Smalyshev has uploaded a new change for review. (
https://gerrit.wikimedia.org/r/336911 )
Change subject: [WIP] Add indexing for descriptions
......................................................................
[WIP] Add indexing for descriptions
Descriptions follow the same pattern as labels, 'descriptions'
field includes per-language fields. Only one description per language.
Bug: T125500
Change-Id: I459c619daf8f7e6c47c1a486428f0ca38f7c8faf
---
M repo/Wikibase.php
M repo/includes/Search/Elastic/ElasticTermResult.php
M repo/includes/Search/Elastic/EntitySearchElastic.php
M repo/includes/Search/Elastic/Fields/DescriptionProviderFieldDefinitions.php
A repo/includes/Search/Elastic/Fields/DescriptionsField.php
M repo/tests/phpunit/includes/Search/Elastic/ElasticTermResultTest.php
M repo/tests/phpunit/includes/Search/Elastic/EntitySearchElasticTest.php
7 files changed, 79 insertions(+), 61 deletions(-)
git pull ssh://gerrit.wikimedia.org:29418/mediawiki/extensions/Wikibase
refs/changes/11/336911/1
diff --git a/repo/Wikibase.php b/repo/Wikibase.php
index 94007b7..2c18d85 100644
--- a/repo/Wikibase.php
+++ b/repo/Wikibase.php
@@ -212,10 +212,6 @@
$entitySearchHelper = new
Wikibase\Repo\Search\Elastic\EntitySearchElastic(
$repo->getLanguageFallbackChainFactory(),
$repo->getEntityIdParser(),
- new
Wikibase\Lib\Store\LanguageFallbackLabelDescriptionLookup(
- $repo->getTermLookup(),
-
$repo->getLanguageFallbackChainFactory()->newFromLanguage(
$repo->getUserLanguage() )
- ),
$repo->getContentModelMappings(),
$mainModule->getRequest(),
$settings
diff --git a/repo/includes/Search/Elastic/ElasticTermResult.php
b/repo/includes/Search/Elastic/ElasticTermResult.php
index acb8af4..0e2db98 100644
--- a/repo/includes/Search/Elastic/ElasticTermResult.php
+++ b/repo/includes/Search/Elastic/ElasticTermResult.php
@@ -6,7 +6,6 @@
use CirrusSearch\Search\SearchContext;
use Wikibase\DataModel\Entity\EntityIdParser;
use Wikibase\DataModel\Entity\EntityIdParsingException;
-use Wikibase\DataModel\Services\Lookup\LabelDescriptionLookup;
use Wikibase\DataModel\Term\Term;
use Wikibase\Lib\Interactors\TermSearchResult;
@@ -15,11 +14,6 @@
* a Wikibase entity by its label or alias.
*/
class ElasticTermResult implements ResultsType {
- /**
- * @var LabelDescriptionLookup
- */
- private $labelDescriptionLookup;
-
/**
* @var EntityIdParser
*/
@@ -32,11 +26,9 @@
*/
private $languageCodes;
- public function __construct( EntityIdParser $idParser,
- LabelDescriptionLookup
$labelDescriptionLookup, $languageCodes ) {
+ public function __construct( EntityIdParser $idParser, array
$languageCodes ) {
$this->idParser = $idParser;
$this->languageCodes = $languageCodes;
- $this->labelDescriptionLookup = $labelDescriptionLookup;
}
/**
@@ -48,6 +40,7 @@
$fields = [ 'namespace', 'title' ];
foreach ( $this->languageCodes as $code ) {
$fields[] = "labels.$code";
+ $fields[] = "descriptions.$code";
}
return $fields;
}
@@ -102,11 +95,12 @@
/**
* Locate label for display among the source data, basing on fallback
chain.
* @param array $sourceData
+ * @param string $field
* @return null|Term
*/
- private function findLabelForDisplay( $sourceData ) {
+ private function findLabelForDisplay( $sourceData, $field ) {
foreach ( $this->languageCodes as $code ) {
- if ( !empty( $sourceData['labels'][$code] ) &&
$sourceData['labels'][$code][0] !== '' ) {
+ if ( !empty( $sourceData[$field][$code] ) &&
$sourceData[$field][$code][0] !== '' ) {
return new Term( $code,
$sourceData['labels'][$code][0] );
}
}
@@ -131,7 +125,8 @@
$highlight = $r->getHighlights();
$matchedTermType = 'label';
- $displayLabel = $this->findLabelForDisplay( $sourceData
);
+ $displayLabel = $this->findLabelForDisplay(
$sourceData, 'labels' );
+ $displayDescription = $this->findLabelForDisplay(
$sourceData, 'descriptions' );
if ( !empty( $highlight['title'] ) ) {
$matchedTermType = 'entityId';
@@ -162,8 +157,6 @@
// This should not happen, but just in case,
it's better to return something
$displayLabel = $matchedTerm;
}
- // TODO: eventually this will be fetched from ES.
