/**
* Method to index a content item.
*
* @param FinderIndexerResult $item The content item to index.
* @param string $format The format of the content. [optional]
*
* @return integer The ID of the record in the links table.
*
* @since 2.5
* @throws Exception on database error.
*/
public static function index($item, $format = 'html')
{
// Mark beforeIndexing in the profiler.
self::$profiler ? self::$profiler->mark('beforeIndexing') : null;
$db = JFactory::getDBO();
$nd = $db->getNullDate();
// Check if the item is in the database.
$query = $db->getQuery(true);
$query->select($db->quoteName('link_id') . ', ' . $db->quoteName('md5sum'));
$query->from($db->quoteName('#__finder_links'));
$query->where($db->quoteName('url') . ' = ' . $db->quote($item->url));
// Load the item from the database.
$db->setQuery($query);
$link = $db->loadObject();
// Check for a database error.
if ($db->getErrorNum()) {
// Throw database error exception.
throw new Exception($db->getErrorMsg(), 500);
}
// Get the indexer state.
$state = FinderIndexer::getState();
// Get the signatures of the item.
$curSig = self::getSignature($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();
$query->delete();
$query->from($db->quoteName('#__finder_links_terms' . dechex($i)));
$query->where($db->quoteName('link_id') . ' = ' . (int) $linkId);
$db->setQuery($query);
$db->query();
// Check for a database error.
if ($db->getErrorNum()) {
// Throw database error exception.
throw new Exception($db->getErrorMsg(), 500);
}
}
// Remove the taxonomy maps.
FinderIndexerTaxonomy::removeMaps($linkId);
}
// Mark afterUnmapping in the profiler.
self::$profiler ? self::$profiler->mark('afterUnmapping') : null;
// Perform cleanup on the item data.
$item->publish_start_date = intval($item->publish_start_date) != 0 ? $item->publish_start_date : $nd;
$item->publish_end_date = intval($item->publish_end_date) != 0 ? $item->publish_end_date : $nd;
$item->start_date = intval($item->start_date) != 0 ? $item->start_date : $nd;
$item->end_date = intval($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();
$query->insert($db->quoteName('#__finder_links'));
$query->columns($columnsArray);
$query->values($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) . ', ' . $db->quote($item->list_price) . ', ' . $db->quote($item->sale_price));
$db->setQuery($query);
$db->query();
// Check for a database error.
if ($db->getErrorNum()) {
// Throw database error exception.
throw new Exception($db->getErrorMsg(), 500);
}
// Get the link id.
$linkId = (int) $db->insertid();
} else {
// Update the link.
//@TODO: Implement this
$query->clear();
$query->update($db->qn('#__finder_links'));
$query->set($db->qn('route') . ' = ' . $db->quote($item->route));
$query->set($db->qn('title') . ' = ' . $db->quote($item->title));
$query->set($db->qn('description') . ' = ' . $db->quote($item->description));
$query->set($db->qn('indexdate') . ' = ' . $query->currentTimestamp());
$query->set($db->qn('state') . ' = ' . (int) $item->state);
$query->set($db->qn('access') . ' = ' . (int) $item->access);
$query->set($db->qn('language') . ' = ' . $db->quote($item->language));
$query->set($db->qn('type_id') . ' = ' . (int) $item->type_id);
$query->set($db->qn('object') . ' = ' . $db->quote(serialize($item)));
$query->set($db->qn('publish_start_date') . ' = ' . $db->quote($item->publish_start_date));
$query->set($db->qn('publish_end_date') . ' = ' . $db->quote($item->publish_end_date));
$query->set($db->qn('start_date') . ' = ' . $db->quote($item->start_date));
$query->set($db->qn('end_date') . ' = ' . $db->quote($item->end_date));
$query->set($db->qn('list_price') . ' = ' . $db->quote($item->list_price));
$query->set($db->qn('sale_price') . ' = ' . $db->quote($item->sale_price));
$query->where('link_id = ' . (int) $linkId);
$db->setQuery($query);
$db->query();
// Check for a database error.
if ($db->getErrorNum()) {
// Throw database error exception.
throw new Exception($db->getErrorMsg(), 500);
}
}
// Set up the variables we will need during processing.
