| Drupal | PHP Cross Reference | Content Management Systems |
1 <?php 2 3 /** 4 * @file 5 * Enables site-wide keyword searching. 6 */ 7 8 /** 9 * Matches all 'N' Unicode character classes (numbers) 10 */ 11 define('PREG_CLASS_NUMBERS', 12 '\x{30}-\x{39}\x{b2}\x{b3}\x{b9}\x{bc}-\x{be}\x{660}-\x{669}\x{6f0}-\x{6f9}' . 13 '\x{966}-\x{96f}\x{9e6}-\x{9ef}\x{9f4}-\x{9f9}\x{a66}-\x{a6f}\x{ae6}-\x{aef}' . 14 '\x{b66}-\x{b6f}\x{be7}-\x{bf2}\x{c66}-\x{c6f}\x{ce6}-\x{cef}\x{d66}-\x{d6f}' . 15 '\x{e50}-\x{e59}\x{ed0}-\x{ed9}\x{f20}-\x{f33}\x{1040}-\x{1049}\x{1369}-' . 16 '\x{137c}\x{16ee}-\x{16f0}\x{17e0}-\x{17e9}\x{17f0}-\x{17f9}\x{1810}-\x{1819}' . 17 '\x{1946}-\x{194f}\x{2070}\x{2074}-\x{2079}\x{2080}-\x{2089}\x{2153}-\x{2183}' . 18 '\x{2460}-\x{249b}\x{24ea}-\x{24ff}\x{2776}-\x{2793}\x{3007}\x{3021}-\x{3029}' . 19 '\x{3038}-\x{303a}\x{3192}-\x{3195}\x{3220}-\x{3229}\x{3251}-\x{325f}\x{3280}-' . 20 '\x{3289}\x{32b1}-\x{32bf}\x{ff10}-\x{ff19}'); 21 22 /** 23 * Matches all 'P' Unicode character classes (punctuation) 24 */ 25 define('PREG_CLASS_PUNCTUATION', 26 '\x{21}-\x{23}\x{25}-\x{2a}\x{2c}-\x{2f}\x{3a}\x{3b}\x{3f}\x{40}\x{5b}-\x{5d}' . 27 '\x{5f}\x{7b}\x{7d}\x{a1}\x{ab}\x{b7}\x{bb}\x{bf}\x{37e}\x{387}\x{55a}-\x{55f}' . 28 '\x{589}\x{58a}\x{5be}\x{5c0}\x{5c3}\x{5f3}\x{5f4}\x{60c}\x{60d}\x{61b}\x{61f}' . 29 '\x{66a}-\x{66d}\x{6d4}\x{700}-\x{70d}\x{964}\x{965}\x{970}\x{df4}\x{e4f}' . 30 '\x{e5a}\x{e5b}\x{f04}-\x{f12}\x{f3a}-\x{f3d}\x{f85}\x{104a}-\x{104f}\x{10fb}' . 31 '\x{1361}-\x{1368}\x{166d}\x{166e}\x{169b}\x{169c}\x{16eb}-\x{16ed}\x{1735}' . 32 '\x{1736}\x{17d4}-\x{17d6}\x{17d8}-\x{17da}\x{1800}-\x{180a}\x{1944}\x{1945}' . 33 '\x{2010}-\x{2027}\x{2030}-\x{2043}\x{2045}-\x{2051}\x{2053}\x{2054}\x{2057}' . 34 '\x{207d}\x{207e}\x{208d}\x{208e}\x{2329}\x{232a}\x{23b4}-\x{23b6}\x{2768}-' . 35 '\x{2775}\x{27e6}-\x{27eb}\x{2983}-\x{2998}\x{29d8}-\x{29db}\x{29fc}\x{29fd}' . 36 '\x{3001}-\x{3003}\x{3008}-\x{3011}\x{3014}-\x{301f}\x{3030}\x{303d}\x{30a0}' . 37 '\x{30fb}\x{fd3e}\x{fd3f}\x{fe30}-\x{fe52}\x{fe54}-\x{fe61}\x{fe63}\x{fe68}' . 38 '\x{fe6a}\x{fe6b}\x{ff01}-\x{ff03}\x{ff05}-\x{ff0a}\x{ff0c}-\x{ff0f}\x{ff1a}' . 39 '\x{ff1b}\x{ff1f}\x{ff20}\x{ff3b}-\x{ff3d}\x{ff3f}\x{ff5b}\x{ff5d}\x{ff5f}-' . 40 '\x{ff65}'); 41 42 /** 43 * Matches CJK (Chinese, Japanese, Korean) letter-like characters. 44 * 45 * This list is derived from the "East Asian Scripts" section of 46 * http://www.unicode.org/charts/index.html, as well as a comment on 47 * http://unicode.org/reports/tr11/tr11-11.html listing some character 48 * ranges that are reserved for additional CJK ideographs. 49 * 50 * The character ranges do not include numbers, punctuation, or symbols, since 51 * these are handled separately in search. Note that radicals and strokes are 52 * considered symbols. (See 53 * http://www.unicode.org/Public/UNIDATA/extracted/DerivedGeneralCategory.txt) 54 * 55 * @see search_expand_cjk() 56 */ 57 define('PREG_CLASS_CJK', '\x{1100}-\x{11FF}\x{3040}-\x{309F}\x{30A1}-\x{318E}' . 58 '\x{31A0}-\x{31B7}\x{31F0}-\x{31FF}\x{3400}-\x{4DBF}\x{4E00}-\x{9FCF}' . 59 '\x{A000}-\x{A48F}\x{A4D0}-\x{A4FD}\x{A960}-\x{A97F}\x{AC00}-\x{D7FF}' . 60 '\x{F900}-\x{FAFF}\x{FF21}-\x{FF3A}\x{FF41}-\x{FF5A}\x{FF66}-\x{FFDC}' . 61 '\x{20000}-\x{2FFFD}\x{30000}-\x{3FFFD}'); 62 63 /** 64 * Implements hook_help(). 65 */ 66 function search_help($path, $arg) { 67 switch ($path) { 68 case 'admin/help#search': 69 $output = ''; 70 $output .= '<h3>' . t('About') . '</h3>'; 71 $output .= '<p>' . t('The Search module provides the ability to index and search for content by exact keywords, and for users by username or e-mail. For more information, see the online handbook entry for <a href="@search-module">Search module</a>.', array('@search-module' => 'http://drupal.org/documentation/modules/search/', '@search' => url('search'))) . '</p>'; 72 $output .= '<h3>' . t('Uses') . '</h3>'; 73 $output .= '<dl>'; 74 $output .= '<dt>' . t('Searching content and users') . '</dt>'; 75 $output .= '<dd>' . t('Users with <em>Use search</em> permission can use the search block and <a href="@search">Search page</a>. Users with the <em>View published content</em> permission can search for content containing exact keywords. Users with the <em>View user profiles</em> permission can search for users containing the keyword anywhere in the user name, and users with the <em>Administer users</em> permission can search for users by email address. Additionally, users with <em>Use advanced search</em> permission can find content using more complex search methods and filtering by choosing the <em>Advanced search</em> option on the <a href="@search">Search page</a>.', array('@search' => url('search'))) . '</dd>'; 76 $output .= '<dt>' . t('Indexing content with cron') . '</dt>'; 77 $output .= '<dd>' . t('To provide keyword searching, the search engine maintains an index of words found in the content and its fields, along with text added to your content by other modules (such as comments from the core Comment module, and taxonomy terms from the core Taxonomy module). To build and maintain this index, a correctly configured <a href="@cron">cron maintenance task</a> is required. Users with <em>Administer search</em> permission can further configure the cron settings on the <a href="@searchsettings">Search settings page</a>.', array('@cron' => 'http://drupal.org/cron', '@searchsettings' => url('admin/config/search/settings'))) . '</dd>'; 78 $output .= '<dt>' . t('Content reindexing') . '</dt>'; 79 $output .= '<dd>' . t('Content-related actions on your site (creating, editing, or deleting content and comments) automatically cause affected content items to be marked for indexing or reindexing at the next cron run. When content is marked for reindexing, the previous content remains in the index until cron runs, at which time it is replaced by the new content. Unlike content-related actions, actions related to the structure of your site do not cause affected content to be marked for reindexing. Examples of structure-related actions that affect content include deleting or editing taxonomy terms, enabling or disabling modules that add text to content (such as Taxonomy, Comment, and field-providing modules), and modifying the fields or display parameters of your content types. If you take one of these actions and you want to ensure that the search index is updated to reflect your changed site structure, you can mark all content for reindexing by clicking the "Re-index site" button on the <a href="@searchsettings">Search settings page</a>. If you have a lot of content on your site, it may take several cron runs for the content to be reindexed.', array('@searchsettings' => url('admin/config/search/settings'))) . '</dd>'; 80 $output .