hazza/Phraseanet
1
0
Fork 0
mirror of https://github.com/alchemy-fr/Phraseanet.git synced 2025-10-17 15:03:25 +00:00
Code Issues Packages Projects Releases Wiki Activity

Refactor thesaurus query build

Browse Source 
- Look for text nodes and infer the concepts behind term using thesaurus
- Use value objects for thesaurus terms and concepts
- Pass a QueryContext holding allowed fields and locales informations when building the Elasticsearch query
- Change type hinting and name of query building method on nodes
- Remove unused method Node#isFullTextOnly()
- Move getFieldsStructure from RecordIndexer to RecordHelper for reusing field structure in SearchEngine
This commit is contained in:
Mathieu Darse
2015-01-15 20:04:46 +01:00
parent f283bf01d1
commit dc2c9f8c7f
21 changed files with 391 additions and 267 deletions
Download Patch File Download Diff File Expand all files Collapse all files

132
lib/Alchemy/Phrasea/SearchEngine/Elastic/ElasticSearchEngine.php
View File

@@ -13,7 +13,9 @@ namespace Alchemy\Phrasea\SearchEngine\Elastic;
use Alchemy\Phrasea\SearchEngine\Elastic\Indexer\RecordIndexer;
use Alchemy\Phrasea\SearchEngine\Elastic\Indexer\TermIndexer;
use Alchemy\Phrasea\SearchEngine\Elastic\Search\SearchQuery;
use Alchemy\Phrasea\SearchEngine\Elastic\Search\QueryContext;
use Alchemy\Phrasea\SearchEngine\Elastic\Thesaurus\Concept;
use Alchemy\Phrasea\SearchEngine\Elastic\Thesaurus\Term;
use Alchemy\Phrasea\SearchEngine\SearchEngineInterface;
use Alchemy\Phrasea\SearchEngine\SearchEngineOptions;
use Alchemy\Phrasea\SearchEngine\SearchEngineResult;
@@ -280,83 +282,26 @@ class ElasticSearchEngine implements SearchEngineInterface
$query['_ast'] = $searchQuery->dump();
$thesaurus = $this->app['thesaurus'];
foreach ($searchQuery->getTextNodes() as $textNode) {
$text = $textNode->getText();
$concepts = $thesaurus->findConcepts($text);
$query['_thesaurus_concepts'][$text] = $concepts;
$textNodes = $searchQuery->getTextNodes();
$concepts = $thesaurus->findConceptsBulk($textNodes);
foreach ($concepts as $index => $termConcepts) {
$node = $textNodes[$index];
$node->setConcepts($termConcepts);
$term = Term::dump($node);
$query['_thesaurus_concepts'][$term] = Concept::toPathArray($termConcepts);
}
// $concepts = $thesaurus->findConceptsBulk($terms);
$recordHelper = $this->app['elasticsearch.record_helper'];
// TODO Pass options to getFields to include/exclude private fields
$searchableFields = $recordHelper->getFields();
$queryContext = new QueryContext($searchableFields, $this->locales, $this->app['locale']);
$recordQuery = $searchQuery->build($queryContext);
// Contains the full thesaurus paths to search on
$pathsToFilter = [];
// Contains the thesaurus values by fields (synonyms, translations, etc)
$collectFields = [];
// Only search in thesaurus for full text search
if ($searchQuery->isFullTextOnly()) {
$termFields = $this->expendToAnalyzedFieldsNames('value', null, $this->app['locale']);
$termsQuery = $searchQuery->getElasticsearchQuery($termFields);
$params = $this->createTermQueryParams($termsQuery, $options);
$terms = $this->doExecute('search', $params);
foreach ($terms['hits']['hits'] as $term) {
// Skip paths with very low score
if ($term['_score'] < 1) {
continue;
}
$pathsToFilter[$term['_source']['path']] = $term['_score'];
foreach ($term['_source']['fields'] as $field) {
$collectFields['caption.'.$field][] = $term['_source']['value'];
}
}
$pathsToFilter = array_unique($pathsToFilter);
}
if (empty($collectFields)) {
// @todo a list of field by default? all fields?
$searchFieldNames = ['caption.*'];
} else {
$searchFieldNames = array_keys($collectFields);
}
$recordFields = $this->expendToAnalyzedFieldsNames($searchFieldNames, null, $this->app['locale']);
$recordQuery = [
'bool' => [
'should' => [
$searchQuery->getElasticsearchQuery($recordFields)
]
]
];
foreach ($pathsToFilter as $path => $score) {
// Also match incomplete path. /a/b/c will return /a/b/c/d records
$recordQuery['bool']['should'][] = [
'match' => [
'concept_paths' => array(
'query' => $path,
'boost' => $score,
)
]
];
// Add signal for exact path only
$recordQuery['bool']['should'][] = [
'term' => [
'concept_paths.raw' => array(
'value' => $path,
'boost' => $score,
)
]
];
}
$params = $this->createRecordQueryParams($recordQuery, $options, null);
$params['body']['from'] = $offset;
$params['body']['size'] = $perPage;
@@ -390,10 +335,15 @@ class ElasticSearchEngine implements SearchEngineInterface
$results[] = ElasticsearchRecordHydrator::hydrate($hit['_source'], $n++);
}
$query['_searchable_fields'] = $searchableFields;
$query['_ast'] = $searchQuery->dump();
$query['_paths'] = $pathsToFilter;
$query['_richFields'] = $collectFields;
$query['query'] = json_encode($params);
// $query['_paths'] = $pathsToFilter;
// $query['_richFields'] = $collectFields;
$queryyy = $recordQuery;
// $queryyy = $params['body'];
$query['query'] = $queryyy;
$query['query_as_string'] = json_encode($queryyy);
return new SearchEngineResult($results, json_encode($query), $res['took'], $offset,
$res['hits']['total'], $res['hits']['total'], null, null, $suggestions, [],
@@ -584,36 +534,6 @@ class ElasticSearchEngine implements SearchEngineInterface
return $res;
}
/**
* @param array|string $fields
* @param array|null $locales
* @param null $currentLocale
* @return array
*/
public function expendToAnalyzedFieldsNames($fields, $locales = null, $currentLocale = null)
{
$fieldsExpended = [];
if (!$locales) {
$locales = $this->locales;
}
foreach ((array) $fields as $field) {
foreach ($locales as $locale) {
$boost = "";
if ($locale === $currentLocale) {
$boost = "^5";
}
$fieldsExpended[] = sprintf('%s.%s%s', $field, $locale, $boost);
}
$fieldsExpended[] = sprintf('%s.%s', $field, 'light^10');
}
return $fieldsExpended;
}
private function getFlagsKey(\appbox $appbox)
{
$flags = [];