4}-\x{9f9}\x{a66}-\x{a6f}\x{ae6}-\x{aef}'.
webmaster@1: '\x{b66}-\x{b6f}\x{be7}-\x{bf2}\x{c66}-\x{c6f}\x{ce6}-\x{cef}\x{d66}-\x{d6f}'.
webmaster@1: '\x{e50}-\x{e59}\x{ed0}-\x{ed9}\x{f20}-\x{f33}\x{1040}-\x{1049}\x{1369}-'.
webmaster@1: '\x{137c}\x{16ee}-\x{16f0}\x{17e0}-\x{17e9}\x{17f0}-\x{17f9}\x{1810}-\x{1819}'.
webmaster@1: '\x{1946}-\x{194f}\x{2070}\x{2074}-\x{2079}\x{2080}-\x{2089}\x{2153}-\x{2183}'.
webmaster@1: '\x{2460}-\x{249b}\x{24ea}-\x{24ff}\x{2776}-\x{2793}\x{3007}\x{3021}-\x{3029}'.
webmaster@1: '\x{3038}-\x{303a}\x{3192}-\x{3195}\x{3220}-\x{3229}\x{3251}-\x{325f}\x{3280}-'.
webmaster@1: '\x{3289}\x{32b1}-\x{32bf}\x{ff10}-\x{ff19}');
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Matches all 'P' Unicode character classes (punctuation)
webmaster@1:  */
webmaster@1: define('PREG_CLASS_PUNCTUATION',
webmaster@1: '\x{21}-\x{23}\x{25}-\x{2a}\x{2c}-\x{2f}\x{3a}\x{3b}\x{3f}\x{40}\x{5b}-\x{5d}'.
webmaster@1: '\x{5f}\x{7b}\x{7d}\x{a1}\x{ab}\x{b7}\x{bb}\x{bf}\x{37e}\x{387}\x{55a}-\x{55f}'.
webmaster@1: '\x{589}\x{58a}\x{5be}\x{5c0}\x{5c3}\x{5f3}\x{5f4}\x{60c}\x{60d}\x{61b}\x{61f}'.
webmaster@1: '\x{66a}-\x{66d}\x{6d4}\x{700}-\x{70d}\x{964}\x{965}\x{970}\x{df4}\x{e4f}'.
webmaster@1: '\x{e5a}\x{e5b}\x{f04}-\x{f12}\x{f3a}-\x{f3d}\x{f85}\x{104a}-\x{104f}\x{10fb}'.
webmaster@1: '\x{1361}-\x{1368}\x{166d}\x{166e}\x{169b}\x{169c}\x{16eb}-\x{16ed}\x{1735}'.
webmaster@1: '\x{1736}\x{17d4}-\x{17d6}\x{17d8}-\x{17da}\x{1800}-\x{180a}\x{1944}\x{1945}'.
webmaster@1: '\x{2010}-\x{2027}\x{2030}-\x{2043}\x{2045}-\x{2051}\x{2053}\x{2054}\x{2057}'.
webmaster@1: '\x{207d}\x{207e}\x{208d}\x{208e}\x{2329}\x{232a}\x{23b4}-\x{23b6}\x{2768}-'.
webmaster@1: '\x{2775}\x{27e6}-\x{27eb}\x{2983}-\x{2998}\x{29d8}-\x{29db}\x{29fc}\x{29fd}'.
webmaster@1: '\x{3001}-\x{3003}\x{3008}-\x{3011}\x{3014}-\x{301f}\x{3030}\x{303d}\x{30a0}'.
webmaster@1: '\x{30fb}\x{fd3e}\x{fd3f}\x{fe30}-\x{fe52}\x{fe54}-\x{fe61}\x{fe63}\x{fe68}'.
webmaster@1: '\x{fe6a}\x{fe6b}\x{ff01}-\x{ff03}\x{ff05}-\x{ff0a}\x{ff0c}-\x{ff0f}\x{ff1a}'.
webmaster@1: '\x{ff1b}\x{ff1f}\x{ff20}\x{ff3b}-\x{ff3d}\x{ff3f}\x{ff5b}\x{ff5d}\x{ff5f}-'.
webmaster@1: '\x{ff65}');
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Matches all CJK characters that are candidates for auto-splitting
webmaster@1:  * (Chinese, Japanese, Korean).
webmaster@1:  * Contains kana and BMP ideographs.
webmaster@1:  */
webmaster@1: define('PREG_CLASS_CJK', '\x{3041}-\x{30ff}\x{31f0}-\x{31ff}\x{3400}-\x{4db5}'.
webmaster@1: '\x{4e00}-\x{9fbb}\x{f900}-\x{fad9}');
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Implementation of hook_help().
webmaster@1:  */
webmaster@1: function search_help($path, $arg) {
webmaster@1:   switch ($path) {
webmaster@1:     case 'admin/help#search':
webmaster@1:       $output = '<p>'. t('The search module adds the ability to search for content by keywords. Search is often the only practical way to find content on a large site, and is useful for finding both users and posts.') .'</p>';
webmaster@1:       $output .= '<p>'. t('To provide keyword searching, the search engine maintains an index of words found in your site\'s content. To build and maintain this index, a correctly configured <a href="@cron">cron maintenance task</a> is required. Indexing behavior can be adjusted using the <a href="@searchsettings">search settings page</a>; for example, the <em>Number of items to index per cron run</em> sets the maximum number of items indexed in each pass of a <a href="@cron">cron maintenance task</a>. If necessary, reduce this number to prevent timeouts and memory errors when indexing.', array('@cron' => url('admin/reports/status'), '@searchsettings' => url('admin/settings/search'))) .'</p>';
webmaster@1:       $output .= '<p>'. t('For more information, see the online handbook entry for <a href="@search">Search module</a>.', array('@search' => 'http://drupal.org/handbook/modules/search/')) .'</p>';
webmaster@1:       return $output;
webmaster@1:     case 'admin/settings/search':
webmaster@1:       return '<p>'. t('The search engine maintains an index of words found in your site\'s content. To build and maintain this index, a correctly configured <a href="@cron">cron maintenance task</a> is required. Indexing behavior can be adjusted using the settings below.', array('@cron' => url('admin/reports/status'))) .'</p>';
webmaster@1:     case 'search#noresults':
webmaster@1:       return t('<ul>
webmaster@1: <li>Check if your spelling is correct.</li>
webmaster@1: <li>Remove quotes around phrases to match each word individually: <em>"blue smurf"</em> will match less than <em>blue smurf</em>.</li>
webmaster@1: <li>Consider loosening your query with <em>OR</em>: <em>blue smurf</em> will match less than <em>blue OR smurf</em>.</li>
webmaster@1: </ul>');
webmaster@1:   }
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Implementation of hook_theme()
webmaster@1:  */
webmaster@1: function search_theme() {
webmaster@1:   return array(
webmaster@1:     'search_theme_form' => array(
webmaster@1:       'arguments' => array('form' => NULL),
webmaster@1:       'template' => 'search-theme-form',
webmaster@1:     ),
webmaster@1:     'search_block_form' => array(
webmaster@1:       'arguments' => array('form' => NULL),
webmaster@1:       'template' => 'search-block-form',
webmaster@1:     ),
webmaster@1:     'search_result' => array(
webmaster@1:       'arguments' => array('result' => NULL, 'type' => NULL),
webmaster@1:       'file' => 'search.pages.inc',
webmaster@1:       'template' => 'search-result',
webmaster@1:     ),
webmaster@1:     'search_results' => array(
webmaster@1:       'arguments' => array('results' => NULL, 'type' => NULL),
webmaster@1:       'file' => 'search.pages.inc',
webmaster@1:       'template' => 'search-results',
webmaster@1:     ),
webmaster@1:   );
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Implementation of hook_perm().
webmaster@1:  */
webmaster@1: function search_perm() {
webmaster@1:   return array('search content', 'use advanced search', 'administer search');
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Implementation of hook_block().