= '<dt>' . t('Configuring search settings') . '</dt>'; 81 $output .= '<dd>' . t('Indexing behavior can be adjusted using the <a href="@searchsettings">Search settings page</a>. Users with <em>Administer search</em> permission can control settings such as the <em>Number of items to index per cron run</em>, <em>Indexing settings</em> (word length), <em>Active search modules</em>, and <em>Content ranking</em>, which lets you adjust the priority in which indexed content is returned in results.', array('@searchsettings' => url('admin/config/search/settings'))) . '</dd>'; 82 $output .= '<dt>' . t('Search block') . '</dt>'; 83 $output .= '<dd>' . t('The Search module includes a default <em>Search form</em> block, which can be enabled and configured on the <a href="@blocks">Blocks administration page</a>. The block is available to users with the <em>Search content</em> permission.', array('@blocks' => url('admin/structure/block'))) . '</dd>'; 84 $output .= '<dt>' . t('Extending Search module') . '</dt>'; 85 $output .= '<dd>' . t('By default, the Search module only supports exact keyword matching in content searches. You can modify this behavior by installing a language-specific stemming module for your language (such as <a href="http://drupal.org/project/porterstemmer">Porter Stemmer</a> for American English), which allows words such as walk, walking, and walked to be matched in the Search module. Another approach is to use a third-party search technology with stemming or partial word matching features built in, such as <a href="http://drupal.org/project/apachesolr">Apache Solr</a> or <a href="http://drupal.org/project/sphinx">Sphinx</a>. These and other <a href="@contrib-search">search-related contributed modules</a> can be downloaded by visiting Drupal.org.', array('@contrib-search' => 'http://drupal.org/project/modules?filters=tid%3A105')) . '</dd>'; 86 $output .= '</dl>'; 87 return $output; 88 case 'admin/config/search/settings': 89 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>'; 90 case 'search#noresults': 91 return t('<ul> 92 <li>Check if your spelling is correct.</li> 93 <li>Remove quotes around phrases to search for each word individually. <em>bike shed</em> will often show more results than <em>"bike shed"</em>.</li> 94 <li>Consider loosening your query with <em>OR</em>. <em>bike OR shed</em> will often show more results than <em>bike shed</em>.</li> 95 </ul>'); 96 } 97 } 98 99 /** 100 * Implements hook_theme(). 101 */ 102 function search_theme() { 103 return array( 104 'search_block_form' => array( 105 'render element' => 'form', 106 'template' => 'search-block-form', 107 ), 108 'search_result' => array( 109 'variables' => array('result' => NULL, 'module' => NULL), 110 'file' => 'search.pages.inc', 111 'template' => 'search-result', 112 ), 113 'search_results' => array( 114 'variables' => array('results' => NULL, 'module' => NULL), 115 'file' => 'search.pages.inc', 116 'template' => 'search-results', 117 ), 118 ); 119 } 120 121 /** 122 * Implements hook_permission(). 123 */ 124 function search_permission() { 125 return array( 126 'administer search' => array( 127 'title' => t('Administer search'), 128 ), 129 'search content' => array( 130 'title' => t('Use search'), 131 ), 132 'use advanced search' => array( 133 'title' => t('Use advanced search'), 134 ), 135 ); 136 } 137 138 /** 139 * Implements hook_block_info(). 140 */ 141 function search_block_info() { 142 $blocks['form']['info'] = t('Search form'); 143 // Not worth caching. 144 $blocks['form']['cache'] = DRUPAL_NO_CACHE; 145 $blocks['form']['properties']['administrative'] = TRUE; 146 return $blocks; 147 } 148 149 /** 150 * Implements hook_block_view(). 151 */ 152 function search_block_view($delta = '') { 153 if (user_access('search content')) { 154 $block['content'] = drupal_get_form('search_block_form'); 155 return $block; 156 } 157 } 158 159 /** 160 * Implements hook_menu(). 161 */ 162 function search_menu() { 163 $items['search'] = array( 164 'title' => 'Search', 165 'page callback' => 'search_view', 166 'access callback' => 'search_is_active', 167 'type' => MENU_SUGGESTED_ITEM, 168 'file' => 'search.pages.inc', 169 ); 170 $items['admin/config/search/settings'] = array( 171 'title' => 'Search settings', 172 'description' => 'Configure relevance settings for search and other indexing options.', 173 'page callback' => 'drupal_get_form', 174 'page arguments' => array('search_admin_settings'), 175 'access arguments' => array('administer search'), 176 'weight' => -10, 177 'file' => 'search.admin.inc', 178 ); 179 $items['admin/config/search/settings/reindex'] = array( 180 'title' => 'Clear index', 181 'page callback' => 'drupal_get_form', 182 'page arguments' => array('search_reindex_confirm'), 183 'access arguments' => array('administer search'), 184 'type' => MENU_VISIBLE_IN_BREADCRUMB, 185 'file' => 'search.admin.inc', 186 ); 187 188 // Add paths for searching. We add each module search path twice: once without 189 // and once with %menu_tail appended. The reason for this is that we want to 190 // preserve keywords when switching tabs, and also to have search tabs 191 // highlighted properly. The only way to do that within the Drupal menu 192 // system appears to be having two sets of tabs. See discussion on issue 193 // http://drupal.org/node/245103 for details. 194 195 drupal_static_reset('search_get_info'); 196 $default_info = search_get_default_module_info(); 197 if ($default_info) { 198 foreach (search_get_info() as $module => $search_info) { 199 $path = 'search/' . $search_info['path']; 200 $items[$path] = array( 201 'title' => $search_info['title'], 202 'page callback' => 'search_view', 203 'page arguments' => array($module, ''), 204 'access callback' => '_search_menu_access', 205 'access arguments' => array($module), 206 'type' => MENU_LOCAL_TASK, 207 'file' => 'search.pages.inc', 208 'weight' => $module == $default_info['module'] ? -10 : 0, 209 ); 210 $items["$path/%menu_tail"] = array( 211 'title' => $search_info['title'], 212 'load arguments' => array('%map', '%index'), 213 'page callback' => 'search_view', 214 'page arguments' => array($module, 2), 215 'access callback' => '_search_menu_access', 216 'access arguments' => array($module), 217 // The default local task points to its parent, but this item points to 218 // where it should so it should not be changed. 