mysql.php 18.9 KB
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558
<?php
/**
 * @package     Joomla.Administrator
 * @subpackage  com_finder
 *
 * @copyright   Copyright (C) 2005 - 2019 Open Source Matters, Inc. All rights reserved.
 * @license     GNU General Public License version 2 or later; see LICENSE.txt
 */

defined('_JEXEC') or die;

jimport('joomla.filesystem.file');

/**
 * Indexer class supporting MySQL(i) for the Finder indexer package.
 *
 * The indexer class provides the core functionality of the Finder
 * search engine. It is responsible for adding and updating the
 * content links table; extracting and scoring tokens; and maintaining
 * all referential information for the content.
 *
 * Note: All exceptions thrown from within this class should be caught
 * by the controller.
 *
 * @since  3.0
 */
class FinderIndexerDriverMysql extends FinderIndexer
{
	/**
	 * Method to index a content item.
	 *
	 * @param   FinderIndexerResult  $item    The content item to index.
	 * @param   string               $format  The format of the content. [optional]
	 *
	 * @return  integer  The ID of the record in the links table.
	 *
	 * @since   3.0
	 * @throws  Exception on database error.
	 */
	public function index($item, $format = 'html')
	{
		// Mark beforeIndexing in the profiler.
		static::$profiler ? static::$profiler->mark('beforeIndexing') : null;
		$db = $this->db;
		$nd = $db->getNullDate();

		// Check if the item is in the database.
		$query = $db->getQuery(true)
			->select($db->quoteName('link_id') . ', ' . $db->quoteName('md5sum'))
			->from($db->quoteName('#__finder_links'))
			->where($db->quoteName('url') . ' = ' . $db->quote($item->url));

		// Load the item  from the database.
		$db->setQuery($query);
		$link = $db->loadObject();

		// Get the indexer state.
		$state = static::getState();

		// Get the signatures of the item.
		$curSig = static::getSignature($item);
		$oldSig = isset($link->md5sum) ? $link->md5sum : null;

		// Get the other item information.
		$linkId = empty($link->link_id) ? null : $link->link_id;
		$isNew = empty($link->link_id) ? true : false;

		// Check the signatures. If they match, the item is up to date.
		if (!$isNew && $curSig == $oldSig)
		{
			return $linkId;
		}

		/*
		 * If the link already exists, flush all the term maps for the item.
		 * Maps are stored in 16 tables so we need to iterate through and flush
		 * each table one at a time.
		 */
		if (!$isNew)
		{
			for ($i = 0; $i <= 15; $i++)
			{
				// Flush the maps for the link.
				$query->clear()
					->delete($db->quoteName('#__finder_links_terms' . dechex($i)))
					->where($db->quoteName('link_id') . ' = ' . (int) $linkId);
				$db->setQuery($query);
				$db->execute();
			}

			// Remove the taxonomy maps.
			FinderIndexerTaxonomy::removeMaps($linkId);
		}

		// Mark afterUnmapping in the profiler.
		static::$profiler ? static::$profiler->mark('afterUnmapping') : null;

		// Perform cleanup on the item data.
		$item->publish_start_date = (int) $item->publish_start_date != 0 ? $item->publish_start_date : $nd;
		$item->publish_end_date = (int) $item->publish_end_date != 0 ? $item->publish_end_date : $nd;
		$item->start_date = (int) $item->start_date != 0 ? $item->start_date : $nd;
		$item->end_date = (int) $item->end_date != 0 ? $item->end_date : $nd;

		// Prepare the item description.
		$item->description = FinderIndexerHelper::parse($item->summary);

		/*
		 * Now, we need to enter the item into the links table. If the item
		 * already exists in the database, we need to use an UPDATE query.
		 * Otherwise, we need to use an INSERT to get the link id back.
		 */

		if ($isNew)
		{
			$columnsArray = array(
				$db->quoteName('url'), $db->quoteName('route'), $db->quoteName('title'), $db->quoteName('description'),
				$db->quoteName('indexdate'), $db->quoteName('published'), $db->quoteName('state'), $db->quoteName('access'),
				$db->quoteName('language'), $db->quoteName('type_id'), $db->quoteName('object'), $db->quoteName('publish_start_date'),
				$db->quoteName('publish_end_date'), $db->quoteName('start_date'), $db->quoteName('end_date'), $db->quoteName('list_price'),
				$db->quoteName('sale_price')
			);

