Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
2.67% |
5 / 187 |
|
10.00% |
1 / 10 |
CRAP | |
0.00% |
0 / 1 |
| base | |
2.67% |
5 / 187 |
|
10.00% |
1 / 10 |
2947.12 | |
0.00% |
0 / 1 |
| __construct | |
100.00% |
5 / 5 |
|
100.00% |
1 / 1 |
1 | |||
| obtain_ids | |
0.00% |
0 / 26 |
|
0.00% |
0 / 1 |
132 | |||
| save_ids | |
0.00% |
0 / 51 |
|
0.00% |
0 / 1 |
182 | |||
| destroy_cache | |
0.00% |
0 / 26 |
|
0.00% |
0 / 1 |
90 | |||
| create_index | |
0.00% |
0 / 29 |
|
0.00% |
0 / 1 |
72 | |||
| delete_index | |
0.00% |
0 / 26 |
|
0.00% |
0 / 1 |
56 | |||
| forum_ids_with_indexing_enabled | |
0.00% |
0 / 9 |
|
0.00% |
0 / 1 |
12 | |||
| get_posts_batch_after | |
0.00% |
0 / 8 |
|
0.00% |
0 / 1 |
6 | |||
| get_max_post_id | |
0.00% |
0 / 6 |
|
0.00% |
0 / 1 |
2 | |||
| get_type | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| 1 | <?php |
| 2 | /** |
| 3 | * |
| 4 | * This file is part of the phpBB Forum Software package. |
| 5 | * |
| 6 | * @copyright (c) phpBB Limited <https://www.phpbb.com> |
| 7 | * @license GNU General Public License, version 2 (GPL-2.0) |
| 8 | * |
| 9 | * For full copyright and license information, please see |
| 10 | * the docs/CREDITS.txt file. |
| 11 | * |
| 12 | */ |
| 13 | |
| 14 | namespace phpbb\search\backend; |
| 15 | |
| 16 | use phpbb\cache\service; |
| 17 | use phpbb\config\config; |
| 18 | use phpbb\db\driver\driver_interface; |
| 19 | use phpbb\user; |
| 20 | |
| 21 | /** |
| 22 | * optional base class for search plugins providing simple caching based on ACM |
| 23 | * and functions to retrieve ignore_words and synonyms |
| 24 | */ |
| 25 | abstract class base implements search_backend_interface |
| 26 | { |
| 27 | public const SEARCH_RESULT_NOT_IN_CACHE = 0; |
| 28 | public const SEARCH_RESULT_IN_CACHE = 1; |
| 29 | public const SEARCH_RESULT_INCOMPLETE = 2; |
| 30 | |
| 31 | // Batch size for create_index and delete_index |
| 32 | private const BATCH_SIZE = 100; |
| 33 | |
| 34 | /** |
| 35 | * @var service |
| 36 | */ |
| 37 | protected $cache; |
| 38 | |
| 39 | /** |
| 40 | * @var config |
| 41 | */ |
| 42 | protected $config; |
| 43 | |
| 44 | /** |
| 45 | * @var driver_interface |
| 46 | */ |
| 47 | protected $db; |
| 48 | |
| 49 | /** |
| 50 | * @var user |
| 51 | */ |
| 52 | protected $user; |
| 53 | |
| 54 | /** |
| 55 | * @var string |
| 56 | */ |
| 57 | protected $search_results_table; |
| 58 | |
| 59 | /** |
| 60 | * Constructor. |
| 61 | * |
| 62 | * @param service $cache |
| 63 | * @param config $config |
| 64 | * @param driver_interface $db |
| 65 | * @param user $user |
| 66 | * @param string $search_results_table |
| 67 | */ |
| 68 | public function __construct(service $cache, config $config, driver_interface $db, user $user, string $search_results_table) |
| 69 | { |
| 70 | $this->cache = $cache; |
| 71 | $this->config = $config; |
| 72 | $this->db = $db; |
| 73 | $this->user = $user; |
| 74 | $this->search_results_table = $search_results_table; |
| 75 | } |
| 76 | |
| 77 | /** |
| 78 | * Retrieves cached search results |
| 79 | * |
| 80 | * @param string $search_key an md5 string generated from all the passed search options to identify the results |
| 81 | * @param int &$result_count will contain the number of all results for the search (not only for the current page) |
| 82 | * @param array &$id_ary is filled with the ids belonging to the requested page that are stored in the cache |
| 83 | * @param int &$start indicates the first index of the page |
| 84 | * @param int $per_page number of ids each page is supposed to contain |
| 85 | * @param string $sort_dir is either a or d representing ASC and DESC |