219 'type' => MENU_LOCAL_TASK, 220 'file' => 'search.pages.inc', 221 'weight' => 0, 222 // These tabs are not subtabs. 223 'tab_root' => 'search/' . $default_info['path'] . '/%', 224 // These tabs need to display at the same level. 225 'tab_parent' => 'search/' . $default_info['path'], 226 ); 227 } 228 } 229 return $items; 230 } 231 232 /** 233 * Determines access for the ?q=search path. 234 */ 235 function search_is_active() { 236 // This path cannot be accessed if there are no active modules. 237 return user_access('search content') && search_get_info(); 238 } 239 240 /** 241 * Returns information about available search modules. 242 * 243 * @param $all 244 * If TRUE, information about all enabled modules implementing 245 * hook_search_info() will be returned. If FALSE (default), only modules that 246 * have been set to active on the search settings page will be returned. 247 * 248 * @return 249 * Array of hook_search_info() return values, keyed by module name. The 250 * 'title' and 'path' array elements will be set to defaults for each module 251 * if not supplied by hook_search_info(), and an additional array element of 252 * 'module' will be added (set to the module name). 253 */ 254 function search_get_info($all = FALSE) { 255 $search_hooks = &drupal_static(__FUNCTION__); 256 257 if (!isset($search_hooks)) { 258 foreach (module_implements('search_info') as $module) { 259 $search_hooks[$module] = call_user_func($module . '_search_info'); 260 // Use module name as the default value. 261 $search_hooks[$module] += array('title' => $module, 'path' => $module); 262 // Include the module name itself in the array. 263 $search_hooks[$module]['module'] = $module; 264 } 265 } 266 267 if ($all) { 268 return $search_hooks; 269 } 270 271 $active = variable_get('search_active_modules', array('node', 'user')); 272 return array_intersect_key($search_hooks, array_flip($active)); 273 } 274 275 /** 276 * Returns information about the default search module. 277 * 278 * @return 279 * The search_get_info() array element for the default search module, if any. 280 */ 281 function search_get_default_module_info() { 282 $info = search_get_info(); 283 $default = variable_get('search_default_module', 'node'); 284 if (isset($info[$default])) { 285 return $info[$default]; 286 } 287 // The variable setting does not match any active module, so just return 288 // the info for the first active module (if any). 289 return reset($info); 290 } 291 292 /** 293 * Access callback for search tabs. 294 */ 295 function _search_menu_access($name) { 296 return user_access('search content') && (!function_exists($name . '_search_access') || module_invoke($name, 'search_access')); 297 } 298 299 /** 300 * Clears a part of or the entire search index. 301 * 302 * @param $sid 303 * (optional) The ID of the item to remove from the search index. If 304 * specified, $module must also be given. Omit both $sid and $module to clear 305 * the entire search index. 306 * @param $module 307 * (optional) The machine-readable name of the module for the item to remove 308 * from the search index. 309 */ 310 function search_reindex($sid = NULL, $module = NULL, $reindex = FALSE) { 311 if ($module == NULL && $sid == NULL) { 312 module_invoke_all('search_reset'); 313 } 314 else { 315 db_delete('search_dataset') 316 ->condition('sid', $sid) 317 ->condition('type', $module) 318 ->execute(); 319 db_delete('search_index') 320 ->condition('sid', $sid) 321 ->condition('type', $module) 322 ->execute(); 323 // Don't remove links if re-indexing. 324 if (!$reindex) { 325 db_delete('search_node_links') 326 ->condition('sid', $sid) 327 ->condition('type', $module) 328 ->execute(); 329 } 330 } 331 } 332 333 /** 334 * Marks a word as "dirty" (changed), or retrieves the list of dirty words. 335 * 336 * This is used during indexing (cron). Words that are dirty have outdated 337 * total counts in the search_total table, and need to be recounted. 338 */ 339 function search_dirty($word = NULL) { 340 $dirty = &drupal_static(__FUNCTION__, array()); 341 if ($word !== NULL) { 342 $dirty[$word] = TRUE; 343 } 344 else { 345 return $dirty; 346 } 347 } 348 349 /** 350 * Implements hook_cron(). 351 * 352 * Fires hook_update_index() in all modules and cleans up dirty words. 353 * 354 * @see search_dirty() 355 */ 356 function search_cron() { 357 // We register a shutdown function to ensure that search_total is always up 358 // to date. 359 drupal_register_shutdown_function('search_update_totals'); 360 361 foreach (variable_get('search_active_modules', array('node', 'user')) as $module) { 362 // Update word index 363 module_invoke($module, 'update_index'); 364 } 365 } 366 367 /** 368 * Updates the {search_total} database table. 369 * 370 * This function is called on shutdown to ensure that {search_total} is always 371 * up to date (even if cron times out or otherwise fails). 372 */ 373 function search_update_totals() { 374 // Update word IDF (Inverse Document Frequency) counts for new/changed words. 375 foreach (search_dirty() as $word => $dummy) { 376 // Get total count 377 $total = db_query("SELECT SUM(score) FROM {search_index} WHERE word = :word", array(':word' => $word), array('target' => 'slave'))->fetchField(); 378 // Apply Zipf's law to equalize the probability distribution. 379 $total = log10(1 + 1/(max(1, $total))); 380 db_merge('search_total') 381 ->key(array('word' => $word)) 382 ->fields(array('count' => $total)) 383 ->execute(); 384 } 385 // Find words that were deleted from search_index, but are still in 386 // search_total. We use a LEFT JOIN between the two tables and keep only the 387 // rows which fail to join. 388 $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", array(), array('target' => 'slave')); 389 $or = db_or(); 390 foreach ($result as $word) { 391 $or->condition('word', $word->realword); 392 } 393 if (count($or) > 0) { 394 db_delete('search_total') 395 ->condition($or) 396 ->execute(); 397 } 398 } 399 400 /** 401 * Simplifies a string according to indexing rules. 402 * 403 * @param $text 404 * Text to simplify. 405 * 406 * @return 407 * Simplified text. 408 * 409 * @see hook_search_preprocess() 410 */ 411 function search_simplify($text) { 412 // Decode entities to UTF-8 413 $text = decode_entities($text); 414 415 // Lowercase 416 $text = drupal_strtolower($text); 417 418 // Call an external processor for word handling. 