$tokens = array();
$count = 0;
// Mark afterLinking in the profiler.
self::$profiler ? self::$profiler->mark('afterLinking') : null;
// Truncate the tokens tables.
$db->truncateTable('#__finder_tokens');
// Check for a database error.
if ($db->getErrorNum()) {
// Throw database error exception.
throw new Exception($db->getErrorMsg(), 500);
}
// Truncate the tokens aggregate table.
$db->truncateTable('#__finder_tokens_aggregate');
// Check for a database error.
if ($db->getErrorNum()) {
// Throw database error exception.
throw new Exception($db->getErrorMsg(), 500);
}
/*
* Process the item's content. The items can customize their
* processing instructions to define extra properties to process
* or rearrange how properties are weighted.
*/
foreach ($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 === self::PATH_CONTEXT) {
$ip = JFile::stripExt($ip);
$ip = str_replace('/', ' ', $ip);
$ip = str_replace('-', ' ', $ip);
}
// Tokenize a string of content and add it to the database.
$count += FinderIndexer::tokenizeToDB($ip, $group, $item->language, $format);
// Check if we're approaching the memory limit of the token table.
if ($count > self::$state->options->get('memory_table_limit', 30000)) {
FinderIndexer::toggleTables(false);
}
}
} else {
// If the group is path, we need to a few extra processing
// steps to strip the extension and convert slashes and dashes
// to spaces.
if ($group === self::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.
$count += FinderIndexer::tokenizeToDB($item->{$property}, $group, $item->language, $format);
// Check if we're approaching the memory limit of the token table.
if ($count > self::$state->options->get('memory_table_limit', 30000)) {
FinderIndexer::toggleTables(false);
}
}
}
}
/*
* Process the item's taxonomy. The items can customize their
* taxonomy mappings to define extra properties to map.
*/
foreach ($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);
// Tokenize the node title and add them to the database.
$count += FinderIndexer::tokenizeToDB($node->title, self::META_CONTEXT, $item->language, $format);
}
}
// Mark afterProcessing in the profiler.
self::$profiler ? self::$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('term') . ', ' . $db->quoteName('stem') . ', ' . $db->quoteName('common') . ', ' . $db->quoteName('phrase') . ', ' . $db->quoteName('term_weight') . ', ' . $db->quoteName('context') . ', ' . $db->quoteName('context_weight') . ')' . ' SELECT' . ' t.term_id, t1.term, t1.stem, t1.common, t1.phrase, t1.weight, t1.context,' . ' ROUND( t1.weight * COUNT( t2.term ) * %F, 8 ) AS context_weight' . ' FROM (' . ' SELECT DISTINCT t1.term, t1.stem, t1.common, t1.phrase, t1.weight, t1.context' . ' 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 ' . $db->quoteName('#__finder_terms') . ' AS t ON t.term = t1.term' . ' WHERE t2.context = %d' . ' GROUP BY t1.term' . ' ORDER BY t1.term DESC';
// Iterate through the contexts and aggregate the tokens per context.
foreach ($state->weights as $context => $multiplier) {
// Run the query to aggregate the tokens for this context..
$db->setQuery(sprintf($query, $multiplier, $context, $context));
$db->query();
// Check for a database error.
if ($db->getErrorNum()) {
// Throw database error exception.
throw new Exception($db->getErrorMsg(), 500);
}
}
// Mark afterAggregating in the profiler.
self::$profiler ? self::$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.
*/
//@TODO: PostgreSQL doesn't support SOUNDEX out of the box
/* This edit is causing the indexer to fail.
$queryInsIgn = 'INSERT INTO ' . $db->quoteName('#__finder_terms') .
' (' . $db->quoteName('term') .
', ' . $db->quoteName('stem') .
', ' . $db->quoteName('common') .
', ' . $db->quoteName('phrase') .
', ' . $db->quoteName('weight') .