webmaster@1:  */
webmaster@1: function search_block($op = 'list', $delta = 0) {
webmaster@1:   if ($op == 'list') {
webmaster@1:     $blocks[0]['info'] = t('Search form');
webmaster@1:     // Not worth caching.
webmaster@1:     $blocks[0]['cache'] = BLOCK_NO_CACHE;
webmaster@1:     return $blocks;
webmaster@1:   }
webmaster@1:   else if ($op == 'view' && user_access('search content')) {
webmaster@1:     $block['content'] = drupal_get_form('search_block_form');
webmaster@1:     $block['subject'] = t('Search');
webmaster@1:     return $block;
webmaster@1:   }
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Implementation of hook_menu().
webmaster@1:  */
webmaster@1: function search_menu() {
webmaster@1:   $items['search'] = array(
webmaster@1:     'title' => 'Search',
webmaster@1:     'page callback' => 'search_view',
webmaster@1:     'access arguments' => array('search content'),
webmaster@1:     'type' => MENU_SUGGESTED_ITEM,
webmaster@1:     'file' => 'search.pages.inc',
webmaster@1:   );
webmaster@1:   $items['admin/settings/search'] = array(
webmaster@1:     'title' => 'Search settings',
webmaster@1:     'description' => 'Configure relevance settings for search and other indexing options',
webmaster@1:     'page callback' => 'drupal_get_form',
webmaster@1:     'page arguments' => array('search_admin_settings'),
webmaster@1:     'access arguments' => array('administer search'),
webmaster@1:     'type' => MENU_NORMAL_ITEM,
webmaster@1:     'file' => 'search.admin.inc',
webmaster@1:   );
webmaster@1:   $items['admin/settings/search/wipe'] = array(
webmaster@1:     'title' => 'Clear index',
webmaster@1:     'page callback' => 'drupal_get_form',
webmaster@1:     'page arguments' => array('search_wipe_confirm'),
webmaster@1:     'access arguments' => array('administer search'),
webmaster@1:     'type' => MENU_CALLBACK,
webmaster@1:     'file' => 'search.admin.inc',
webmaster@1:   );
webmaster@1:   $items['admin/reports/search'] = array(
webmaster@1:     'title' => 'Top search phrases',
webmaster@1:     'description' => 'View most popular search phrases.',
webmaster@1:     'page callback' => 'dblog_top',
webmaster@1:     'page arguments' => array('search'),
webmaster@1:     'file' => 'dblog.admin.inc',
webmaster@1:     'file path' => drupal_get_path('module', 'dblog'),
webmaster@1:   );
webmaster@1: 
webmaster@1:   foreach (module_implements('search') as $name) {
webmaster@1:     $items['search/'. $name .'/%menu_tail'] = array(
webmaster@1:       'title callback' => 'module_invoke',
webmaster@1:       'title arguments' => array($name, 'search', 'name', TRUE),
webmaster@1:       'page callback' => 'search_view',
webmaster@1:       'page arguments' => array($name),
webmaster@1:       'access callback' => '_search_menu',
webmaster@1:       'access arguments' => array($name),
webmaster@1:       'type' => MENU_LOCAL_TASK,
webmaster@1:       'parent' => 'search',
webmaster@1:       'file' => 'search.pages.inc',
webmaster@1:     );
webmaster@1:   }
webmaster@1:   return $items;
webmaster@1: }
webmaster@1: 
webmaster@1: function _search_menu($name) {
webmaster@1:   return user_access('search content') && module_invoke($name, 'search', 'name');
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Wipes a part of or the entire search index.
webmaster@1:  *
webmaster@1:  * @param $sid
webmaster@1:  *  (optional) The SID of the item to wipe. If specified, $type must be passed
webmaster@1:  *  too.
webmaster@1:  * @param $type
webmaster@1:  *  (optional) The type of item to wipe.
webmaster@1:  */
webmaster@1: function search_wipe($sid = NULL, $type = NULL, $reindex = FALSE) {
webmaster@1:   if ($type == NULL && $sid == NULL) {
webmaster@1:     module_invoke_all('search', 'reset');
webmaster@1:   }
webmaster@1:   else {
webmaster@1:     db_query("DELETE FROM {search_dataset} WHERE sid = %d AND type = '%s'", $sid, $type);
webmaster@1:     db_query("DELETE FROM {search_index} WHERE sid = %d AND type = '%s'", $sid, $type);
webmaster@1:     // Don't remove links if re-indexing.
webmaster@1:     if (!$reindex) {
webmaster@1:       db_query("DELETE FROM {search_node_links} WHERE sid = %d AND type = '%s'", $sid, $type);
webmaster@1:     }
webmaster@1:   }
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Marks a word as dirty (or retrieves the list of dirty words). This is used
webmaster@1:  * during indexing (cron). Words which are dirty have outdated total counts in
webmaster@1:  * the search_total table, and need to be recounted.
webmaster@1:  */
webmaster@1: function search_dirty($word = NULL) {
webmaster@1:   static $dirty = array();
webmaster@1:   if ($word !== NULL) {
webmaster@1:     $dirty[$word] = TRUE;
webmaster@1:   }
webmaster@1:   else {
webmaster@1:     return $dirty;
webmaster@1:   }
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Implementation of hook_cron().
webmaster@1:  *
webmaster@1:  * Fires hook_update_index() in all modules and cleans up dirty words (see
webmaster@1:  * search_dirty).
webmaster@1:  */
webmaster@1: function search_cron() {
webmaster@1:   // We register a shutdown function to ensure that search_total is always up
webmaster@1:   // to date.
webmaster@1:   register_shutdown_function('search_update_totals');
webmaster@1: 
webmaster@1:   // Update word index
webmaster@1:   foreach (module_list() as $module) {
webmaster@1:     module_invoke($module, 'update_index');
webmaster@1:   }
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * This function is called on shutdown to ensure that search_total is always
webmaster@1:  * up to date (even if cron times out or otherwise fails).
webmaster@1:  */
webmaster@1: function search_update_totals() {
webmaster@1:   // Update word IDF (Inverse Document Frequency) counts for new/changed words
webmaster@1:   foreach (search_dirty() as $word => $dummy) {
webmaster@1:     // Get total count
webmaster@1:     $total = db_result(db_query("SELECT SUM(score) FROM {search_index} WHERE word = '%s'", $word));
webmaster@1:     // Apply Zipf's law to equalize the probability distribution
webmaster@1:     $total = log10(1 + 1/(max(1, $total)));
webmaster@1:     db_query("UPDATE {search_total} SET count = %f WHERE word = '%s'", $total, $word);
webmaster@1:     if (!db_affected_rows()) {
webmaster@1:       db_query("INSERT INTO {search_total} (word, count) VALUES ('%s', %f)", $word, $total);
webmaster@1:     }
webmaster@1:   }
webmaster@1:   // Find words that were deleted from search_index, but are still in
webmaster@1:   // search_total. We use a LEFT JOIN between the two tables and keep only the
webmaster@1:   // rows which fail to join.
webmaster@1:   $result = db_query("SELECT t.word AS realword, i.word FROM {search_total} t LEFT JOIN {search_index} i ON t.word = i.word WHERE i.word IS NULL");
webmaster@1:   while ($word = db_fetch_object($result)) {
webmaster@1:     db_query("DELETE FROM {search_total} WHERE word = '%s'", $word->realword);
webmaster@1:   }
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Simplifies a string according to indexing rules.
webmaster@1:  */
webmaster@1: function search_simplify($text) {
webmaster@1:   // Decode entities to UTF-8
webmaster@1:   $text = decode_entities($text);
webmaster@1: 
webmaster@1:   // Lowercase
webmaster@1:   $text = drupal_strtolower($text);
webmaster@1: 
webmaster@1:   // Call an external processor for word handling.
webmaster@1:   search_invoke_preprocess($text);
webmaster@1: 
webmaster@1:   // Simple CJK handling
webmaster@1:   if (variable_get('overlap_cjk', TRUE)) {
webmaster@1:     $text = preg_replace_callback('/['. PREG_CLASS_CJK .']+/u', 'search_expand_cjk', $text);
webmaster@1:   }
webmaster@1: 
webmaster@1:   // To improve searching for numerical data such as dates, IP addresses
webmaster@1:   // or version numbers, we consider a group of numerical characters
webmaster@1:   // separated only by punctuation characters to be one piece.