419 search_invoke_preprocess($text); 420 421 // Simple CJK handling 422 if (variable_get('overlap_cjk', TRUE)) { 423 $text = preg_replace_callback('/[' . PREG_CLASS_CJK . ']+/u', 'search_expand_cjk', $text); 424 } 425 426 // To improve searching for numerical data such as dates, IP addresses 427 // or version numbers, we consider a group of numerical characters 428 // separated only by punctuation characters to be one piece. 429 // This also means that searching for e.g. '20/03/1984' also returns 430 // results with '20-03-1984' in them. 431 // Readable regexp: ([number]+)[punctuation]+(?=[number]) 432 $text = preg_replace('/([' . PREG_CLASS_NUMBERS . ']+)[' . PREG_CLASS_PUNCTUATION . ']+(?=[' . PREG_CLASS_NUMBERS . '])/u', '\1', $text); 433 434 // Multiple dot and dash groups are word boundaries and replaced with space. 435 // No need to use the unicode modifer here because 0-127 ASCII characters 436 // can't match higher UTF-8 characters as the leftmost bit of those are 1. 437 $text = preg_replace('/[.-]{2,}/', ' ', $text); 438 439 // The dot, underscore and dash are simply removed. This allows meaningful 440 // search behavior with acronyms and URLs. See unicode note directly above. 441 $text = preg_replace('/[._-]+/', '', $text); 442 443 // With the exception of the rules above, we consider all punctuation, 444 // marks, spacers, etc, to be a word boundary. 445 $text = preg_replace('/[' . PREG_CLASS_UNICODE_WORD_BOUNDARY . ']+/u', ' ', $text); 446 447 // Truncate everything to 50 characters. 448 $words = explode(' ', $text); 449 array_walk($words, '_search_index_truncate'); 450 $text = implode(' ', $words); 451 452 return $text; 453 } 454 455 /** 456 * Splits CJK (Chinese, Japanese, Korean) text into tokens. 457 * 458 * The Search module matches exact words, where a word is defined to be a 459 * sequence of characters delimited by spaces or punctuation. CJK languages are 460 * written in long strings of characters, though, not split up into words. So 461 * in order to allow search matching, we split up CJK text into tokens 462 * consisting of consecutive, overlapping sequences of characters whose length 463 * is equal to the 'minimum_word_size' variable. This tokenizing is only done if 464 * the 'overlap_cjk' variable is TRUE. 465 * 466 * @param $matches 467 * This function is a callback for preg_replace_callback(), which is called 468 * from search_simplify(). So, $matches is an array of regular expression 469 * matches, which means that $matches[0] contains the matched text -- a string 470 * of CJK characters to tokenize. 471 * 472 * @return 473 * Tokenized text, starting and ending with a space character. 474 */ 475 function search_expand_cjk($matches) { 476 $min = variable_get('minimum_word_size', 3); 477 $str = $matches[0]; 478 $length = drupal_strlen($str); 479 // If the text is shorter than the minimum word size, don't tokenize it. 480 if ($length <= $min) { 481 return ' ' . $str . ' '; 482 } 483 $tokens = ' '; 484 // Build a FIFO queue of characters. 485 $chars = array(); 486 for ($i = 0; $i < $length; $i++) { 487 // Add the next character off the beginning of the string to the queue. 488 $current = drupal_substr($str, 0, 1); 489 $str = substr($str, strlen($current)); 490 $chars[] = $current; 491 if ($i >= $min - 1) { 492 // Make a token of $min characters, and add it to the token string. 493 $tokens .= implode('', $chars) . ' '; 494 // Shift out the first character in the queue. 495 array_shift($chars); 496 } 497 } 498 return $tokens; 499 } 500 501 /** 502 * Simplifies and splits a string into tokens for indexing. 503 */ 504 function search_index_split($text) { 505 $last = &drupal_static(__FUNCTION__); 506 $lastsplit = &drupal_static(__FUNCTION__ . ':lastsplit'); 507 508 if ($last == $text) { 509 return $lastsplit; 510 } 511 // Process words 512 $text = search_simplify($text); 513 $words = explode(' ', $text); 514 515 // Save last keyword result 516 $last = $text; 517 $lastsplit = $words; 518 519 return $words; 520 } 521 522 /** 523 * Helper function for array_walk in search_index_split. 524 */ 525 function _search_index_truncate(&$text) { 526 if (is_numeric($text)) { 527 $text = ltrim($text, '0'); 528 } 529 $text = truncate_utf8($text, 50); 530 } 531 532 /** 533 * Invokes hook_search_preprocess() in modules. 534 */ 535 function search_invoke_preprocess(&$text) { 536 foreach (module_implements('search_preprocess') as $module) { 537 $text = module_invoke($module, 'search_preprocess', $text); 538 } 539 } 540 541 /** 542 * Update the full-text search index for a particular item. 543 * 544 * @param $sid 545 * An ID number identifying this particular item (e.g., node ID). 546 * @param $module 547 * The machine-readable name of the module that this item comes from (a module 548 * that implements hook_search_info()). 549 * @param $text 550 * The content of this item. Must be a piece of HTML or plain text. 551 * 552 * @ingroup search 553 */ 554 function search_index($sid, $module, $text) { 555 $minimum_word_size = variable_get('minimum_word_size', 3); 556 557 // Link matching 558 global $base_url; 559 $node_regexp = '@href=[\'"]?(?:' . preg_quote($base_url, '@') . '/|' . preg_quote(base_path(), '@') . ')(?:\?q=)?/?((?![a-z]+:)[^\'">]+)[\'">]@i'; 560 561 // Multipliers for scores of words inside certain HTML tags. The weights are stored 562 // in a variable so that modules can overwrite the default weights. 563 // Note: 'a' must be included for link ranking to work. 564 $tags = variable_get('search_tag_weights', array( 565 'h1' => 25, 566 'h2' => 18, 567 'h3' => 15, 568 'h4' => 12, 569 'h5' => 9, 570 'h6' => 6, 571 'u' => 3, 572 'b' => 3, 573 'i' => 3, 574 'strong' => 3, 575 'em' => 3, 576 'a' => 10)); 577 578 // Strip off all ignored tags to speed up processing, but insert space before/after 579 // them to keep word boundaries. 580 $text = str_replace(array('<', '>'), array(' <', '> '), $text); 581 $text = strip_tags($text, '<' . implode('><', array_keys($tags)) . '>'); 582 583 // Split HTML tags from plain text. 584 $split = preg_split('/\s*<([^>]+?)>\s*/', $text, -1, PREG_SPLIT_DELIM_CAPTURE); 585 // Note: PHP ensures the array consists of alternating delimiters and literals 586 // and begins and ends with a literal (inserting $null as required). 587 588 $tag = FALSE; // Odd/even counter. Tag or no tag. 589 $link = FALSE; // State variable for link analyzer 590 $score = 1; // Starting score per word 591 $accum = ' '; // Accumulator for cleaned up data 592 $tagstack = array(); // Stack with open tags 593 $tagwords = 0; // Counter for consecutive words 594 $focus = 1; // Focus state 595 596 $results = array(0 => array()); // Accumulator for words for index 597 598 foreach ($split as $value) { 599 if ($tag) { 600 // Increase or decrease score per word based on tag 601 list($tagname) = explode(' ', $value, 2); 602 $tagname = drupal_strtolower($tagname); 603 // Closing or opening tag? 604 if ($tagname[0] == '/') { 605 $tagname = substr($tagname, 1); 606 // If we encounter unexpected tags, reset score to avoid incorrect boosting. 