| 86 | * |
| 87 | * @return int self::SEARCH_RESULT_NOT_IN_CACHE or self::SEARCH_RESULT_IN_CACHE or self::SEARCH_RESULT_INCOMPLETE |
| 88 | */ |
| 89 | protected function obtain_ids(string $search_key, int &$result_count, array &$id_ary, int &$start, int $per_page, string $sort_dir): int |
| 90 | { |
| 91 | if (!($stored_ids = $this->cache->get('_search_results_' . $search_key))) |
| 92 | { |
| 93 | // no search results cached for this search_key |
| 94 | return self::SEARCH_RESULT_NOT_IN_CACHE; |
| 95 | } |
| 96 | else |
| 97 | { |
| 98 | $result_count = $stored_ids[-1]; |
| 99 | $reverse_ids = $stored_ids[-2] != $sort_dir; |
| 100 | $complete = true; |
| 101 | |
| 102 | // Change start parameter in case out of bounds |
| 103 | if ($result_count) |
| 104 | { |
| 105 | if ($start < 0) |
| 106 | { |
| 107 | $start = 0; |
| 108 | } |
| 109 | else if ($start >= $result_count) |
| 110 | { |
| 111 | $start = floor(($result_count - 1) / $per_page) * $per_page; |
| 112 | } |
| 113 | } |
| 114 | |
| 115 | // If the sort direction differs from the direction in the cache, then recalculate array keys |
| 116 | if ($reverse_ids) |
| 117 | { |
| 118 | $keys = array_keys($stored_ids); |
| 119 | array_walk($keys, function (&$value, $key) use ($result_count) |
| 120 | { |
| 121 | $value = ($value >= 0) ? $result_count - $value - 1 : $value; |
| 122 | } |
| 123 | ); |
| 124 | $stored_ids = array_combine($keys, $stored_ids); |
| 125 | } |
| 126 | |
| 127 | for ($i = $start, $n = $start + $per_page; ($i < $n) && ($i < $result_count); $i++) |
| 128 | { |
| 129 | if (!isset($stored_ids[$i])) |
| 130 | { |
| 131 | $complete = false; |
| 132 | } |
| 133 | else |
| 134 | { |
| 135 | $id_ary[] = $stored_ids[$i]; |
| 136 | } |
| 137 | } |
| 138 | unset($stored_ids); |
| 139 | |
| 140 | if (!$complete) |
| 141 | { |
| 142 | return self::SEARCH_RESULT_INCOMPLETE; |
| 143 | } |
| 144 | return self::SEARCH_RESULT_IN_CACHE; |
| 145 | } |
| 146 | } |
| 147 | |
| 148 | /** |
| 149 | * Caches post/topic ids |
| 150 | * |
| 151 | * @param string $search_key an md5 string generated from all the passed search options to identify the results |
| 152 | * @param string $keywords contains the keywords as entered by the user |
| 153 | * @param array $author_ary an array of author ids, if the author should be ignored during the search the array is empty |
| 154 | * @param int $result_count contains the number of all results for the search (not only for the current page) |
| 155 | * @param array &$id_ary contains a list of post or topic ids that shall be cached, the first element |
| 156 | * must have the absolute index $start in the result set. |
| 157 | * @param int $start indicates the first index of the page |
| 158 | * @param string $sort_dir is either a or d representing ASC and DESC |
| 159 | * |
| 160 | * @return void |
| 161 | */ |
| 162 | protected function save_ids(string $search_key, string $keywords, array $author_ary, int $result_count, array &$id_ary, int $start, string $sort_dir): void |
| 163 | { |
| 164 | global $user; |
| 165 | |
| 166 | $length = min(count($id_ary), $this->config['search_block_size']); |
| 167 | |
| 168 | // nothing to cache so exit |
| 169 | if (!$length) |
| 170 | { |
| 171 | return; |
| 172 | } |
| 173 | |
| 174 | $store_ids = array_slice($id_ary, 0, $length); |
| 175 | $id_range = range($start, $start + $length - 1); |
| 176 | $store_ids = array_combine($id_range, $store_ids); |
| 177 | |
| 178 | // create a new resultset if there is none for this search_key yet |
| 179 | // or add the ids to the existing resultset |