webmaster@1:   // This also means that searching for e.g. '20/03/1984' also returns
webmaster@1:   // results with '20-03-1984' in them.
webmaster@1:   // Readable regexp: ([number]+)[punctuation]+(?=[number])
webmaster@1:   $text = preg_replace('/(['. PREG_CLASS_NUMBERS .']+)['. PREG_CLASS_PUNCTUATION .']+(?=['. PREG_CLASS_NUMBERS .'])/u', '\1', $text);
webmaster@1: 
webmaster@1:   // The dot, underscore and dash are simply removed. This allows meaningful
webmaster@1:   // search behavior with acronyms and URLs.
webmaster@1:   $text = preg_replace('/[._-]+/', '', $text);
webmaster@1: 
webmaster@1:   // With the exception of the rules above, we consider all punctuation,
webmaster@1:   // marks, spacers, etc, to be a word boundary.
webmaster@1:   $text = preg_replace('/['. PREG_CLASS_SEARCH_EXCLUDE .']+/u', ' ', $text);
webmaster@1: 
webmaster@1:   return $text;
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Basic CJK tokenizer. Simply splits a string into consecutive, overlapping
webmaster@1:  * sequences of characters ('minimum_word_size' long).
webmaster@1:  */
webmaster@1: function search_expand_cjk($matches) {
webmaster@1:   $min = variable_get('minimum_word_size', 3);
webmaster@1:   $str = $matches[0];
webmaster@1:   $l = drupal_strlen($str);
webmaster@1:   // Passthrough short words
webmaster@1:   if ($l <= $min) {
webmaster@1:     return ' '. $str .' ';
webmaster@1:   }
webmaster@1:   $tokens = ' ';
webmaster@1:   // FIFO queue of characters
webmaster@1:   $chars = array();
webmaster@1:   // Begin loop
webmaster@1:   for ($i = 0; $i < $l; ++$i) {
webmaster@1:     // Grab next character
webmaster@1:     $current = drupal_substr($str, 0, 1);
webmaster@1:     $str = substr($str, strlen($current));
webmaster@1:     $chars[] = $current;
webmaster@1:     if ($i >= $min - 1) {
webmaster@1:       $tokens .= implode('', $chars) .' ';
webmaster@1:       array_shift($chars);
webmaster@1:     }
webmaster@1:   }
webmaster@1:   return $tokens;
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Splits a string into tokens for indexing.
webmaster@1:  */
webmaster@1: function search_index_split($text) {
webmaster@1:   static $last = NULL;
webmaster@1:   static $lastsplit = NULL;
webmaster@1: 
webmaster@1:   if ($last == $text) {
webmaster@1:     return $lastsplit;
webmaster@1:   }
webmaster@1:   // Process words
webmaster@1:   $text = search_simplify($text);
webmaster@1:   $words = explode(' ', $text);
webmaster@1:   array_walk($words, '_search_index_truncate');
webmaster@1: 
webmaster@1:   // Save last keyword result
webmaster@1:   $last = $text;
webmaster@1:   $lastsplit = $words;
webmaster@1: 
webmaster@1:   return $words;
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Helper function for array_walk in search_index_split.
webmaster@1:  */
webmaster@1: function _search_index_truncate(&$text) {
webmaster@1:   $text = truncate_utf8($text, 50);
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Invokes hook_search_preprocess() in modules.
webmaster@1:  */
webmaster@1: function search_invoke_preprocess(&$text) {
webmaster@1:   foreach (module_implements('search_preprocess') as $module) {
webmaster@1:     $text = module_invoke($module, 'search_preprocess', $text);
webmaster@1:   }
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Update the full-text search index for a particular item.
webmaster@1:  *
webmaster@1:  * @param $sid
webmaster@1:  *   A number identifying this particular item (e.g. node id).
webmaster@1:  *
webmaster@1:  * @param $type
webmaster@1:  *   A string defining this type of item (e.g. 'node')
webmaster@1:  *
webmaster@1:  * @param $text
webmaster@1:  *   The content of this item. Must be a piece of HTML text.
webmaster@1:  *
webmaster@1:  * @ingroup search
webmaster@1:  */
webmaster@1: function search_index($sid, $type, $text) {
webmaster@1:   $minimum_word_size = variable_get('minimum_word_size', 3);
webmaster@1: 
webmaster@1:   // Link matching
webmaster@1:   global $base_url;
webmaster@1:   $node_regexp = '@href=[\'"]?(?:'. preg_quote($base_url, '@') .'/|'. preg_quote(base_path(), '@') .')(?:\?q=)?/?((?![a-z]+:)[^\'">]+)[\'">]@i';
webmaster@1: 
webmaster@1:   // Multipliers for scores of words inside certain HTML tags.
webmaster@1:   // Note: 'a' must be included for link ranking to work.
webmaster@1:   $tags = array('h1' => 25,
webmaster@1:                 'h2' => 18,
webmaster@1:                 'h3' => 15,
webmaster@1:                 'h4' => 12,
webmaster@1:                 'h5' => 9,
webmaster@1:                 'h6' => 6,
webmaster@1:                 'u' => 3,
webmaster@1:                 'b' => 3,
webmaster@1:                 'i' => 3,
webmaster@1:                 'strong' => 3,
webmaster@1:                 'em' => 3,
webmaster@1:                 'a' => 10);
webmaster@1: 
webmaster@1:   // Strip off all ignored tags to speed up processing, but insert space before/after
webmaster@1:   // them to keep word boundaries.
webmaster@1:   $text = str_replace(array('<', '>'), array(' <', '> '), $text);
webmaster@1:   $text = strip_tags($text, '<'. implode('><', array_keys($tags)) .'>');
webmaster@1: 
webmaster@1:   // Split HTML tags from plain text.
webmaster@1:   $split = preg_split('/\s*<([^>]+?)>\s*/', $text, -1, PREG_SPLIT_DELIM_CAPTURE);
webmaster@1:   // Note: PHP ensures the array consists of alternating delimiters and literals
webmaster@1:   // and begins and ends with a literal (inserting $null as required).
webmaster@1: 
webmaster@1:   $tag = FALSE; // Odd/even counter. Tag or no tag.
webmaster@1:   $link = FALSE; // State variable for link analyser
webmaster@1:   $score = 1; // Starting score per word
webmaster@1:   $accum = ' '; // Accumulator for cleaned up data
webmaster@1:   $tagstack = array(); // Stack with open tags
webmaster@1:   $tagwords = 0; // Counter for consecutive words
webmaster@1:   $focus = 1; // Focus state
webmaster@1: 
webmaster@1:   $results = array(0 => array()); // Accumulator for words for index
webmaster@1: 
webmaster@1:   foreach ($split as $value) {
webmaster@1:     if ($tag) {
webmaster@1:       // Increase or decrease score per word based on tag
webmaster@1:       list($tagname) = explode(' ', $value, 2);
webmaster@1:       $tagname = drupal_strtolower($tagname);
webmaster@1:       // Closing or opening tag?
webmaster@1:       if ($tagname[0] == '/') {
webmaster@1:         $tagname = substr($tagname, 1);
webmaster@1:         // If we encounter unexpected tags, reset score to avoid incorrect boosting.
webmaster@1:         if (!count($tagstack) || $tagstack[0] != $tagname) {
webmaster@1:           $tagstack = array();
webmaster@1:           $score = 1;
webmaster@1:         }
webmaster@1:         else {
webmaster@1:           // Remove from tag stack and decrement score
webmaster@1:           $score = max(1, $score - $tags[array_shift($tagstack)]);
webmaster@1:         }
webmaster@1:         if ($tagname == 'a') {
webmaster@1:           $link = FALSE;
webmaster@1:         }
webmaster@1:       }
webmaster@1:       else {
webmaster@1:         if (isset($tagstack[0]) && $tagstack[0] == $tagname) {
webmaster@1:           // None of the tags we look for make sense when nested identically.