', ' . $db->quoteName('soundex') . ')' .
' SELECT ta.term, ta.stem, ta.common, ta.phrase, ta.term_weight, SOUNDEX(ta.term)' .
' FROM ' . $db->quoteName('#__finder_tokens_aggregate') . ' AS ta' .
' WHERE 1 NOT IN ' .
'( SELECT 1 FROM ' . $db->quoteName('#__finder_terms') .
' WHERE ta.term_id = 0 )' .
' AND ta.term_id = 0' .
' GROUP BY ta.term';
$db->setQuery($queryInsIgn);
$db->query();
// Check for a database error.
if ($db->getErrorNum())
{
//@TODO: PostgreSQL doesn't support SOUNDEX out of the box
$query->clear();
$query->select('ta.term, ta.stem, ta.common, ta.phrase, ta.term_weight, SOUNDEX(ta.term)')
->from($db->quoteName('#__finder_tokens_aggregate') . ' AS ta')
->where('ta.term_id = 0');
$db->setQuery($query);
$subQuVal = $db->loadObject();
$quRepl_p1 = 'UPDATE ' . $db->quoteName('#__finder_terms') . ' AS ta' .
' SET ' .
' (' . $db->quoteName('term') .
', ' . $db->quoteName('stem') .
', ' . $db->quoteName('common') .
', ' . $db->quoteName('phrase') .
', ' . $db->quoteName('weight') .
', ' . $db->quoteName('soundex') . ')' .
' = ' .
' (' . $db->quote($subQuVal->term) .
', ' . $db->quote($subQuVal->stem) .
', ' . $db->quote($subQuVal->common) .
', ' . $db->quote($subQuVal->phrase) .
', ' . $db->quote($subQuVal->weight) .
', ' . $db->quote($subQuVal->soundex) . ')' .
' WHERE ' .
$db->quoteName('term') . ' = ' . $db->quote($subQuVal->term) . ' AND ' .
$db->quoteName('stem') . ' = ' . $db->quote($subQuVal->stem) . ' AND ' .
$db->quoteName('common') . ' = ' . $db->quote($subQuVal->common) . ' AND ' .
$db->quoteName('phrase') . ' = ' . $db->quote($subQuVal->phrase) . ' AND ' .
$db->quoteName('weight') . ' = ' . $db->quote($subQuVal->weight) . ' AND ' .
$db->quoteName('soundex') . ' = ' . $db->quote($subQuVal->soundex);
$db->setQuery($quRepl_p1);
$db->query();
$quRepl_p2 = 'INSERT INTO ' . $db->quoteName('#__finder_terms') .
' (' . $db->quoteName('term') .
', ' . $db->quoteName('stem') .
', ' . $db->quoteName('common') .
', ' . $db->quoteName('phrase') .
', ' . $db->quoteName('weight') .
', ' . $db->quoteName('soundex') . ')' .
' SELECT ta.term, ta.stem, ta.common, ta.phrase, ta.term_weight, SOUNDEX(ta.term)' .
' FROM ' . $db->quoteName('#__finder_tokens_aggregate') . ' AS ta' .
' WHERE 1 NOT IN ' .
'( SELECT 1 FROM ' . $db->quoteName('#__finder_terms') .
' WHERE ta.term_id = 0 )' .
' AND ta.term_id = 0' .
' GROUP BY ta.term';
$db->setQuery($quRepl_p2);
$db->query();
// Check for a database error.
if ($db->getErrorNum())
{
throw new Exception($db->getErrorMsg(), 500);
}
}
End of failing edit */
//@TODO: PostgreSQL doesn't support INSERT IGNORE INTO
//@TODO: PostgreSQL doesn't support SOUNDEX out of the box
$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') . ')' . ' SELECT ta.term, ta.stem, ta.common, ta.phrase, ta.term_weight, SOUNDEX(ta.term)' . ' FROM ' . $db->quoteName('#__finder_tokens_aggregate') . ' AS ta' . ' WHERE ta.term_id = 0' . ' GROUP BY ta.term');
$db->query();
// Check for a database error.