607 if (!count($tagstack) || $tagstack[0] != $tagname) { 608 $tagstack = array(); 609 $score = 1; 610 } 611 else { 612 // Remove from tag stack and decrement score 613 $score = max(1, $score - $tags[array_shift($tagstack)]); 614 } 615 if ($tagname == 'a') { 616 $link = FALSE; 617 } 618 } 619 else { 620 if (isset($tagstack[0]) && $tagstack[0] == $tagname) { 621 // None of the tags we look for make sense when nested identically. 622 // If they are, it's probably broken HTML. 623 $tagstack = array(); 624 $score = 1; 625 } 626 else { 627 // Add to open tag stack and increment score 628 array_unshift($tagstack, $tagname); 629 $score += $tags[$tagname]; 630 } 631 if ($tagname == 'a') { 632 // Check if link points to a node on this site 633 if (preg_match($node_regexp, $value, $match)) { 634 $path = drupal_get_normal_path($match[1]); 635 if (preg_match('!(?:node|book)/(?:view/)?([0-9]+)!i', $path, $match)) { 636 $linknid = $match[1]; 637 if ($linknid > 0) { 638 $node = db_query('SELECT title, nid, vid FROM {node} WHERE nid = :nid', array(':nid' => $linknid), array('target' => 'slave'))->fetchObject(); 639 $link = TRUE; 640 $linktitle = $node->title; 641 } 642 } 643 } 644 } 645 } 646 // A tag change occurred, reset counter. 647 $tagwords = 0; 648 } 649 else { 650 // Note: use of PREG_SPLIT_DELIM_CAPTURE above will introduce empty values 651 if ($value != '') { 652 if ($link) { 653 // Check to see if the node link text is its URL. If so, we use the target node title instead. 654 if (preg_match('!^https?://!i', $value)) { 655 $value = $linktitle; 656 } 657 } 658 $words = search_index_split($value); 659 foreach ($words as $word) { 660 // Add word to accumulator 661 $accum .= $word . ' '; 662 // Check wordlength 663 if (is_numeric($word) || drupal_strlen($word) >= $minimum_word_size) { 664 // Links score mainly for the target. 665 if ($link) { 666 if (!isset($results[$linknid])) { 667 $results[$linknid] = array(); 668 } 669 $results[$linknid][] = $word; 670 // Reduce score of the link caption in the source. 671 $focus *= 0.2; 672 } 673 // Fall-through 674 if (!isset($results[0][$word])) { 675 $results[0][$word] = 0; 676 } 677 $results[0][$word] += $score * $focus; 678 679 // Focus is a decaying value in terms of the amount of unique words up to this point. 680 // From 100 words and more, it decays, to e.g. 0.5 at 500 words and 0.3 at 1000 words. 681 $focus = min(1, .01 + 3.5 / (2 + count($results[0]) * .015)); 682 } 683 $tagwords++; 684 // Too many words inside a single tag probably mean a tag was accidentally left open. 685 if (count($tagstack) && $tagwords >= 15) { 686 $tagstack = array(); 687 $score = 1; 688 } 689 } 690 } 691 } 692 $tag = !$tag; 693 } 694 695 search_reindex($sid, $module, TRUE); 696 697 // Insert cleaned up data into dataset 698 db_insert('search_dataset') 699 ->fields(array( 700 'sid' => $sid, 701 'type' => $module, 702 'data' => $accum, 703 'reindex' => 0, 704 )) 705 ->execute(); 706 707 // Insert results into search index 708 foreach ($results[0] as $word => $score) { 709 // If a word already exists in the database, its score gets increased 710 // appropriately. If not, we create a new record with the appropriate 711 // starting score. 712 db_merge('search_index') 713 ->key(array( 714 'word' => $word, 715 'sid' => $sid, 716 'type' => $module, 717 )) 718 ->fields(array('score' => $score)) 719 ->expression('score', 'score + :score', array(':score' => $score)) 720 ->execute(); 721 search_dirty($word); 722 } 723 unset($results[0]); 724 725 // Get all previous links from this item. 726 $result = db_query("SELECT nid, caption FROM {search_node_links} WHERE sid = :sid AND type = :type", array( 727 ':sid' => $sid, 728 ':type' => $module 729 ), array('target' => 'slave')); 730 $links = array(); 731 foreach ($result as $link) { 732 $links[$link->nid] = $link->caption; 733 } 734 735 // Now store links to nodes. 736 foreach ($results as $nid => $words) { 737 $caption = implode(' ', $words); 738 if (isset($links[$nid])) { 739 if ($links[$nid] != $caption) { 740 // Update the existing link and mark the node for reindexing. 741 db_update('search_node_links') 742 ->fields(array('caption' => $caption)) 743 ->condition('sid', $sid) 744 ->condition('type', $module) 745 ->condition('nid', $nid) 746 ->execute(); 747 search_touch_node($nid); 748 } 749 // Unset the link to mark it as processed. 750 unset($links[$nid]); 751 } 752 elseif ($sid != $nid || $module != 'node') { 753 // Insert the existing link and mark the node for reindexing, but don't 754 // reindex if this is a link in a node pointing to itself. 755 db_insert('search_node_links') 756 ->fields(array( 757 'caption' => $caption, 758 'sid' => $sid, 759 'type' => $module, 760 'nid' => $nid, 761 )) 762 ->execute(); 763 search_touch_node($nid); 764 } 765 } 766 // Any left-over links in $links no longer exist. Delete them and mark the nodes for reindexing. 767 foreach ($links as $nid => $caption) { 768 db_delete('search_node_links') 769 ->condition('sid', $sid) 770 ->condition('type', $module) 771 ->condition('nid', $nid) 772 ->execute(); 773 search_touch_node($nid); 774 } 775 } 776 777 /** 778 * Changes a node's changed timestamp to 'now' to force reindexing. 779 * 780 * @param $nid 781 * The node ID of the node that needs reindexing. 782 */ 783 function search_touch_node($nid) { 784 db_update('search_dataset') 785 ->fields(array('reindex' => REQUEST_TIME)) 786 ->condition('type', 'node') 787 ->condition('sid', $nid) 788 ->execute(); 789 } 790 791 /** 792 * Implements hook_node_update_index(). 793 */ 794 function search_node_update_index($node) { 795 // Transplant links to a node into the target node. 796 $result = db_query("SELECT caption FROM {search_node_links} WHERE nid = :nid", array(':nid' => $node->nid), array('target' => 'slave')); 797 $output = array(); 798 foreach ($result as $link) { 799 $output[] = $link->caption; 800 } 801 if (count($output)) { 802 return '<a>(' . implode(', ', $output) . ')</a>'; 803 } 804 } 805 806 /** 807 * Implements hook_node_update(). 808 */ 809 function search_node_update($node) { 810 // Reindex the node when it is updated. The node is automatically indexed 811 // when it is added, simply by being added to the node table. 