| 180 | if (!($store = $this->cache->get('_search_results_' . $search_key))) |
| 181 | { |
| 182 | // add the current keywords to the recent searches in the cache which are listed on the search page |
| 183 | if (!empty($keywords) || count($author_ary)) |
| 184 | { |
| 185 | $sql = 'SELECT search_time |
| 186 | FROM ' . $this->search_results_table . ' |
| 187 | WHERE search_key = \'' . $this->db->sql_escape($search_key) . '\''; |
| 188 | $result = $this->db->sql_query($sql); |
| 189 | |
| 190 | if (!$this->db->sql_fetchrow($result)) |
| 191 | { |
| 192 | $sql_ary = array( |
| 193 | 'search_key' => $search_key, |
| 194 | 'search_time' => time(), |
| 195 | 'search_keywords' => $keywords, |
| 196 | 'search_authors' => ' ' . implode(' ', $author_ary) . ' ' |
| 197 | ); |
| 198 | |
| 199 | $sql = 'INSERT INTO ' . $this->search_results_table . ' ' . $this->db->sql_build_array('INSERT', $sql_ary); |
| 200 | $this->db->sql_query($sql); |
| 201 | } |
| 202 | $this->db->sql_freeresult($result); |
| 203 | } |
| 204 | |
| 205 | $sql = 'UPDATE ' . USERS_TABLE . ' |
| 206 | SET user_last_search = ' . time() . ' |
| 207 | WHERE user_id = ' . $user->data['user_id']; |
| 208 | $this->db->sql_query($sql); |
| 209 | |
| 210 | $store = array(-1 => $result_count, -2 => $sort_dir); |
| 211 | } |
| 212 | else |
| 213 | { |
| 214 | // we use one set of results for both sort directions so we have to calculate the indizes |
| 215 | // for the reversed array |
| 216 | if ($store[-2] != $sort_dir) |
| 217 | { |
| 218 | $keys = array_keys($store_ids); |
| 219 | array_walk($keys, function (&$value, $key) use ($store) { |
| 220 | $value = $store[-1] - $value - 1; |
| 221 | }); |
| 222 | $store_ids = array_combine($keys, $store_ids); |
| 223 | } |
| 224 | } |
| 225 | |
| 226 | // append the ids |
| 227 | if (is_array($store_ids)) |
| 228 | { |
| 229 | $store += $store_ids; |
| 230 | ksort($store); |
| 231 | |
| 232 | // if the cache is too big |
| 233 | if (count($store) - 2 > 20 * $this->config['search_block_size']) |
| 234 | { |
| 235 | // remove everything in front of two blocks in front of the current start index |
| 236 | for ($i = 0, $n = $id_range[0] - 2 * $this->config['search_block_size']; $i < $n; $i++) |
| 237 | { |
| 238 | if (isset($store[$i])) |
| 239 | { |
| 240 | unset($store[$i]); |
| 241 | } |
| 242 | } |
| 243 | |
| 244 | // remove everything after two blocks after the current stop index |
| 245 | end($id_range); |
| 246 | for ($i = $store[-1] - 1, $n = current($id_range) + 2 * $this->config['search_block_size']; $i > $n; $i--) |
| 247 | { |
| 248 | if (isset($store[$i])) |
| 249 | { |
| 250 | unset($store[$i]); |
| 251 | } |
| 252 | } |
| 253 | } |
| 254 | $this->cache->put('_search_results_' . $search_key, $store, $this->config['search_store_results']); |
| 255 | |
| 256 | $sql = 'UPDATE ' . $this->search_results_table . ' |
| 257 | SET search_time = ' . time() . ' |
| 258 | WHERE search_key = \'' . $this->db->sql_escape($search_key) . '\''; |
| 259 | $this->db->sql_query($sql); |
| 260 | } |
| 261 | |
| 262 | unset($store, $store_ids, $id_range); |
| 263 | } |
| 264 | |
| 265 | /** |
| 266 | * Removes old entries from the search results table and removes searches with keywords that contain a word in $words. |
| 267 | * |
| 268 | * @param array $words |
| 269 | * @param array|bool $authors |
| 270 | */ |
| 271 | protected function destroy_cache(array $words, $authors = false): void |
| 272 | { |
| 273 | // clear all searches that searched for the specified words |
| 274 | if (count($words)) |
| 275 | { |
| 276 | $sql_where = ''; |
| 277 | foreach ($words as $word) |
| 278 | { |