if ($db->getErrorNum()) {
throw new Exception($db->getErrorMsg(), 500);
}
/*
* 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);
$query->update($db->quoteName('#__finder_tokens_aggregate') . ' AS ta');
$query->join('INNER', $db->quoteName('#__finder_terms') . ' AS t ON t.term = ta.term');
$query->set('ta.term_id = t.term_id');
$query->where('ta.term_id = 0');
$db->setQuery($query);
$db->query();
// Check for a database error.
if ($db->getErrorNum()) {
// Throw database error exception.
throw new Exception($db->getErrorMsg(), 500);
}
// Mark afterTerms in the profiler.
self::$profiler ? self::$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();
$query->update($db->quoteName('#__finder_terms') . ' AS t');
$query->join('INNER', $db->quoteName('#__finder_tokens_aggregate') . ' AS ta ON ta.term_id = t.term_id');
$query->set('t.' . $db->quoteName('links') . ' = t.links + 1');
$db->setQuery($query);
$db->query();
// Check for a database error.
if ($db->getErrorNum()) {
// Throw database error exception.
throw new Exception($db->getErrorMsg(), 500);
}
// Mark afterTerms in the profiler.
self::$profiler ? self::$profiler->mark('afterTerms') : null;
/*
* Before we can insert all of the mapping rows, we have to figure out
* which mapping table the rows need to be inserted into. The mapping
* table for each term is based on the first character of the md5 of
* the first character of the term. In php, it would be expressed as
* substr(md5(substr($token, 0, 1)), 0, 1)
*/
$query->clear();
$query->update($db->quoteName('#__finder_tokens_aggregate'));
$query->set($db->quoteName('map_suffix') . ' = SUBSTR(MD5(SUBSTR(' . $db->quoteName('term') . ', 1, 1)), 1, 1)');
$db->setQuery($query);
$db->query();
// Check for a database error.
if ($db->getErrorNum()) {
// Throw database error exception.
throw new Exception($db->getErrorMsg(), 500);
}
/*
* At this point, the aggregate table contains a record for each
* term in each context. So, we're going to pull down all of that
* data while grouping the records by term and add all of the
* sub-totals together to arrive at the final total for each token for
* this link. Then, we insert all of that data into the appropriate
* mapping table.
*/
for ($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.
*/
//@TODO: Convert to JDatabaseQuery
$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') . ' ORDER BY ' . $db->quoteName('term') . ' DESC');
$db->query();
// Check for a database error.
if ($db->getErrorNum()) {
// Throw database error exception.
throw new Exception($db->getErrorMsg(), 500);
}
}
// Mark afterMapping in the profiler.
self::$profiler ? self::$profiler->mark('afterMapping') : null;
// Update the signature.
$query->clear();
$query->update($db->quoteName('#__finder_links'));
$query->set($db->quoteName('md5sum') . ' = ' . $db->quote($curSig));
$query->where($db->quoteName('link_id') . ' = ' . $db->quote($linkId));
$db->setQuery($query);
$db->query();
// Check for a database error.
if ($db->getErrorNum()) {
// Throw database error exception.
throw new Exception($db->getErrorMsg(), 500);
}
// Mark afterSigning in the profiler.
self::$profiler ? self::$profiler->mark('afterSigning') : null;
// Truncate the tokens tables.
$db->truncateTable('#__finder_tokens');
// Check for a database error.
if ($db->getErrorNum()) {
// Throw database error exception.
throw new Exception($db->getErrorMsg(), 500);
}
// Truncate the tokens aggregate table.
$db->truncateTable('#__finder_tokens_aggregate');
// Check for a database error.
if ($db->getErrorNum()) {
// Throw database error exception.
throw new Exception($db->getErrorMsg(), 500);
}
// Toggle the token tables back to memory tables.
FinderIndexer::toggleTables(true);
// Mark afterTruncating in the profiler.
self::$profiler ? self::$profiler->mark('afterTruncating') : null;
return $linkId;
}