812 search_touch_node($node->nid); 813 } 814 815 /** 816 * Implements hook_comment_insert(). 817 */ 818 function search_comment_insert($comment) { 819 // Reindex the node when comments are added. 820 search_touch_node($comment->nid); 821 } 822 823 /** 824 * Implements hook_comment_update(). 825 */ 826 function search_comment_update($comment) { 827 // Reindex the node when comments are changed. 828 search_touch_node($comment->nid); 829 } 830 831 /** 832 * Implements hook_comment_delete(). 833 */ 834 function search_comment_delete($comment) { 835 // Reindex the node when comments are deleted. 836 search_touch_node($comment->nid); 837 } 838 839 /** 840 * Implements hook_comment_publish(). 841 */ 842 function search_comment_publish($comment) { 843 // Reindex the node when comments are published. 844 search_touch_node($comment->nid); 845 } 846 847 /** 848 * Implements hook_comment_unpublish(). 849 */ 850 function search_comment_unpublish($comment) { 851 // Reindex the node when comments are unpublished. 852 search_touch_node($comment->nid); 853 } 854 855 /** 856 * Extracts a module-specific search option from a search expression. 857 * 858 * Search options are added using search_expression_insert(), and retrieved 859 * using search_expression_extract(). They take the form option:value, and 860 * are added to the ordinary keywords in the search expression. 861 * 862 * @param $expression 863 * The search expression to extract from. 864 * @param $option 865 * The name of the option to retrieve from the search expression. 866 * 867 * @return 868 * The value previously stored in the search expression for option $option, 869 * if any. Trailing spaces in values will not be included. 870 */ 871 function search_expression_extract($expression, $option) { 872 if (preg_match('/(^| )' . $option . ':([^ ]*)( |$)/i', $expression, $matches)) { 873 return $matches[2]; 874 } 875 } 876 877 /** 878 * Adds a module-specific search option to a search expression. 879 * 880 * Search options are added using search_expression_insert(), and retrieved 881 * using search_expression_extract(). They take the form option:value, and 882 * are added to the ordinary keywords in the search expression. 883 * 884 * @param $expression 885 * The search expression to add to. 886 * @param $option 887 * The name of the option to add to the search expression. 888 * @param $value 889 * The value to add for the option. If present, it will replace any previous 890 * value added for the option. Cannot contain any spaces or | characters, as 891 * these are used as delimiters. If you want to add a blank value $option: to 892 * the search expression, pass in an empty string or a string that is composed 893 * of only spaces. To clear a previously-stored option without adding a 894 * replacement, pass in NULL for $value or omit. 895 * 896 * @return 897 * $expression, with any previous value for this option removed, and a new 898 * $option:$value pair added if $value was provided. 899 */ 900 function search_expression_insert($expression, $option, $value = NULL) { 901 // Remove any previous values stored with $option. 902 $expression = trim(preg_replace('/(^| )' . $option . ':[^ ]*/i', '', $expression)); 903 904 // Set new value, if provided. 905 if (isset($value)) { 906 $expression .= ' ' . $option . ':' . trim($value); 907 } 908 return $expression; 909 } 910 911 /** 912 * @defgroup search Search interface 913 * @{ 914 * The Drupal search interface manages a global search mechanism. 915 * 916 * Modules may plug into this system to provide searches of different types of 917 * data. Most of the system is handled by search.module, so this must be enabled 918 * for all of the search features to work. 919 * 920 * There are three ways to interact with the search system: 921 * - Specifically for searching nodes, you can implement 922 * hook_node_update_index() and hook_node_search_result(). However, note that 923 * the search system already indexes all visible output of a node; i.e., 924 * everything displayed normally by hook_view() and hook_node_view(). This is 925 * usually sufficient. You should only use this mechanism if you want 926 * additional, non-visible data to be indexed. 927 * - Implement hook_search_info(). This will create a search tab for your module 928 * on the /search page with a simple keyword search form. You will also need 929 * to implement hook_search_execute() to perform the search. 930 * - Implement hook_update_index(). This allows your module to use Drupal's 931 * HTML indexing mechanism for searching full text efficiently. 932 * 933 * If your module needs to provide a more complicated search form, then you need 934 * to implement it yourself without hook_search_info(). In that case, you should 935 * define it as a local task (tab) under the /search page (e.g. /search/mymodule) 936 * so that users can easily find it. 937 */ 938 939 /** 940 * Builds a search form. 941 * 942 * @param $action 943 * Form action. Defaults to "search/$path", where $path is the search path 944 * associated with the module in its hook_search_info(). This will be 945 * run through url(). 946 * @param $keys 947 * The search string entered by the user, containing keywords for the search. 948 * @param $module 949 * The search module to render the form for: a module that implements 950 * hook_search_info(). If not supplied, the default search module is used. 951 * @param $prompt 952 * Label for the keywords field. Defaults to t('Enter your keywords') if NULL. 953 * Supply '' to omit. 954 * 955 * @return 956 * A Form API array for the search form. 957 */ 958 function search_form($form, &$form_state, $action = '', $keys = '', $module = NULL, $prompt = NULL) { 959 $module_info = FALSE; 960 if (!$module) { 961 $module_info = search_get_default_module_info(); 962 } 963 else { 964 $info = search_get_info(); 965 $module_info = isset($info[$module]) ? $info[$module] : FALSE; 966 } 967 968 // Sanity check. 969 if (!$module_info) { 970 form_set_error(NULL, t('Search is currently disabled.'), 'error'); 971 return $form; 972 } 973 974 if (!$action) { 975 $action = 'search/' . $module_info['path']; 976 } 977 if (!isset($prompt)) { 978 $prompt = t('Enter your keywords'); 979 } 980 981 $form['#action'] = url($action); 982 // Record the $action for later use in redirecting. 