webmaster@1:           // If they are, it's probably broken HTML.
webmaster@1:           $tagstack = array();
webmaster@1:           $score = 1;
webmaster@1:         }
webmaster@1:         else {
webmaster@1:           // Add to open tag stack and increment score
webmaster@1:           array_unshift($tagstack, $tagname);
webmaster@1:           $score += $tags[$tagname];
webmaster@1:         }
webmaster@1:         if ($tagname == 'a') {
webmaster@1:           // Check if link points to a node on this site
webmaster@1:           if (preg_match($node_regexp, $value, $match)) {
webmaster@1:             $path = drupal_get_normal_path($match[1]);
webmaster@1:             if (preg_match('!(?:node|book)/(?:view/)?([0-9]+)!i', $path, $match)) {
webmaster@1:               $linknid = $match[1];
webmaster@1:               if ($linknid > 0) {
webmaster@1:                 // Note: ignore links to uncachable nodes to avoid redirect bugs.
webmaster@1:                 $node = db_fetch_object(db_query('SELECT n.title, n.nid, n.vid, r.format FROM {node} n INNER JOIN {node_revisions} r ON n.vid = r.vid WHERE n.nid = %d', $linknid));
webmaster@1:                 if (filter_format_allowcache($node->format)) {
webmaster@1:                   $link = TRUE;
webmaster@1:                   $linktitle = $node->title;
webmaster@1:                 }
webmaster@1:               }
webmaster@1:             }
webmaster@1:           }
webmaster@1:         }
webmaster@1:       }
webmaster@1:       // A tag change occurred, reset counter.
webmaster@1:       $tagwords = 0;
webmaster@1:     }
webmaster@1:     else {
webmaster@1:       // Note: use of PREG_SPLIT_DELIM_CAPTURE above will introduce empty values
webmaster@1:       if ($value != '') {
webmaster@1:         if ($link) {
webmaster@1:           // Check to see if the node link text is its URL. If so, we use the target node title instead.
webmaster@1:           if (preg_match('!^https?://!i', $value)) {
webmaster@1:             $value = $linktitle;
webmaster@1:           }
webmaster@1:         }
webmaster@1:         $words = search_index_split($value);
webmaster@1:         foreach ($words as $word) {
webmaster@1:           // Add word to accumulator
webmaster@1:           $accum .= $word .' ';
webmaster@1:           $num = is_numeric($word);
webmaster@1:           // Check wordlength
webmaster@1:           if ($num || drupal_strlen($word) >= $minimum_word_size) {
webmaster@1:             // Normalize numbers
webmaster@1:             if ($num) {
webmaster@1:               $word = (int)ltrim($word, '-0');
webmaster@1:             }
webmaster@1: 
webmaster@1:             // Links score mainly for the target.
webmaster@1:             if ($link) {
webmaster@1:               if (!isset($results[$linknid])) {
webmaster@1:                 $results[$linknid] = array();
webmaster@1:               }
webmaster@1:               $results[$linknid][] = $word;
webmaster@1:               // Reduce score of the link caption in the source.
webmaster@1:               $focus *= 0.2;
webmaster@1:             }
webmaster@1:             // Fall-through
webmaster@1:             if (!isset($results[0][$word])) {
webmaster@1:               $results[0][$word] = 0;
webmaster@1:             }
webmaster@1:             $results[0][$word] += $score * $focus;
webmaster@1: 
webmaster@1:             // Focus is a decaying value in terms of the amount of unique words up to this point.
webmaster@1:             // From 100 words and more, it decays, to e.g. 0.5 at 500 words and 0.3 at 1000 words.
webmaster@1:             $focus = min(1, .01 + 3.5 / (2 + count($results[0]) * .015));
webmaster@1:           }
webmaster@1:           $tagwords++;
webmaster@1:           // Too many words inside a single tag probably mean a tag was accidentally left open.
webmaster@1:           if (count($tagstack) && $tagwords >= 15) {
webmaster@1:             $tagstack = array();
webmaster@1:             $score = 1;
webmaster@1:           }
webmaster@1:         }
webmaster@1:       }
webmaster@1:     }
webmaster@1:     $tag = !$tag;
webmaster@1:   }
webmaster@1: 
webmaster@1:   search_wipe($sid, $type, TRUE);
webmaster@1: 
webmaster@1:   // Insert cleaned up data into dataset
webmaster@1:   db_query("INSERT INTO {search_dataset} (sid, type, data, reindex) VALUES (%d, '%s', '%s', %d)", $sid, $type, $accum, 0);
webmaster@1: 
webmaster@1:   // Insert results into search index
webmaster@1:   foreach ($results[0] as $word => $score) {
webmaster@1:     // The database will collate similar words (accented and non-accented forms, etc.),
webmaster@1:     // and the score is additive, so first add and then insert.
webmaster@1:     db_query("UPDATE {search_index} SET score = score + %d WHERE word = '%s' AND sid = '%d' AND type = '%s'", $score, $word, $sid, $type);
webmaster@1:     if (!db_affected_rows()) {
webmaster@1:       db_query("INSERT INTO {search_index} (word, sid, type, score) VALUES ('%s', %d, '%s', %f)", $word, $sid, $type, $score);
webmaster@1:     }
webmaster@1:     search_dirty($word);
webmaster@1:   }
webmaster@1:   unset($results[0]);
webmaster@1: 
webmaster@1:   // Get all previous links from this item.
webmaster@1:   $result = db_query("SELECT nid, caption FROM {search_node_links} WHERE sid = %d AND type = '%s'", $sid, $type);
webmaster@1:   $links = array();
webmaster@1:   while ($link = db_fetch_object($result)) {
webmaster@1:     $links[$link->nid] = $link->caption;
webmaster@1:   }
webmaster@1: 
webmaster@1:   // Now store links to nodes.
webmaster@1:   foreach ($results as $nid => $words) {
webmaster@1:     $caption = implode(' ', $words);
webmaster@1:     if (isset($links[$nid])) {
webmaster@1:       if ($links[$nid] != $caption) {
webmaster@1:         // Update the existing link and mark the node for reindexing.
webmaster@1:         db_query("UPDATE {search_node_links} SET caption = '%s' WHERE sid = %d AND type = '%s' AND nid = %d", $caption, $sid, $type, $nid);
webmaster@1:         search_touch_node($nid);
webmaster@1:       }
webmaster@1:       // Unset the link to mark it as processed.
webmaster@1:       unset($links[$nid]);
webmaster@1:     }
webmaster@1:     else {
webmaster@1:       // Insert the existing link and mark the node for reindexing.
webmaster@1:       db_query("INSERT INTO {search_node_links} (caption, sid, type, nid) VALUES ('%s', %d, '%s', %d)", $caption, $sid, $type, $nid);
webmaster@1:       search_touch_node($nid);
webmaster@1:     }
webmaster@1:   }
webmaster@1:   // Any left-over links in $links no longer exist. Delete them and mark the nodes for reindexing.
webmaster@1:   foreach ($links as $nid => $caption) {
webmaster@1:     db_query("DELETE FROM {search_node_links} WHERE sid = %d AND type = '%s' AND nid = %d", $sid, $type, $nid);
webmaster@1:     search_touch_node($nid);
webmaster@1:   }
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Change a node's changed timestamp to 'now' to force reindexing.
webmaster@1:  *
webmaster@1:  * @param $nid
webmaster@1:  *   The nid of the node that needs reindexing.
webmaster@1:  */
webmaster@1: function search_touch_node($nid) {
webmaster@1:   db_query("UPDATE {search_dataset} SET reindex = %d WHERE sid = %d AND type = 'node'", time(), $nid);
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Implementation of hook_nodeapi().
webmaster@1:  */
webmaster@1: function search_nodeapi(&$node, $op, $teaser = NULL, $page = NULL) {
webmaster@1:   switch ($op) {
webmaster@1:     // Transplant links to a node into the target node.
