mod/elgg_solr/lib/hooks.php ➔ elgg_solr_object_search()

Complexity

Conditions	27
Paths	> 20000

Size

Total Lines

187

Duplication

Lines	79
Ratio	42.25 %

Code Coverage

Tests	0
CRAP Score	756

Importance

Changes

0

<?php
function elgg_solr_object_search($hook, $type, $return, $params) {

    $select = array(
        'start'  => $params['offset'],
        'rows'   => $params['limit'] ? $params['limit'] : 10,
        'fields' => array('id','title','description','score')
    );
	
	if ($params['select'] && is_array($params['select'])) {
        $select = array_merge($select, $params['select']);
    }
	
    // create a client instance
    $client = elgg_solr_get_client($select);

The call to elgg_solr_get_client() has too many arguments starting with $select.

    // get an update query instance
    $query = $client->createSelect($select);
	
	$title_boost = elgg_solr_get_title_boost();
	$description_boost = elgg_solr_get_description_boost();
	
	// get the dismax component and set a boost query
	$dismax = $query->getEDisMax();
	$qf = "title^{$title_boost} description^{$description_boost}";
	if ($params['qf']) {
		$qf = $params['qf'];
	}
	$dismax->setQueryFields($qf);
	$dismax->setQueryAlternative('*:*');
	
	$boostQuery = elgg_solr_get_boost_query();
	if ($boostQuery) {
		$dismax->setBoostQuery($boostQuery);
	}
	
	// this query is now a dismax query
	$query->setQuery($params['query']);
	
	$default_sorts = array(
		'score' => 'desc',
		'time_created' => 'desc'
	);
	
	$sorts = $params['sorts'] ? $params['sorts'] : $default_sorts;
	$query->addSorts($sorts);
	
	// make sure we're only getting objectss
	$params['fq']['type'] = 'type:object';

	$default_fq = elgg_solr_get_default_fq($params);
	if ($params['fq']) {
		$filter_queries = array_merge($default_fq, $params['fq']);
	}
	else {
		$filter_queries = $default_fq;
	}

    if (!empty($filter_queries)) {
        foreach ($filter_queries as $key => $value) {
            $query->createFilterQuery($key)->setQuery($value);
        }
    }

    // get highlighting component and apply settings
    $hl = $query->getHighlighting();
	$hlfields = array('title', 'description');
	if ($params['hlfields']) {
		$hlfields = $params['hlfields'];
	}
    $hl->setFields($hlfields);
	$hl->setSimplePrefix('<span data-hl="elgg-solr">');
	$hl->setSimplePostfix('</span>');
	
	$fragsize = elgg_solr_get_fragsize();
	if (isset($params['fragsize'])) {
		$fragsize = (int) $params['fragsize'];
	}
	$hl->setFragSize($fragsize);

	
    // this executes the query and returns the result
    try {
        $resultset = $client->select($query);
    } catch (Exception $e) {
        error_log($e->getMessage());
        return null;
    }

    // Get the highlighted snippet
    try {
        $highlighting = $resultset->getHighlighting();
    } catch (Exception $e) {
        error_log($e->getMessage());
        return null;
    }

    // Count the total number of documents found by solr
    $count = $resultset->getNumFound();
	$hl_prefix = elgg_solr_get_hl_prefix();
	$hl_suffix = elgg_solr_get_hl_suffix();
	
	$config = HTMLPurifier_Config::createDefault();
	$purifier = new HTMLPurifier($config);

    $search_results = array();
    foreach ($resultset as $document) {
		$search_results[$document->id] = array();
		$snippet = '';
            
		// highlighting results can be fetched by document id (the field defined as uniquekey in this schema)
        $highlightedDoc = $highlighting->getResult($document->id);

        if($highlightedDoc){
            foreach($highlightedDoc as $field => $highlight) {
                $snippet = implode(' (...) ', $highlight);
				// get our highlight based on the wrapped tokens
				// note, this is to prevent partial html from breaking page layouts
				preg_match_all('/<span data-hl="elgg-solr">(.*?)<\/span>/', $snippet, $match);

				if ($match[1]) {
					$matches = array_unique($match[1]);
					foreach ($matches as $m) {
						$snippet = str_replace($m, $hl_prefix . $m . $hl_suffix, $snippet);	
					}
					$snippet = $purifier->purify($snippet);
				}
				
				$search_results[$document->id][$field] = $snippet;
            }
        }
		
		$search_results[$document->id]['score'] = $document->score;
    }
	
	// get the entities
	$entities = array();
	$entities_unsorted = array();
	if ($search_results) {

The expression $search_results of type array is implicitly converted to a boolean; are you sure this is intended? If so, consider using ! empty($expr) instead to make it clear that you intend to check for an array without elements.

		$entities_unsorted = elgg_get_entities(array(
			'guids' => array_keys($search_results),
			'limit' => false
		));
	}
	
	$show_score = elgg_get_plugin_setting('show_score', 'elgg_solr');
	foreach ($search_results as $guid => $matches) {
		foreach ($entities_unsorted as $e) {
			
			$desc_suffix = '';
			if ($show_score == 'yes' && elgg_is_admin_logged_in()) {
				$desc_suffix .= elgg_view('output/longtext', array(
					'value' => elgg_echo('elgg_solr:relevancy', array($matches['score'])),
					'class' => 'elgg-subtext'
				));
			}
			
			if ($e->guid == $guid) {
				if ($matches['title']) {
					$e->setVolatileData('search_matched_title', $matches['title']);
				}
				else {
					$e->setVolatileData('search_matched_title', $e->title);
				}
				
				if ($matches['description']) {
					$desc = $matches['description'];
				}
				else {
					$desc = elgg_get_excerpt($e->description, 100);
				}
				
				unset($matches['title']);
				unset($matches['description']);
				unset($matches['score']);
				$desc .= implode('...', $matches);
				
				$e->setVolatileData('search_matched_description', $desc . $desc_suffix);
				$entities[] = $e;
			}
		}
	}

    return array(
        'entities' => $entities,
        'count' => $count,
    );
}