983 $form_state['action'] = $action; 984 $form['#attributes']['class'][] = 'search-form'; 985 $form['module'] = array('#type' => 'value', '#value' => $module); 986 $form['basic'] = array('#type' => 'container', '#attributes' => array('class' => array('container-inline'))); 987 $form['basic']['keys'] = array( 988 '#type' => 'textfield', 989 '#title' => $prompt, 990 '#default_value' => $keys, 991 '#size' => $prompt ? 40 : 20, 992 '#maxlength' => 255, 993 ); 994 // processed_keys is used to coordinate keyword passing between other forms 995 // that hook into the basic search form. 996 $form['basic']['processed_keys'] = array('#type' => 'value', '#value' => ''); 997 $form['basic']['submit'] = array('#type' => 'submit', '#value' => t('Search')); 998 999 return $form; 1000 } 1001 1002 /** 1003 * Form builder; Output a search form for the search block's search box. 1004 * 1005 * @ingroup forms 1006 * @see search_box_form_submit() 1007 * @see search-block-form.tpl.php 1008 */ 1009 function search_box($form, &$form_state, $form_id) { 1010 $form[$form_id] = array( 1011 '#type' => 'textfield', 1012 '#title' => t('Search'), 1013 '#title_display' => 'invisible', 1014 '#size' => 15, 1015 '#default_value' => '', 1016 '#attributes' => array('title' => t('Enter the terms you wish to search for.')), 1017 ); 1018 $form['actions'] = array('#type' => 'actions'); 1019 $form['actions']['submit'] = array('#type' => 'submit', '#value' => t('Search')); 1020 $form['#submit'][] = 'search_box_form_submit'; 1021 1022 return $form; 1023 } 1024 1025 /** 1026 * Process a block search form submission. 1027 */ 1028 function search_box_form_submit($form, &$form_state) { 1029 // The search form relies on control of the redirect destination for its 1030 // functionality, so we override any static destination set in the request, 1031 // for example by drupal_access_denied() or drupal_not_found() 1032 // (see http://drupal.org/node/292565). 1033 if (isset($_GET['destination'])) { 1034 unset($_GET['destination']); 1035 } 1036 1037 // Check to see if the form was submitted empty. 1038 // If it is empty, display an error message. 1039 // (This method is used instead of setting #required to TRUE for this field 1040 // because that results in a confusing error message. It would say a plain 1041 // "field is required" because the search keywords field has no title. 1042 // The error message would also complain about a missing #title field.) 1043 if ($form_state['values']['search_block_form'] == '') { 1044 form_set_error('keys', t('Please enter some keywords.')); 1045 } 1046 1047 $form_id = $form['form_id']['#value']; 1048 $info = search_get_default_module_info(); 1049 if ($info) { 1050 $form_state['redirect'] = 'search/' . $info['path'] . '/' . trim($form_state['values'][$form_id]); 1051 } 1052 else { 1053 form_set_error(NULL, t('Search is currently disabled.'), 'error'); 1054 } 1055 } 1056 1057 /** 1058 * Process variables for search-block-form.tpl.php. 1059 * 1060 * The $variables array contains the following arguments: 1061 * - $form 1062 * 1063 * @see search-block-form.tpl.php 1064 */ 1065 function template_preprocess_search_block_form(&$variables) { 1066 $variables['search'] = array(); 1067 $hidden = array(); 1068 // Provide variables named after form keys so themers can print each element independently. 1069 foreach (element_children($variables['form']) as $key) { 1070 $type = $variables['form'][$key]['#type']; 1071 if ($type == 'hidden' || $type == 'token') { 1072 $hidden[] = drupal_render($variables['form'][$key]); 1073 } 1074 else { 1075 $variables['search'][$key] = drupal_render($variables['form'][$key]); 1076 } 1077 } 1078 // Hidden form elements have no value to themers. No need for separation. 1079 $variables['search']['hidden'] = implode($hidden); 1080 // Collect all form elements to make it easier to print the whole form. 1081 $variables['search_form'] = implode($variables['search']); 1082 } 1083 1084 /** 1085 * Performs a search by calling hook_search_execute(). 1086 * 1087 * @param $keys 1088 * Keyword query to search on. 1089 * @param $module 1090 * Search module to search. 1091 * @param $conditions 1092 * Optional array of additional search conditions. 1093 * 1094 * @return 1095 * Renderable array of search results. No return value if $keys are not 1096 * supplied or if the given search module is not active. 1097 */ 1098 function search_data($keys, $module, $conditions = NULL) { 1099 if (module_hook($module, 'search_execute')) { 1100 $results = module_invoke($module, 'search_execute', $keys, $conditions); 1101 if (module_hook($module, 'search_page')) { 1102 return module_invoke($module, 'search_page', $results); 1103 } 1104 else { 1105 return array( 1106 '#theme' => 'search_results', 1107 '#results' => $results, 1108 '#module' => $module, 1109 ); 1110 } 1111 } 1112 } 1113 1114 /** 1115 * Returns snippets from a piece of text, with certain keywords highlighted. 1116 * Used for formatting search results. 1117 * 1118 * @param $keys 1119 * A string containing a search query. 1120 * 1121 * @param $text 1122 * The text to extract fragments from. 1123 * 1124 * @return 1125 * A string containing HTML for the excerpt. 1126 */ 1127 function search_excerpt($keys, $text) { 1128 // We highlight around non-indexable or CJK characters. 1129 $boundary = '(?:(?<=[' . PREG_CLASS_UNICODE_WORD_BOUNDARY . PREG_CLASS_CJK . '])|(?=[' . PREG_CLASS_UNICODE_WORD_BOUNDARY . PREG_CLASS_CJK . ']))'; 1130 1131 // Extract positive keywords and phrases 1132 preg_match_all('/ ("([^"]+)"|(?!OR)([^" ]+))/', ' ' . $keys, $matches); 1133 $keys = array_merge($matches[2], $matches[3]); 1134 1135 // Prepare text by stripping HTML tags and decoding HTML entities. 1136 $text = strip_tags(str_replace(array('<', '>'), array(' <', '> '), $text)); 1137 $text = decode_entities($text); 1138 1139 // Slash-escape quotes in the search keyword string. 1140 array_walk($keys, '_search_excerpt_replace'); 1141 $workkeys = $keys; 1142 1143 // Extract fragments around keywords. 1144 // First we collect ranges of text around each keyword, starting/ending 1145 // at spaces, trying to get to 256 characters. 1146 // If the sum of all fragments is too short, we look for second occurrences. 1147 $ranges = array(); 1148 $included = array(); 1149 $foundkeys = array(); 1150 $length = 0; 1151 while ($length < 256 && count($workkeys)) { 1152 foreach ($workkeys as $k => $key) { 1153 if (strlen($key) == 0) { 1154 unset($workkeys[$k]); 1155 unset($keys[$k]); 1156 continue; 1157 } 1158 if ($length >= 256) { 1159 break; 1160 } 1161 // Remember occurrence of key so we can skip over it if more occurrences 1162 // are desired. 