webmaster@1:     case 'update index':
webmaster@1:       $result = db_query("SELECT caption FROM {search_node_links} WHERE nid = %d", $node->nid);
webmaster@1:       $output = array();
webmaster@1:       while ($link = db_fetch_object($result)) {
webmaster@1:         $output[] = $link->caption;
webmaster@1:       }
webmaster@1:       return '<a>('. implode(', ', $output) .')</a>';
webmaster@1:     // Reindex the node when it is updated.  The node is automatically indexed
webmaster@1:     // when it is added, simply by being added to the node table.
webmaster@1:     case 'update':
webmaster@1:       search_touch_node($node->nid);
webmaster@1:       break;
webmaster@1:   }
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Implementation of hook_comment().
webmaster@1:  */
webmaster@1: function search_comment($a1, $op) {
webmaster@1:   switch ($op) {
webmaster@1:     // Reindex the node when comments are added or changed
webmaster@1:     case 'insert':
webmaster@1:     case 'update':
webmaster@1:     case 'delete':
webmaster@1:     case 'publish':
webmaster@1:     case 'unpublish':
webmaster@1:       search_touch_node(is_array($a1) ? $a1['nid'] : $a1->nid);
webmaster@1:       break;
webmaster@1:   }
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Extract a module-specific search option from a search query. e.g. 'type:book'
webmaster@1:  */
webmaster@1: function search_query_extract($keys, $option) {
webmaster@1:   if (preg_match('/(^| )'. $option .':([^ ]*)( |$)/i', $keys, $matches)) {
webmaster@1:     return $matches[2];
webmaster@1:   }
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Return a query with the given module-specific search option inserted in.
webmaster@1:  * e.g. 'type:book'.
webmaster@1:  */
webmaster@1: function search_query_insert($keys, $option, $value = '') {
webmaster@1:   if (search_query_extract($keys, $option)) {
webmaster@1:     $keys = trim(preg_replace('/(^| )'. $option .':[^ ]*/i', '', $keys));
webmaster@1:   }
webmaster@1:   if ($value != '') {
webmaster@1:     $keys .= ' '. $option .':'. $value;
webmaster@1:   }
webmaster@1:   return $keys;
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Parse a search query into SQL conditions.
webmaster@1:  *
webmaster@1:  * We build two queries that matches the dataset bodies. @See do_search for
webmaster@1:  * more about these.
webmaster@1:  *
webmaster@1:  * @param $text
webmaster@1:  *   The search keys.
webmaster@1:  * @return
webmaster@1:  *   A list of six elements.
webmaster@1:  *    * A series of statements AND'd together which will be used to provide all
webmaster@1:  *      possible matches.
webmaster@1:  *    * Arguments for this query part.
webmaster@1:  *    * A series of exact word matches OR'd together.
webmaster@1:  *    * Arguments for this query part.
webmaster@1:  *    * A bool indicating whether this is a simple query or not. Negative
webmaster@1:  *      terms, presence of both AND / OR make this FALSE.
webmaster@1:  *    * A bool indicating the presence of a lowercase or. Maybe the user
webmaster@1:  *      wanted to use OR.
webmaster@1:  */
webmaster@1: function search_parse_query($text) {
webmaster@1:   $keys = array('positive' => array(), 'negative' => array());
webmaster@1: 
webmaster@1:   // Tokenize query string
webmaster@1:   preg_match_all('/ (-?)("[^"]+"|[^" ]+)/i', ' '. $text, $matches, PREG_SET_ORDER);
webmaster@1: 
webmaster@1:   if (count($matches) < 1) {
webmaster@1:     return NULL;
webmaster@1:   }
webmaster@1: 
webmaster@1:   // Classify tokens
webmaster@1:   $or = FALSE;
webmaster@1:   $warning = '';
webmaster@1:   $simple = TRUE;
webmaster@1:   foreach ($matches as $match) {
webmaster@1:     $phrase = FALSE;
webmaster@1:     // Strip off phrase quotes
webmaster@1:     if ($match[2]{0} == '"') {
webmaster@1:       $match[2] = substr($match[2], 1, -1);
webmaster@1:       $phrase = TRUE;
webmaster@1:       $simple = FALSE;
webmaster@1:     }
webmaster@1:     // Simplify keyword according to indexing rules and external preprocessors
webmaster@1:     $words = search_simplify($match[2]);
webmaster@1:     // Re-explode in case simplification added more words, except when matching a phrase
webmaster@1:     $words = $phrase ? array($words) : preg_split('/ /', $words, -1, PREG_SPLIT_NO_EMPTY);
webmaster@1:     // Negative matches
webmaster@1:     if ($match[1] == '-') {
webmaster@1:       $keys['negative'] = array_merge($keys['negative'], $words);
webmaster@1:     }
webmaster@1:     // OR operator: instead of a single keyword, we store an array of all
webmaster@1:     // OR'd keywords.
webmaster@1:     elseif ($match[2] == 'OR' && count($keys['positive'])) {
webmaster@1:       $last = array_pop($keys['positive']);
webmaster@1:       // Starting a new OR?
webmaster@1:       if (!is_array($last)) {
webmaster@1:         $last = array($last);
webmaster@1:       }
webmaster@1:       $keys['positive'][] = $last;
webmaster@1:       $or = TRUE;
webmaster@1:       continue;
webmaster@1:     }
webmaster@1:     // AND operator: implied, so just ignore it
webmaster@1:     elseif ($match[2] == 'AND' || $match[2] == 'and') {
webmaster@1:       $warning = $match[2];
webmaster@1:       continue;
webmaster@1:     }
webmaster@1: 
webmaster@1:     // Plain keyword
webmaster@1:     else {
webmaster@1:       if ($match[2] == 'or') {
webmaster@1:         $warning = $match[2];
webmaster@1:       }
webmaster@1:       if ($or) {
webmaster@1:         // Add to last element (which is an array)
webmaster@1:         $keys['positive'][count($keys['positive']) - 1] = array_merge($keys['positive'][count($keys['positive']) - 1], $words);
webmaster@1:       }
webmaster@1:       else {
webmaster@1:         $keys['positive'] = array_merge($keys['positive'], $words);
webmaster@1:       }
webmaster@1:     }
webmaster@1:     $or = FALSE;
webmaster@1:   }
webmaster@1: 
webmaster@1:   // Convert keywords into SQL statements.