			// Insert the link.
			$query->clear()
				->insert($db->quoteName('#__finder_links'))
				->columns($columnsArray)
				->values(
					$db->quote($item->url) . ', '
					. $db->quote($item->route) . ', '
					. $db->quote($item->title) . ', '
					. $db->quote($item->description) . ', '
					. $query->currentTimestamp() . ', '
					. '1, '
					. (int) $item->state . ', '
					. (int) $item->access . ', '
					. $db->quote($item->language) . ', '
					. (int) $item->type_id . ', '
					. $db->quote(serialize($item)) . ', '
					. $db->quote($item->publish_start_date) . ', '
					. $db->quote($item->publish_end_date) . ', '
					. $db->quote($item->start_date) . ', '
					. $db->quote($item->end_date) . ', '
					. (double) ($item->list_price ?: 0) . ', '
					. (double) ($item->sale_price ?: 0)
				);
			$db->setQuery($query);
			$db->execute();

			// Get the link id.
			$linkId = (int) $db->insertid();
		}
		else
		{
			// Update the link.
			$query->clear()
				->update($db->quoteName('#__finder_links'))
				->set($db->quoteName('route') . ' = ' . $db->quote($item->route))
				->set($db->quoteName('title') . ' = ' . $db->quote($item->title))
				->set($db->quoteName('description') . ' = ' . $db->quote($item->description))
				->set($db->quoteName('indexdate') . ' = ' . $query->currentTimestamp())
				->set($db->quoteName('state') . ' = ' . (int) $item->state)
				->set($db->quoteName('access') . ' = ' . (int) $item->access)
				->set($db->quoteName('language') . ' = ' . $db->quote($item->language))
				->set($db->quoteName('type_id') . ' = ' . (int) $item->type_id)
				->set($db->quoteName('object') . ' = ' . $db->quote(serialize($item)))
				->set($db->quoteName('publish_start_date') . ' = ' . $db->quote($item->publish_start_date))
				->set($db->quoteName('publish_end_date') . ' = ' . $db->quote($item->publish_end_date))
				->set($db->quoteName('start_date') . ' = ' . $db->quote($item->start_date))
				->set($db->quoteName('end_date') . ' = ' . $db->quote($item->end_date))
				->set($db->quoteName('list_price') . ' = ' . (double) ($item->list_price ?: 0))
				->set($db->quoteName('sale_price') . ' = ' . (double) ($item->sale_price ?: 0))
				->where('link_id = ' . (int) $linkId);
			$db->setQuery($query);
			$db->execute();
		}

		// Set up the variables we will need during processing.
		$count = 0;

		// Mark afterLinking in the profiler.
		static::$profiler ? static::$profiler->mark('afterLinking') : null;

		// Truncate the tokens tables.
		$db->truncateTable('#__finder_tokens');

		// Truncate the tokens aggregate table.
		$db->truncateTable('#__finder_tokens_aggregate');

		/*
		 * Process the item's content. The items can customize their
		 * processing instructions to define extra properties to process
		 * or rearrange how properties are weighted.
		 */
		foreach ($item->getInstructions() as $group => $properties)
		{
			// Iterate through the properties of the group.
			foreach ($properties as $property)
			{
				// Check if the property exists in the item.
				if (empty($item->$property))
				{
					continue;
				}

				// Tokenize the property.
				if (is_array($item->$property))
				{
					// Tokenize an array of content and add it to the database.
					foreach ($item->$property as $ip)
					{
						/*
						 * If the group is path, we need to a few extra processing
						 * steps to strip the extension and convert slashes and dashes
						 * to spaces.
						 */
						if ($group === static::PATH_CONTEXT)
						{
							$ip = JFile::stripExt($ip);
							$ip = str_replace(array('/', '-'), ' ', $ip);
						}

						// Tokenize a string of content and add it to the database.
						$count += $this->tokenizeToDb($ip, $group, $item->language, $format);