1163 if (!isset($included[$key])) { 1164 $included[$key] = 0; 1165 } 1166 // Locate a keyword (position $p, always >0 because $text starts with a 1167 // space). First try bare keyword, but if that doesn't work, try to find a 1168 // derived form from search_simplify(). 1169 $p = 0; 1170 if (preg_match('/' . $boundary . $key . $boundary . '/iu', $text, $match, PREG_OFFSET_CAPTURE, $included[$key])) { 1171 $p = $match[0][1]; 1172 } 1173 else { 1174 $info = search_simplify_excerpt_match($key, $text, $included[$key], $boundary); 1175 if ($info['where']) { 1176 $p = $info['where']; 1177 if ($info['keyword']) { 1178 $foundkeys[] = $info['keyword']; 1179 } 1180 } 1181 } 1182 // Now locate a space in front (position $q) and behind it (position $s), 1183 // leaving about 60 characters extra before and after for context. 1184 // Note that a space was added to the front and end of $text above. 1185 if ($p) { 1186 if (($q = strpos(' ' . $text, ' ', max(0, $p - 61))) !== FALSE) { 1187 $end = substr($text . ' ', $p, 80); 1188 if (($s = strrpos($end, ' ')) !== FALSE) { 1189 // Account for the added spaces. 1190 $q = max($q - 1, 0); 1191 $s = min($s, strlen($end) - 1); 1192 $ranges[$q] = $p + $s; 1193 $length += $p + $s - $q; 1194 $included[$key] = $p + 1; 1195 } 1196 else { 1197 unset($workkeys[$k]); 1198 } 1199 } 1200 else { 1201 unset($workkeys[$k]); 1202 } 1203 } 1204 else { 1205 unset($workkeys[$k]); 1206 } 1207 } 1208 } 1209 1210 if (count($ranges) == 0) { 1211 // We didn't find any keyword matches, so just return the first part of the 1212 // text. We also need to re-encode any HTML special characters that we 1213 // entity-decoded above. 1214 return check_plain(truncate_utf8($text, 256, TRUE, TRUE)); 1215 } 1216 1217 // Sort the text ranges by starting position. 1218 ksort($ranges); 1219 1220 // Now we collapse overlapping text ranges into one. The sorting makes it O(n). 1221 $newranges = array(); 1222 foreach ($ranges as $from2 => $to2) { 1223 if (!isset($from1)) { 1224 $from1 = $from2; 1225 $to1 = $to2; 1226 continue; 1227 } 1228 if ($from2 <= $to1) { 1229 $to1 = max($to1, $to2); 1230 } 1231 else { 1232 $newranges[$from1] = $to1; 1233 $from1 = $from2; 1234 $to1 = $to2; 1235 } 1236 } 1237 $newranges[$from1] = $to1; 1238 1239 // Fetch text 1240 $out = array(); 1241 foreach ($newranges as $from => $to) { 1242 $out[] = substr($text, $from, $to - $from); 1243 } 1244 1245 // Let translators have the ... separator text as one chunk. 1246 $dots = explode('!excerpt', t('... !excerpt ... !excerpt ...')); 1247 1248 $text = (isset($newranges[0]) ? '' : $dots[0]) . implode($dots[1], $out) . $dots[2]; 1249 $text = check_plain($text); 1250 1251 // Slash-escape quotes in keys found in a derived form and merge with original keys. 1252 array_walk($foundkeys, '_search_excerpt_replace'); 1253 $keys = array_merge($keys, $foundkeys); 1254 1255 // Highlight keywords. Must be done at once to prevent conflicts ('strong' and '<strong>'). 1256 $text = preg_replace('/' . $boundary . '(' . implode('|', $keys) . ')' . $boundary . '/iu', '<strong>\0</strong>', $text); 1257 return $text; 1258 } 1259 1260 /** 1261 * @} End of "defgroup search". 1262 */ 1263 1264 /** 1265 * Helper function for array_walk() in search_excerpt(). 1266 */ 1267 function _search_excerpt_replace(&$text) { 1268 $text = preg_quote($text, '/'); 1269 } 1270 1271 /** 1272 * Find words in the original text that matched via search_simplify(). 1273 * 1274 * This is called in search_excerpt() if an exact match is not found in the 1275 * text, so that we can find the derived form that matches. 1276 * 1277 * @param $key 1278 * The keyword to find. 1279 * @param $text 1280 * The text to search for the keyword. 1281 * @param $offset 1282 * Offset position in $text to start searching at. 1283 * @param $boundary 1284 * Text to include in a regular expression that will match a word boundary. 1285 * 1286 * @return 1287 * FALSE if no match is found. If a match is found, return an associative 1288 * array with element 'where' giving the position of the match, and element 1289 * 'keyword' giving the actual word found in the text at that position. 1290 */ 1291 function search_simplify_excerpt_match($key, $text, $offset, $boundary) { 1292 $pos = NULL; 1293 $simplified_key = search_simplify($key); 1294 $simplified_text = search_simplify($text); 1295 1296 // Return immediately if simplified key or text are empty. 1297 if (!$simplified_key || !$simplified_text) { 1298 return FALSE; 1299 } 1300 1301 // Check if we have a match after simplification in the text. 1302 if (!preg_match('/' . $boundary . $simplified_key . $boundary . '/iu', $simplified_text, $match, PREG_OFFSET_CAPTURE, $offset)) { 1303 return FALSE; 1304 } 1305 1306 // If we get here, we have a match. Now find the exact location of the match 1307 // and the original text that matched. Start by splitting up the text by all 1308 // potential starting points of the matching text and iterating through them. 1309 $split = array_filter(preg_split('/' . $boundary . '/iu', $text, -1, PREG_SPLIT_OFFSET_CAPTURE), '_search_excerpt_match_filter'); 1310 foreach ($split as $value) { 1311 // Skip starting points before the offset. 1312 if ($value[1] < $offset) { 1313 continue; 1314 } 1315 1316 // Check a window of 80 characters after the starting point for a match, 1317 // based on the size of the excerpt window. 1318 $window = substr($text, $value[1], 80); 1319 $simplified_window = search_simplify($window); 1320 if (strpos($simplified_window, $simplified_key) === 0) { 1321 // We have a match in this window. Store the position of the match. 1322 $pos = $value[1]; 1323 // Iterate through the text in the window until we find the full original 1324 // matching text. 1325 $length = strlen($window); 1326 for ($i = 1; $i <= $length; $i++) { 1327 $keyfound = substr($text, $value[1], $i); 1328 if ($simplified_key == search_simplify($keyfound)) { 1329 break; 1330 } 1331 } 1332 break; 1333 } 1334 } 1335 1336 return $pos ? array('where' => $pos, 'keyword' => $keyfound) : FALSE; 1337 } 1338 1339 /** 1340 * Helper function for array_filter() in search_search_excerpt_match(). 1341 */ 1342 function _search_excerpt_match_filter($var) { 1343 return strlen(trim($var[0])); 1344 } 1345 1346 /** 1347 * Implements hook_forms(). 1348 */ 1349 function search_forms() { 1350 $forms['search_block_form']= array( 1351 'callback' => 'search_box', 1352 'callback arguments' => array('search_block_form'), 1353 ); 1354 return $forms; 1355 } 1356
title
Description
Body
title
Description
Body
title
Description
Body
title
Body
title