webmaster@1:   $query = array();
webmaster@1:   $query2 = array();
webmaster@1:   $arguments = array();
webmaster@1:   $arguments2 = array();
webmaster@1:   $matches = 0;
webmaster@1:   $simple_and = FALSE;
webmaster@1:   $simple_or = FALSE;
webmaster@1:   // Positive matches
webmaster@1:   foreach ($keys['positive'] as $key) {
webmaster@1:     // Group of ORed terms
webmaster@1:     if (is_array($key) && count($key)) {
webmaster@1:       $simple_or = TRUE;
webmaster@1:       $queryor = array();
webmaster@1:       $any = FALSE;
webmaster@1:       foreach ($key as $or) {
webmaster@1:         list($q, $num_new_scores) = _search_parse_query($or, $arguments2);
webmaster@1:         $any |= $num_new_scores;
webmaster@1:         if ($q) {
webmaster@1:           $queryor[] = $q;
webmaster@1:           $arguments[] = $or;
webmaster@1:         }
webmaster@1:       }
webmaster@1:       if (count($queryor)) {
webmaster@1:         $query[] = '('. implode(' OR ', $queryor) .')';
webmaster@1:         // A group of OR keywords only needs to match once
webmaster@1:         $matches += ($any > 0);
webmaster@1:       }
webmaster@1:     }
webmaster@1:     // Single ANDed term
webmaster@1:     else {
webmaster@1:       $simple_and = TRUE;
webmaster@1:       list($q, $num_new_scores, $num_valid_words) = _search_parse_query($key, $arguments2);
webmaster@1:       if ($q) {
webmaster@1:         $query[] = $q;
webmaster@1:         $arguments[] = $key;
webmaster@1:         if (!$num_valid_words) {
webmaster@1:           $simple = FALSE;
webmaster@1:         }
webmaster@1:         // Each AND keyword needs to match at least once
webmaster@1:         $matches += $num_new_scores;
webmaster@1:       }
webmaster@1:     }
webmaster@1:   }
webmaster@1:   if ($simple_and && $simple_or) {
webmaster@1:     $simple = FALSE;
webmaster@1:   }
webmaster@1:   // Negative matches
webmaster@1:   foreach ($keys['negative'] as $key) {
webmaster@1:     list($q) = _search_parse_query($key, $arguments2, TRUE);
webmaster@1:     if ($q) {
webmaster@1:       $query[] = $q;
webmaster@1:       $arguments[] = $key;
webmaster@1:       $simple = FALSE;
webmaster@1:     }
webmaster@1:   }
webmaster@1:   $query = implode(' AND ', $query);
webmaster@1: 
webmaster@1:   // Build word-index conditions for the first pass
webmaster@1:   $query2 = substr(str_repeat("i.word = '%s' OR ", count($arguments2)), 0, -4);
webmaster@1: 
webmaster@1:   return array($query, $arguments, $query2, $arguments2, $matches, $simple, $warning);
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Helper function for search_parse_query();
webmaster@1:  */
webmaster@1: function _search_parse_query(&$word, &$scores, $not = FALSE) {
webmaster@1:   $num_new_scores = 0;
webmaster@1:   $num_valid_words = 0;
webmaster@1:   // Determine the scorewords of this word/phrase
webmaster@1:   if (!$not) {
webmaster@1:     $split = explode(' ', $word);
webmaster@1:     foreach ($split as $s) {
webmaster@1:       $num = is_numeric($s);
webmaster@1:       if ($num || drupal_strlen($s) >= variable_get('minimum_word_size', 3)) {
webmaster@1:         $s = $num ? ((int)ltrim($s, '-0')) : $s;
webmaster@1:         if (!isset($scores[$s])) {
webmaster@1:           $scores[$s] = $s;
webmaster@1:           $num_new_scores++;
webmaster@1:         }
webmaster@1:         $num_valid_words++;
webmaster@1:       }
webmaster@1:     }
webmaster@1:   }
webmaster@1:   // Return matching snippet and number of added words
webmaster@1:   return array("d.data ". ($not ? 'NOT ' : '') ."LIKE '%% %s %%'", $num_new_scores, $num_valid_words);
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Do a query on the full-text search index for a word or words.
webmaster@1:  *
webmaster@1:  * This function is normally only called by each module that support the
webmaster@1:  * indexed search (and thus, implements hook_update_index()).
webmaster@1:  *
webmaster@1:  * Results are retrieved in two logical passes. However, the two passes are
webmaster@1:  * joined together into a single query.  And in the case of most simple
webmaster@1:  * queries the second pass is not even used.
webmaster@1:  *
webmaster@1:  * The first pass selects a set of all possible matches, which has the benefit
webmaster@1:  * of also providing the exact result set for simple "AND" or "OR" searches.
webmaster@1:  *
webmaster@1:  * The second portion of the query further refines this set by verifying
webmaster@1:  * advanced text conditions (such negative or phrase matches)
webmaster@1:  *
webmaster@1:  * @param $keywords
webmaster@1:  *   A search string as entered by the user.
webmaster@1:  *
webmaster@1:  * @param $type
webmaster@1:  *   A string identifying the calling module.
webmaster@1:  *
webmaster@1:  * @param $join1
webmaster@1:  *   (optional) Inserted into the JOIN part of the first SQL query.
webmaster@1:  *   For example "INNER JOIN {node} n ON n.nid = i.sid".
webmaster@1:  *
webmaster@1:  * @param $where1
webmaster@1:  *   (optional) Inserted into the WHERE part of the first SQL query.
webmaster@1:  *   For example "(n.status > %d)".
webmaster@1:  *
webmaster@1:  * @param $arguments1
webmaster@1:  *   (optional) Extra SQL arguments belonging to the first query.
webmaster@1:  *
webmaster@1:  * @param $columns2
webmaster@1:  *   (optional) Inserted into the SELECT pat of the second query. Must contain
webmaster@1:  *   a column selected as 'score'.
webmaster@1:  *   defaults to 'i.relevance AS score'
webmaster@1:  *
webmaster@1:  * @param $join2
webmaster@1:  *   (optional) Inserted into the JOIN par of the second SQL query.
webmaster@1:  *   For example "INNER JOIN {node_comment_statistics} n ON n.nid = i.sid"
webmaster@1:  *
webmaster@1:  * @param $arguments2
webmaster@1:  *   (optional) Extra SQL arguments belonging to the second query parameter.
webmaster@1:  *
webmaster@1:  * @param $sort_parameters
webmaster@1:  *   (optional) SQL arguments for sorting the final results.
webmaster@1:  *              Default: 'ORDER BY score DESC'
webmaster@1:  *
webmaster@1:  * @return
webmaster@1:  *   An array of SIDs for the search results.
webmaster@1:  *
webmaster@1:  * @ingroup search
webmaster@1:  */
webmaster@1: function do_search($keywords, $type, $join1 = '', $where1 = '1', $arguments1 = array(), $columns2 = 'i.relevance AS score', $join2 = '', $arguments2 = array(), $sort_parameters = 'ORDER BY score DESC') {
webmaster@1:   $query = search_parse_query($keywords);
webmaster@1: 
webmaster@1:   if ($query[2] == '') {
webmaster@1:     form_set_error('keys', t('You must include at least one positive keyword with @count characters or more.', array('@count' => variable_get('minimum_word_size', 3))));
webmaster@1:   }
webmaster@1:   if ($query[6]) {
webmaster@1:     if ($query[6] == 'or') {
webmaster@1:       drupal_set_message(t('Search for either of the two terms with uppercase <strong>OR</strong>. For example, <strong>cats OR dogs</strong>.'));
webmaster@1:     }
webmaster@1:   }
webmaster@1:   if ($query === NULL || $query[0] == '' || $query[2] == '') {
webmaster@1:     return array();
webmaster@1:   }
webmaster@1: 
webmaster@1:   // Build query for keyword normalization.
webmaster@1:   $conditions = "$where1 AND ($query[2]) AND i.type = '%s'";
webmaster@1:   $arguments1 = array_merge($arguments1, $query[3], array($type));
webmaster@1:   $join = "INNER JOIN {search_total} t ON i.word = t.word $join1";
webmaster@1:   if (!$query[5]) {
webmaster@1:     $conditions .= " AND ($query[0])";
webmaster@1:     $arguments1 = array_merge($arguments1, $query[1]);
webmaster@1:     $join .= " INNER JOIN {search_dataset} d ON i.sid = d.sid AND i.type = d.type";
webmaster@1:   }
webmaster@1: 
webmaster@1:   // Calculate maximum keyword relevance, to normalize it.
webmaster@1:   $select = "SELECT SUM(i.score * t.count) AS score FROM {search_index} i $join WHERE $conditions GROUP BY i.type, i.sid HAVING COUNT(*) >= %d ORDER BY score DESC";
webmaster@1:   $arguments = array_merge($arguments1, array($query[4]));
webmaster@1:   $normalize = db_result(db_query_range($select, $arguments, 0, 1));
webmaster@1:   if (!$normalize) {
webmaster@1:     return array();
webmaster@1:   }
webmaster@1:   $columns2 = str_replace('i.relevance', '('. (1.0 / $normalize) .' * SUM(i.score * t.count))', $columns2);
webmaster@1: 
webmaster@1:   // Build query to retrieve results.
webmaster@1:   $select = "SELECT i.type, i.sid, $columns2 FROM {search_index} i $join $join2 WHERE $conditions GROUP BY i.type, i.sid HAVING COUNT(*) >= %d";
webmaster@1:   $count_select =  "SELECT COUNT(*) FROM ($select) n1";
webmaster@1:   $arguments = array_merge($arguments2, $arguments1, array($query[4]));
webmaster@1: 
webmaster@1:   // Do actual search query
webmaster@1:   $result = pager_query("$select $sort_parameters", 10, 0, $count_select, $arguments);
webmaster@1:   $results = array();
webmaster@1:   while ($item = db_fetch_object($result)) {
webmaster@1:     $results[] = $item;
webmaster@1:   }
webmaster@1:   return $results;
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Helper function for grabbing search keys.