| 279 | $sql_where .= " OR search_keywords " . $this->db->sql_like_expression($this->db->get_any_char() . $word . $this->db->get_any_char()); |
| 280 | } |
| 281 | |
| 282 | $sql = 'SELECT search_key |
| 283 | FROM ' . $this->search_results_table . " |
| 284 | WHERE search_keywords LIKE '%*%' $sql_where"; |
| 285 | $result = $this->db->sql_query($sql); |
| 286 | |
| 287 | while ($row = $this->db->sql_fetchrow($result)) |
| 288 | { |
| 289 | $this->cache->destroy('_search_results_' . $row['search_key']); |
| 290 | } |
| 291 | $this->db->sql_freeresult($result); |
| 292 | } |
| 293 | |
| 294 | // clear all searches that searched for the specified authors |
| 295 | if (is_array($authors) && count($authors)) |
| 296 | { |
| 297 | $sql_where = ''; |
| 298 | foreach ($authors as $author) |
| 299 | { |
| 300 | $sql_where .= (($sql_where) ? ' OR ' : '') . 'search_authors ' . $this->db->sql_like_expression($this->db->get_any_char() . ' ' . (int) $author . ' ' . $this->db->get_any_char()); |
| 301 | } |
| 302 | |
| 303 | $sql = 'SELECT search_key |
| 304 | FROM ' . $this->search_results_table . " |
| 305 | WHERE $sql_where"; |
| 306 | $result = $this->db->sql_query($sql); |
| 307 | |
| 308 | while ($row = $this->db->sql_fetchrow($result)) |
| 309 | { |
| 310 | $this->cache->destroy('_search_results_' . $row['search_key']); |
| 311 | } |
| 312 | $this->db->sql_freeresult($result); |
| 313 | } |
| 314 | |
| 315 | $sql = 'DELETE |
| 316 | FROM ' . $this->search_results_table . ' |
| 317 | WHERE search_time < ' . (time() - (int) $this->config['search_store_results']); |
| 318 | $this->db->sql_query($sql); |
| 319 | } |
| 320 | |
| 321 | /** |
| 322 | * {@inheritdoc} |
| 323 | */ |
| 324 | public function create_index(int &$post_counter = 0): array|null |
| 325 | { |
| 326 | $max_post_id = $this->get_max_post_id(); |
| 327 | $forums_indexing_enabled = $this->forum_ids_with_indexing_enabled(); |
| 328 | |
| 329 | $starttime = microtime(true); |
| 330 | $row_count = 0; |
| 331 | |
| 332 | while (still_on_time() && $post_counter < $max_post_id) |
| 333 | { |
| 334 | $rows = $this->get_posts_batch_after($post_counter); |
| 335 | |
| 336 | if ($this->db->sql_buffer_nested_transactions()) |
| 337 | { |
| 338 | $rows = iterator_to_array($rows); |
| 339 | } |
| 340 | |
| 341 | foreach ($rows as $row) |
| 342 | { |
| 343 | // Indexing enabled for this forum |
| 344 | if (in_array($row['forum_id'], $forums_indexing_enabled, true)) |
| 345 | { |
| 346 | $this->index('post', (int) $row['post_id'], $row['post_text'], $row['post_subject'], (int) $row['poster_id'], (int) $row['forum_id']); |
| 347 | } |
| 348 | $row_count++; |
| 349 | $post_counter = (int) $row['post_id']; |
| 350 | } |
| 351 | |
| 352 | // With cli process only one batch each time to be able to track progress |
| 353 | if (PHP_SAPI === 'cli') |
| 354 | { |
| 355 | break; |
| 356 | } |
| 357 | } |
| 358 | |
| 359 | // pretend the number of posts was as big as the number of ids we indexed so far |
| 360 | // just an estimation as it includes deleted posts |
| 361 | $num_posts = $this->config['num_posts']; |
| 362 | $this->config['num_posts'] = min($this->config['num_posts'], $post_counter); |
| 363 | $this->tidy(); |
| 364 | $this->config['num_posts'] = $num_posts; |
| 365 | |
| 366 | if ($post_counter < $max_post_id) // If there are still post to index |
| 367 | { |
| 368 | $totaltime = microtime(true) - $starttime; |
| 369 | $rows_per_second = $row_count / $totaltime; |
| 370 | |
| 371 | return [ |
| 372 | 'row_count' => $row_count, |
| 373 | 'post_counter' => $post_counter, |
| 374 | 'max_post_id' => $max_post_id, |
| 375 | 'rows_per_second' => $rows_per_second, |