						// Check if we're approaching the memory limit of the token table.
						if ($count > static::$state->options->get('memory_table_limit', 30000))
						{
							$this->toggleTables(false);
						}
					}
				}
				else
				{
					/*
					 * If the group is path, we need to a few extra processing
					 * steps to strip the extension and convert slashes and dashes
					 * to spaces.
					 */
					if ($group === static::PATH_CONTEXT)
					{
						$item->$property = JFile::stripExt($item->$property);
						$item->$property = str_replace('/', ' ', $item->$property);
						$item->$property = str_replace('-', ' ', $item->$property);
					}

					// Tokenize a string of content and add it to the database.
					$count += $this->tokenizeToDb($item->$property, $group, $item->language, $format);

					// Check if we're approaching the memory limit of the token table.
					if ($count > static::$state->options->get('memory_table_limit', 30000))
					{
						$this->toggleTables(false);
					}
				}
			}
		}

		/*
		 * Process the item's taxonomy. The items can customize their
		 * taxonomy mappings to define extra properties to map.
		 */
		foreach ($item->getTaxonomy() as $branch => $nodes)
		{
			// Iterate through the nodes and map them to the branch.
			foreach ($nodes as $node)
			{
				// Add the node to the tree.
				$nodeId = FinderIndexerTaxonomy::addNode($branch, $node->title, $node->state, $node->access);

				// Add the link => node map.
				FinderIndexerTaxonomy::addMap($linkId, $nodeId);
			}
		}

		// Mark afterProcessing in the profiler.
		static::$profiler ? static::$profiler->mark('afterProcessing') : null;

		/*
		 * At this point, all of the item's content has been parsed, tokenized
		 * and inserted into the #__finder_tokens table. Now, we need to
		 * aggregate all the data into that table into a more usable form. The
		 * aggregated data will be inserted into #__finder_tokens_aggregate
		 * table.
		 */
		$query = 'INSERT INTO ' . $db->quoteName('#__finder_tokens_aggregate') .
			' (' . $db->quoteName('term_id') .
			', ' . $db->quoteName('map_suffix') .
				', ' . $db->quoteName('term') .
			', ' . $db->quoteName('stem') .
			', ' . $db->quoteName('common') .
			', ' . $db->quoteName('phrase') .
			', ' . $db->quoteName('term_weight') .
			', ' . $db->quoteName('context') .
			', ' . $db->quoteName('context_weight') .
			', ' . $db->quoteName('total_weight') .
				', ' . $db->quoteName('language') . ')' .
			' SELECT' .
			' COALESCE(t.term_id, 0), \'\', t1.term, t1.stem, t1.common, t1.phrase, t1.weight, t1.context,' .
			' ROUND( t1.weight * COUNT( t2.term ) * %F, 8 ) AS context_weight, 0, t1.language' .
			' FROM (' .
			'   SELECT DISTINCT t1.term, t1.stem, t1.common, t1.phrase, t1.weight, t1.context, t1.language' .
			'   FROM ' . $db->quoteName('#__finder_tokens') . ' AS t1' .
			'   WHERE t1.context = %d' .
			' ) AS t1' .
			' JOIN ' . $db->quoteName('#__finder_tokens') . ' AS t2 ON t2.term = t1.term' .
			' LEFT JOIN ' . $db->quoteName('#__finder_terms') . ' AS t ON t.term = t1.term' .
			' WHERE t2.context = %d' .
			' GROUP BY t1.term, t.term_id, t1.term, t1.stem, t1.common, t1.phrase, t1.weight, t1.context, t1.language' .
			' ORDER BY t1.term DESC';

		// Iterate through the contexts and aggregate the tokens per context.
		foreach ($state->weights as $context => $multiplier)
		{
			// Run the query to aggregate the tokens for this context..
			$db->setQuery(sprintf($query, $multiplier, $context, $context));
			$db->execute();
		}

		// Mark afterAggregating in the profiler.
		static::$profiler ? static::$profiler->mark('afterAggregating') : null;