webmaster@1:  */
webmaster@1: function search_get_keys() {
webmaster@1:   static $return;
webmaster@1:   if (!isset($return)) {
webmaster@1:     // Extract keys as remainder of path
webmaster@1:     // Note: support old GET format of searches for existing links.
webmaster@1:     $path = explode('/', $_GET['q'], 3);
webmaster@1:     $keys = empty($_REQUEST['keys']) ? '' : $_REQUEST['keys'];
webmaster@1:     $return = count($path) == 3 ? $path[2] : $keys;
webmaster@1:   }
webmaster@1:   return $return;
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * @defgroup search Search interface
webmaster@1:  * @{
webmaster@1:  * The Drupal search interface manages a global search mechanism.
webmaster@1:  *
webmaster@1:  * Modules may plug into this system to provide searches of different types of
webmaster@1:  * data. Most of the system is handled by search.module, so this must be enabled
webmaster@1:  * for all of the search features to work.
webmaster@1:  *
webmaster@1:  * There are three ways to interact with the search system:
webmaster@1:  * - Specifically for searching nodes, you can implement nodeapi('update index')
webmaster@1:  *   and nodeapi('search result'). However, note that the search system already
webmaster@1:  *   indexes all visible output of a node, i.e. everything displayed normally
webmaster@1:  *   by hook_view() and hook_nodeapi('view'). This is usually sufficient.
webmaster@1:  *   You should only use this mechanism if you want additional, non-visible data
webmaster@1:  *   to be indexed.
webmaster@1:  * - Implement hook_search(). This will create a search tab for your module on
webmaster@1:  *   the /search page with a simple keyword search form. You may optionally
webmaster@1:  *   implement hook_search_item() to customize the display of your results.
webmaster@1:  * - Implement hook_update_index(). This allows your module to use Drupal's
webmaster@1:  *   HTML indexing mechanism for searching full text efficiently.
webmaster@1:  *
webmaster@1:  * If your module needs to provide a more complicated search form, then you need
webmaster@1:  * to implement it yourself without hook_search(). In that case, you should
webmaster@1:  * define it as a local task (tab) under the /search page (e.g. /search/mymodule)
webmaster@1:  * so that users can easily find it.
webmaster@1:  */
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Render a search form.
webmaster@1:  *
webmaster@1:  * @param $action
webmaster@1:  *   Form action. Defaults to "search".
webmaster@1:  * @param $keys
webmaster@1:  *   The search string entered by the user, containing keywords for the search.
webmaster@1:  * @param $type
webmaster@1:  *   The type of search to render the node for. Must be the name of module
webmaster@1:  *   which implements hook_search(). Defaults to 'node'.
webmaster@1:  * @param $prompt
webmaster@1:  *   A piece of text to put before the form (e.g. "Enter your keywords")
webmaster@1:  * @return
webmaster@1:  *   An HTML string containing the search form.
webmaster@1:  */
webmaster@1: function search_form(&$form_state, $action = '', $keys = '', $type = NULL, $prompt = NULL) {
webmaster@1: 
webmaster@1:   // Add CSS
webmaster@1:   drupal_add_css(drupal_get_path('module', 'search') .'/search.css', 'module', 'all', FALSE);
webmaster@1: 
webmaster@1:   if (!$action) {
webmaster@1:     $action = url('search/'. $type);
webmaster@1:   }
webmaster@1:   if (is_null($prompt)) {
webmaster@1:     $prompt = t('Enter your keywords');
webmaster@1:   }
webmaster@1: 
webmaster@1:   $form = array(
webmaster@1:     '#action' => $action,
webmaster@1:     '#attributes' => array('class' => 'search-form'),
webmaster@1:   );
webmaster@1:   $form['module'] = array('#type' => 'value', '#value' => $type);
webmaster@1:   $form['basic'] = array('#type' => 'item', '#title' => $prompt);
webmaster@1:   $form['basic']['inline'] = array('#prefix' => '<div class="container-inline">', '#suffix' => '</div>');
webmaster@1:   $form['basic']['inline']['keys'] = array(
webmaster@1:     '#type' => 'textfield',
webmaster@1:     '#title' => '',
webmaster@1:     '#default_value' => $keys,
webmaster@1:     '#size' => $prompt ? 40 : 20,
webmaster@1:     '#maxlength' => 255,
webmaster@1:   );
webmaster@1:   // processed_keys is used to coordinate keyword passing between other forms
webmaster@1:   // that hook into the basic search form.
webmaster@1:   $form['basic']['inline']['processed_keys'] = array('#type' => 'value', '#value' => array());
webmaster@1:   $form['basic']['inline']['submit'] = array('#type' => 'submit', '#value' => t('Search'));
webmaster@1: 
webmaster@1:   return $form;
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Form builder; Output a search form for the search block and the theme's search box.
webmaster@1:  *
webmaster@1:  * @ingroup forms
webmaster@1:  * @see search_box_form_submit()
webmaster@1:  * @see theme_search_box_form()
webmaster@1:  */
webmaster@1: function search_box(&$form_state, $form_id) {
webmaster@1:   $form[$form_id] = array(
webmaster@1:     '#title' => t('Search this site'),
webmaster@1:     '#type' => 'textfield',
webmaster@1:     '#size' => 15,
webmaster@1:     '#default_value' => '',
webmaster@1:     '#attributes' => array('title' => t('Enter the terms you wish to search for.')),
webmaster@1:   );
webmaster@1:   $form['submit'] = array('#type' => 'submit', '#value' => t('Search'));
webmaster@1:   $form['#submit'][] = 'search_box_form_submit';
webmaster@1:   $form['#validate'][] = 'search_box_form_validate';
webmaster@1: 
webmaster@1:   return $form;
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Process a block search form submission.
webmaster@1:  */
webmaster@1: function search_box_form_submit($form, &$form_state) {
webmaster@1:   $form_id = $form['form_id']['#value'];
webmaster@1:   $form_state['redirect'] = 'search/node/'. trim($form_state['values'][$form_id]);
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Process variables for search-theme-form.tpl.php.
webmaster@1:  *
webmaster@1:  * The $variables array contains the following arguments:
webmaster@1:  * - $form
webmaster@1:  *
webmaster@1:  * @see search-theme-form.tpl.php
webmaster@1:  */
webmaster@1: function template_preprocess_search_theme_form(&$variables) {
webmaster@1:   $variables['search'] = array();
webmaster@1:   $hidden = array();
webmaster@1:   // Provide variables named after form keys so themers can print each element independently.
webmaster@1:   foreach (element_children($variables['form']) as $key) {
webmaster@1:     $type = $variables['form'][$key]['#type'];
webmaster@1:     if ($type == 'hidden' || $type == 'token') {
webmaster@1:       $hidden[] = drupal_render($variables['form'][$key]);
webmaster@1:     }
webmaster@1:     else {
webmaster@1:       $variables['search'][$key] = drupal_render($variables['form'][$key]);
webmaster@1:     }
webmaster@1:   }
webmaster@1:   // Hidden form elements have no value to themers. No need for separation.
webmaster@1:   $variables['search']['hidden'] = implode($hidden);
webmaster@1:   // Collect all form elements to make it easier to print the whole form.
webmaster@1:   $variables['search_form'] = implode($variables['search']);
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Process variables for search-block-form.tpl.php.
webmaster@1:  *
webmaster@1:  * The $variables array contains the following arguments:
webmaster@1:  * - $form
webmaster@1:  *
webmaster@1:  * @see search-block-form.tpl.php
webmaster@1:  */
webmaster@1: function template_preprocess_search_block_form(&$variables) {
webmaster@1:   $variables['search'] = array();
webmaster@1:   $hidden = array();
webmaster@1:   // Provide variables named after form keys so themers can print each element independently.