| 376 | ]; |
| 377 | } |
| 378 | |
| 379 | return null; |
| 380 | } |
| 381 | |
| 382 | /** |
| 383 | * {@inheritdoc} |
| 384 | */ |
| 385 | public function delete_index(int|null &$post_counter = null): array|null |
| 386 | { |
| 387 | $max_post_id = $this->get_max_post_id(); |
| 388 | |
| 389 | $starttime = microtime(true); |
| 390 | $row_count = 0; |
| 391 | |
| 392 | while (still_on_time() && $post_counter < $max_post_id) |
| 393 | { |
| 394 | $rows = $this->get_posts_batch_after($post_counter); |
| 395 | $ids = $posters = $forum_ids = array(); |
| 396 | foreach ($rows as $row) |
| 397 | { |
| 398 | $ids[] = $row['post_id']; |
| 399 | $posters[] = $row['poster_id']; |
| 400 | $forum_ids[] = $row['forum_id']; |
| 401 | } |
| 402 | $row_count += count($ids); |
| 403 | |
| 404 | if (count($ids)) |
| 405 | { |
| 406 | $this->index_remove($ids, $posters, $forum_ids); |
| 407 | $post_counter = $ids[count($ids) - 1]; |
| 408 | } |
| 409 | |
| 410 | // With cli process only one batch each time to be able to track progress |
| 411 | if (PHP_SAPI === 'cli') |
| 412 | { |
| 413 | break; |
| 414 | } |
| 415 | } |
| 416 | |
| 417 | if ($post_counter < $max_post_id) // If there are still post delete from index |
| 418 | { |
| 419 | $totaltime = microtime(true) - $starttime; |
| 420 | $rows_per_second = $row_count / $totaltime; |
| 421 | |
| 422 | return [ |
| 423 | 'row_count' => $row_count, |
| 424 | 'post_counter' => $post_counter, |
| 425 | 'max_post_id' => $max_post_id, |
| 426 | 'rows_per_second' => $rows_per_second, |
| 427 | ]; |
| 428 | } |
| 429 | |
| 430 | return null; |
| 431 | } |
| 432 | |
| 433 | /** |
| 434 | * Return the ids of the forums that have indexing enabled |
| 435 | * |
| 436 | * @return array |
| 437 | */ |
| 438 | protected function forum_ids_with_indexing_enabled(): array |
| 439 | { |
| 440 | $forums = []; |
| 441 | |
| 442 | $sql = 'SELECT forum_id, enable_indexing |
| 443 | FROM ' . FORUMS_TABLE; |
| 444 | $result = $this->db->sql_query($sql, 3600); |
| 445 | |
| 446 | while ($row = $this->db->sql_fetchrow($result)) |
| 447 | { |
| 448 | if ((bool) $row['enable_indexing']) |
| 449 | { |
| 450 | $forums[] = $row['forum_id']; |
| 451 | } |
| 452 | } |
| 453 | $this->db->sql_freeresult($result); |
| 454 | |
| 455 | return $forums; |
| 456 | } |
| 457 | |
| 458 | /** |
| 459 | * Get batch of posts after id |
| 460 | * |
| 461 | * @param int $post_id |
| 462 | * @return \Generator |
| 463 | */ |
| 464 | protected function get_posts_batch_after(int $post_id): \Generator |
| 465 | { |
| 466 | $sql = 'SELECT post_id, post_subject, post_text, poster_id, forum_id |
| 467 | FROM ' . POSTS_TABLE . ' |
| 468 | WHERE post_id > ' . (int) $post_id . ' |
| 469 | ORDER BY post_id ASC'; |
| 470 | $result = $this->db->sql_query_limit($sql, self::BATCH_SIZE); |
| 471 | |
| 472 | while ($row = $this->db->sql_fetchrow($result)) |
| 473 | { |
| 474 | yield $row; |
| 475 | } |
| 476 | |
| 477 | $this->db->sql_freeresult($result); |
| 478 | } |
| 479 | |
| 480 | /** |
| 481 | * Get post with higher id |
| 482 | */ |
| 483 | protected function get_max_post_id(): int |
| 484 | { |
| 485 | $sql = 'SELECT MAX(post_id) as max_post_id |
| 486 | FROM '. POSTS_TABLE; |
| 487 | $result = $this->db->sql_query($sql); |
| 488 | $max_post_id = (int) $this->db->sql_fetchfield('max_post_id'); |
| 489 | $this->db->sql_freeresult($result); |
| 490 | |
| 491 | return $max_post_id; |
| 492 | } |
| 493 | |
| 494 | /** |
| 495 | * {@inheritdoc} |
| 496 | */ |
| 497 | public function get_type(): string |
| 498 | { |
| 499 | return static::class; |
| 500 | } |
| 501 | } |