		/*
		 * When we pulled down all of the aggregate data, we did a LEFT JOIN
		 * over the terms table to try to find all the term ids that
		 * already exist for our tokens. If any of the rows in the aggregate
		 * table have a term of 0, then no term record exists for that
		 * term so we need to add it to the terms table.
		 */
		$db->setQuery(
			'INSERT IGNORE INTO ' . $db->quoteName('#__finder_terms') .
			' (' . $db->quoteName('term') .
			', ' . $db->quoteName('stem') .
			', ' . $db->quoteName('common') .
			', ' . $db->quoteName('phrase') .
			', ' . $db->quoteName('weight') .
			', ' . $db->quoteName('soundex') .
			', ' . $db->quoteName('language') . ')' .
			' SELECT ta.term, ta.stem, ta.common, ta.phrase, ta.term_weight, SOUNDEX(ta.term), ta.language' .
			' FROM ' . $db->quoteName('#__finder_tokens_aggregate') . ' AS ta' .
			' WHERE ta.term_id = 0' .
			' GROUP BY ta.term, ta.stem, ta.common, ta.phrase, ta.term_weight, SOUNDEX(ta.term), ta.language'
		);
		$db->execute();

		/*
		 * Now, we just inserted a bunch of new records into the terms table
		 * so we need to go back and update the aggregate table with all the
		 * new term ids.
		 */
		$query = $db->getQuery(true)
			->update($db->quoteName('#__finder_tokens_aggregate') . ' AS ta')
			->join('INNER', $db->quoteName('#__finder_terms') . ' AS t ON t.term = ta.term')
			->set('ta.term_id = t.term_id')
			->where('ta.term_id = 0');
		$db->setQuery($query);
		$db->execute();

		// Mark afterTerms in the profiler.
		static::$profiler ? static::$profiler->mark('afterTerms') : null;

		/*
		 * After we've made sure that all of the terms are in the terms table
		 * and the aggregate table has the correct term ids, we need to update
		 * the links counter for each term by one.
		 */
		$query->clear()
			->update($db->quoteName('#__finder_terms') . ' AS t')
			->join('INNER', $db->quoteName('#__finder_tokens_aggregate') . ' AS ta ON ta.term_id = t.term_id')
			->set('t.' . $db->quoteName('links') . ' = t.links + 1');
		$db->setQuery($query);
		$db->execute();

		// Mark afterTerms in the profiler.
		static::$profiler ? static::$profiler->mark('afterTerms') : null;

		/*
		 * Before we can insert all of the mapping rows, we have to figure out
		 * which mapping table the rows need to be inserted into. The mapping
		 * table for each term is based on the first character of the md5 of
		 * the first character of the term. In php, it would be expressed as
		 * substr(md5(substr($token, 0, 1)), 0, 1)
		 */
		$query->clear()
			->update($db->quoteName('#__finder_tokens_aggregate'))
			->set($db->quoteName('map_suffix') . ' = SUBSTR(MD5(SUBSTR(' . $db->quoteName('term') . ', 1, 1)), 1, 1)');
		$db->setQuery($query);
		$db->execute();

		/*
		 * At this point, the aggregate table contains a record for each
		 * term in each context. So, we're going to pull down all of that
		 * data while grouping the records by term and add all of the
		 * sub-totals together to arrive at the final total for each token for
		 * this link. Then, we insert all of that data into the appropriate
		 * mapping table.
		 */
		for ($i = 0; $i <= 15; $i++)
		{
			// Get the mapping table suffix.
			$suffix = dechex($i);

			/*
			 * We have to run this query 16 times, one for each link => term
			 * mapping table.
			 */
			$db->setQuery(
				'INSERT INTO ' . $db->quoteName('#__finder_links_terms' . $suffix) .
				' (' . $db->quoteName('link_id') .
				', ' . $db->quoteName('term_id') .
				', ' . $db->quoteName('weight') . ')' .
				' SELECT ' . (int) $linkId . ', ' . $db->quoteName('term_id') . ',' .
				' ROUND(SUM(' . $db->quoteName('context_weight') . '), 8)' .
				' FROM ' . $db->quoteName('#__finder_tokens_aggregate') .
				' WHERE ' . $db->quoteName('map_suffix') . ' = ' . $db->quote($suffix) .
				' GROUP BY ' . $db->quoteName('term') . ', ' . $db->quoteName('term_id') .
				' ORDER BY ' . $db->quoteName('term') . ' DESC'
			);
			$db->execute();
		}

		// Mark afterMapping in the profiler.
		static::$profiler ? static::$profiler->mark('afterMapping') : null;