webmaster@1:   foreach (element_children($variables['form']) as $key) {
webmaster@1:     $type = $variables['form'][$key]['#type'];
webmaster@1:     if ($type == 'hidden' || $type == 'token') {
webmaster@1:       $hidden[] = drupal_render($variables['form'][$key]);
webmaster@1:     }
webmaster@1:     else {
webmaster@1:       $variables['search'][$key] = drupal_render($variables['form'][$key]);
webmaster@1:     }
webmaster@1:   }
webmaster@1:   // Hidden form elements have no value to themers. No need for separation.
webmaster@1:   $variables['search']['hidden'] = implode($hidden);
webmaster@1:   // Collect all form elements to make it easier to print the whole form.
webmaster@1:   $variables['search_form'] = implode($variables['search']);
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Perform a standard search on the given keys, and return the formatted results.
webmaster@1:  */
webmaster@1: function search_data($keys = NULL, $type = 'node') {
webmaster@1: 
webmaster@1:   if (isset($keys)) {
webmaster@1:     if (module_hook($type, 'search')) {
webmaster@1:       $results = module_invoke($type, 'search', 'search', $keys);
webmaster@1:       if (isset($results) && is_array($results) && count($results)) {
webmaster@1:         if (module_hook($type, 'search_page')) {
webmaster@1:           return module_invoke($type, 'search_page', $results);
webmaster@1:         }
webmaster@1:         else {
webmaster@1:           return theme('search_results', $results, $type);
webmaster@1:         }
webmaster@1:       }
webmaster@1:     }
webmaster@1:   }
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Returns snippets from a piece of text, with certain keywords highlighted.
webmaster@1:  * Used for formatting search results.
webmaster@1:  *
webmaster@1:  * @param $keys
webmaster@1:  *   A string containing a search query.
webmaster@1:  *
webmaster@1:  * @param $text
webmaster@1:  *   The text to extract fragments from.
webmaster@1:  *
webmaster@1:  * @return
webmaster@1:  *   A string containing HTML for the excerpt.
webmaster@1:  */
webmaster@1: function search_excerpt($keys, $text) {
webmaster@1:   // We highlight around non-indexable or CJK characters.
webmaster@1:   $boundary = '(?:(?<=['. PREG_CLASS_SEARCH_EXCLUDE . PREG_CLASS_CJK .'])|(?=['. PREG_CLASS_SEARCH_EXCLUDE . PREG_CLASS_CJK .']))';
webmaster@1: 
webmaster@1:   // Extract positive keywords and phrases
webmaster@1:   preg_match_all('/ ("([^"]+)"|(?!OR)([^" ]+))/', ' '. $keys, $matches);
webmaster@1:   $keys = array_merge($matches[2], $matches[3]);
webmaster@1: 
webmaster@1:   // Prepare text
webmaster@1:   $text = ' '. strip_tags(str_replace(array('<', '>'), array(' <', '> '), $text)) .' ';
webmaster@1:   array_walk($keys, '_search_excerpt_replace');
webmaster@1:   $workkeys = $keys;
webmaster@1: 
webmaster@1:   // Extract a fragment per keyword for at most 4 keywords.
webmaster@1:   // First we collect ranges of text around each keyword, starting/ending
webmaster@1:   // at spaces.
webmaster@1:   // If the sum of all fragments is too short, we look for second occurrences.
webmaster@1:   $ranges = array();
webmaster@1:   $included = array();
webmaster@1:   $length = 0;
webmaster@1:   while ($length < 256 && count($workkeys)) {
webmaster@1:     foreach ($workkeys as $k => $key) {
webmaster@1:       if (strlen($key) == 0) {
webmaster@1:         unset($workkeys[$k]);
webmaster@1:         unset($keys[$k]);
webmaster@1:         continue;
webmaster@1:       }
webmaster@1:       if ($length >= 256) {
webmaster@1:         break;
webmaster@1:       }
webmaster@1:       // Remember occurrence of key so we can skip over it if more occurrences
webmaster@1:       // are desired.
webmaster@1:       if (!isset($included[$key])) {
webmaster@1:         $included[$key] = 0;
webmaster@1:       }
webmaster@1:       // Locate a keyword (position $p), then locate a space in front (position
webmaster@1:       // $q) and behind it (position $s)
webmaster@1:       if (preg_match('/'. $boundary . $key . $boundary .'/iu', $text, $match, PREG_OFFSET_CAPTURE, $included[$key])) {
webmaster@1:         $p = $match[0][1];
webmaster@1:         if (($q = strpos($text, ' ', max(0, $p - 60))) !== FALSE) {
webmaster@1:           $end = substr($text, $p, 80);
webmaster@1:           if (($s = strrpos($end, ' ')) !== FALSE) {
webmaster@1:             $ranges[$q] = $p + $s;
webmaster@1:             $length += $p + $s - $q;
webmaster@1:             $included[$key] = $p + 1;
webmaster@1:           }
webmaster@1:           else {
webmaster@1:             unset($workkeys[$k]);
webmaster@1:           }
webmaster@1:         }
webmaster@1:         else {
webmaster@1:           unset($workkeys[$k]);
webmaster@1:         }
webmaster@1:       }
webmaster@1:       else {
webmaster@1:         unset($workkeys[$k]);
webmaster@1:       }
webmaster@1:     }
webmaster@1:   }
webmaster@1: 
webmaster@1:   // If we didn't find anything, return the beginning.
webmaster@1:   if (count($ranges) == 0) {
webmaster@1:     return truncate_utf8($text, 256) .' ...';
webmaster@1:   }
webmaster@1: 
webmaster@1:   // Sort the text ranges by starting position.
webmaster@1:   ksort($ranges);
webmaster@1: 
webmaster@1:   // Now we collapse overlapping text ranges into one. The sorting makes it O(n).
webmaster@1:   $newranges = array();
webmaster@1:   foreach ($ranges as $from2 => $to2) {
webmaster@1:     if (!isset($from1)) {
webmaster@1:       $from1 = $from2;
webmaster@1:       $to1 = $to2;
webmaster@1:       continue;
webmaster@1:     }
webmaster@1:     if ($from2 <= $to1) {
webmaster@1:       $to1 = max($to1, $to2);
webmaster@1:     }
webmaster@1:     else {
webmaster@1:       $newranges[$from1] = $to1;
webmaster@1:       $from1 = $from2;
webmaster@1:       $to1 = $to2;
webmaster@1:     }
webmaster@1:   }
webmaster@1:   $newranges[$from1] = $to1;
webmaster@1: 
webmaster@1:   // Fetch text
webmaster@1:   $out = array();
webmaster@1:   foreach ($newranges as $from => $to) {
webmaster@1:     $out[] = substr($text, $from, $to - $from);
webmaster@1:   }
webmaster@1:   $text = (isset($newranges[0]) ? '' : '... ') . implode(' ... ', $out) .' ...';
webmaster@1: 
webmaster@1:   // Highlight keywords. Must be done at once to prevent conflicts ('strong' and '<strong>').
webmaster@1:   $text = preg_replace('/'. $boundary .'('. implode('|', $keys) .')'. $boundary .'/iu', '<strong>\0</strong>', $text);
webmaster@1:   return $text;
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * @} End of "defgroup search".
webmaster@1:  */
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Helper function for array_walk in search_except.
webmaster@1:  */
webmaster@1: function _search_excerpt_replace(&$text) {
webmaster@1:   $text = preg_quote($text, '/');
webmaster@1: }
webmaster@1: 
webmaster@1: function search_forms() {
webmaster@1:   $forms['search_theme_form']= array(
webmaster@1:     'callback' => 'search_box',
webmaster@1:     'callback arguments' => array('search_theme_form'),
webmaster@1:   );
webmaster@1:   $forms['search_block_form']= array(
webmaster@1:     'callback' => 'search_box',
webmaster@1:     'callback arguments' => array('search_block_form'),
webmaster@1:   );
webmaster@1:   return $forms;
webmaster@1: }