		// Update the signature.
		$query->clear()
			->update($db->quoteName('#__finder_links'))
			->set($db->quoteName('md5sum') . ' = ' . $db->quote($curSig))
			->where($db->quoteName('link_id') . ' = ' . $db->quote($linkId));
		$db->setQuery($query);
		$db->execute();

		// Mark afterSigning in the profiler.
		static::$profiler ? static::$profiler->mark('afterSigning') : null;

		// Truncate the tokens tables.
		$db->truncateTable('#__finder_tokens');

		// Truncate the tokens aggregate table.
		$db->truncateTable('#__finder_tokens_aggregate');

		// Toggle the token tables back to memory tables.
		$this->toggleTables(true);

		// Mark afterTruncating in the profiler.
		static::$profiler ? static::$profiler->mark('afterTruncating') : null;

		return $linkId;
	}

	/**
	 * Method to optimize the index. We use this method to remove unused terms
	 * and any other optimizations that might be necessary.
	 *
	 * @return  boolean  True on success.
	 *
	 * @since   3.0
	 * @throws  Exception on database error.
	 */
	public function optimize()
	{
		// Get the database object.
		$db = $this->db;
		$query = $db->getQuery(true);

		// Delete all orphaned terms.
		$query->delete($db->quoteName('#__finder_terms'))
			->where($db->quoteName('links') . ' <= 0');
		$db->setQuery($query);
		$db->execute();

		// Optimize the links table.
		$db->setQuery('OPTIMIZE TABLE ' . $db->quoteName('#__finder_links'));
		$db->execute();

		for ($i = 0; $i <= 15; $i++)
		{
			// Optimize the terms mapping table.
			$db->setQuery('OPTIMIZE TABLE ' . $db->quoteName('#__finder_links_terms' . dechex($i)));
			$db->execute();
		}

		// Optimize the filters table.
		$db->setQuery('OPTIMIZE TABLE ' . $db->quoteName('#__finder_filters'));
		$db->execute();

		// Optimize the terms common table.
		$db->setQuery('OPTIMIZE TABLE ' . $db->quoteName('#__finder_terms_common'));
		$db->execute();

		// Optimize the types table.
		$db->setQuery('OPTIMIZE TABLE ' . $db->quoteName('#__finder_types'));
		$db->execute();

		// Remove the orphaned taxonomy nodes.
		FinderIndexerTaxonomy::removeOrphanNodes();

		// Optimize the taxonomy mapping table.
		$db->setQuery('OPTIMIZE TABLE ' . $db->quoteName('#__finder_taxonomy_map'));
		$db->execute();

		// Optimize the taxonomy table.
		$db->setQuery('OPTIMIZE TABLE ' . $db->quoteName('#__finder_taxonomy'));
		$db->execute();

		return true;
	}


	/**
	 * Method to switch the token tables from Memory tables to MyISAM tables
	 * when they are close to running out of memory.
	 *
	 * @param   boolean  $memory  Flag to control how they should be toggled.
	 *
	 * @return  boolean  True on success.
	 *
	 * @since   3.0
	 * @throws  Exception on database error.
	 */
	protected function toggleTables($memory)
	{
		static $state;

		// Get the database adapter.
		$db = $this->db;

		// Check if we are setting the tables to the Memory engine.
		if ($memory === true && $state !== true)
		{
			// Set the tokens table to Memory.
			$db->setQuery('ALTER TABLE ' . $db->quoteName('#__finder_tokens') . ' ENGINE = MEMORY');
			$db->execute();

			// Set the tokens aggregate table to Memory.
			$db->setQuery('ALTER TABLE ' . $db->quoteName('#__finder_tokens_aggregate') . ' ENGINE = MEMORY');
			$db->execute();

			// Set the internal state.
			$state = $memory;
		}
		// We must be setting the tables to the MyISAM engine.
		elseif ($memory === false && $state !== false)
		{
			// Set the tokens table to MyISAM.
			$db->setQuery('ALTER TABLE ' . $db->quoteName('#__finder_tokens') . ' ENGINE = MYISAM');
			$db->execute();

			// Set the tokens aggregate table to MyISAM.
			$db->setQuery('ALTER TABLE ' . $db->quoteName('#__finder_tokens_aggregate') . ' ENGINE = MYISAM');
			$db->execute();

			// Set the internal state.
			$state = $memory;
		}

		return true;
	}
}