- $displayDescription =
$this->labelDescriptionLookup->getDescription( $entityId );
$results[$entityId->getSerialization()] = new
TermSearchResult(
$matchedTerm, $matchedTermType, $entityId,
$displayLabel,
diff --git a/repo/includes/Search/Elastic/EntitySearchElastic.php
b/repo/includes/Search/Elastic/EntitySearchElastic.php
index d468aea..ece804e 100644
--- a/repo/includes/Search/Elastic/EntitySearchElastic.php
+++ b/repo/includes/Search/Elastic/EntitySearchElastic.php
@@ -10,7 +10,6 @@
use RequestContext;
use WebRequest;
use Wikibase\DataModel\Entity\EntityIdParser;
-use Wikibase\DataModel\Services\Lookup\LabelDescriptionLookup;
use Wikibase\LanguageFallbackChainFactory;
use Wikibase\Lib\Interactors\TermSearchResult;
use Wikibase\Repo\Api\EntitySearcher;
@@ -29,11 +28,6 @@
* @var LanguageFallbackChainFactory
*/
private $languageChainFactory;
- /**
- * @var LabelDescriptionLookup
- */
- private $labelDescriptionLookup;
-
/**
* @var EntityIdParser
*/
@@ -60,7 +54,6 @@
/**
* @param LanguageFallbackChainFactory $languageChainFactory
* @param EntityIdParser $idParser
- * @param LabelDescriptionLookup $labelDescriptionLookup
* @param $contentModelMap
* @param WebRequest $request
* @param array $settings
@@ -68,14 +61,12 @@
public function __construct(
LanguageFallbackChainFactory $languageChainFactory,
EntityIdParser $idParser,
- LabelDescriptionLookup $labelDescriptionLookup,
$contentModelMap,
WebRequest $request,
array $settings
) {
$this->languageChainFactory = $languageChainFactory;
$this->idParser = $idParser;
- $this->labelDescriptionLookup = $labelDescriptionLookup;
$this->contentModelMap = $contentModelMap;
$this->request = $request;
$this->settings = $settings;
@@ -203,7 +194,6 @@
$searcher = new WikibasePrefixSearcher( 0, $limit );
$searcher->setResultsType( new ElasticTermResult(
$this->idParser,
- $this->labelDescriptionLookup,
$this->fallbackCodes
) );
diff --git
a/repo/includes/Search/Elastic/Fields/DescriptionProviderFieldDefinitions.php
b/repo/includes/Search/Elastic/Fields/DescriptionProviderFieldDefinitions.php
index 6d07ed2..6df0bb5 100644
---
a/repo/includes/Search/Elastic/Fields/DescriptionProviderFieldDefinitions.php
+++
b/repo/includes/Search/Elastic/Fields/DescriptionProviderFieldDefinitions.php
@@ -25,7 +25,7 @@
public function getFields() {
$fields = [];
- // TODO: nothing for now, will add later
+ $fields['descriptions'] = new DescriptionsField(
$this->languageCodes );
return $fields;
}
diff --git a/repo/includes/Search/Elastic/Fields/DescriptionsField.php
b/repo/includes/Search/Elastic/Fields/DescriptionsField.php
new file mode 100644
index 0000000..361fdaa
--- /dev/null
+++ b/repo/includes/Search/Elastic/Fields/DescriptionsField.php
@@ -0,0 +1,71 @@
+<?php
+namespace Wikibase\Repo\Search\Elastic\Fields;
+
+use Flow\Import\SourceStore\Exception;
+use SearchEngine;
+use SearchIndexFieldDefinition;
+use Wikibase\DataModel\Entity\EntityDocument;
+use Wikibase\DataModel\Term\AliasesProvider;
+use Wikibase\DataModel\Term\DescriptionsProvider;
+use Wikibase\DataModel\Term\LabelsProvider;
+
+/**
+ * Field which contains per-language specific descriptions.
+ */
+class DescriptionsField implements WikibaseIndexField {
+
+ /**
+ * List of available languages
+ * @var string[]
+ */
+ private $languages;
+
+ public function __construct( $languages ) {
+ $this->languages = $languages;
+ }
+
+ /**
+ * @param SearchEngine $engine
+ * @param string $name
+ * @return null|\SearchIndexField
+ */
+ public function getMappingField( SearchEngine $engine, $name ) {
+ $field = $engine->makeSearchFieldMapping( $name,
\SearchIndexField::INDEX_TYPE_NESTED );
+ if ( is_callable( [ $field, 'setMergeCallback' ] ) ) {
+ $field->setMergeCallback( function ( $that ) use (
$field ) {
+ return $field;
+ } );
+
+ }
+ /**
+ * @var SearchIndexFieldDefinition
+ */
+ foreach ( $this->languages as $language ) {
+ $langConfig = $engine->makeSearchFieldMapping(
"$name.$language",
+ \SearchIndexField::INDEX_TYPE_TEXT );
+
+ $field->addSubfield( $language, $langConfig );
+ }
+
+ return $field;
+ }
+
+ /**
+ * @param EntityDocument $entity
+ *
+ * @return mixed Get the value of the field to be indexed when a
page/document
+ * is indexed. This might be an array with nested data,
if the field
+ * is defined with nested type or an int or string for
simple field types.
+ */
+ public function getFieldData( EntityDocument $entity ) {
+ if ( !( $entity instanceof DescriptionsProvider ) ) {
+ return [];
+ }
+ $data = [];
+ foreach ( $entity->getDescriptions() as $language => $desc ) {
+ $data[$language] = $desc->getText();
+ }
+ return $data;
+ }
+
+}
diff --git
a/repo/tests/phpunit/includes/Search/Elastic/ElasticTermResultTest.php
b/repo/tests/phpunit/includes/Search/Elastic/ElasticTermResultTest.php
index 5b28efb..d21578f 100644
--- a/repo/tests/phpunit/includes/Search/Elastic/ElasticTermResultTest.php
+++ b/repo/tests/phpunit/includes/Search/Elastic/ElasticTermResultTest.php
@@ -164,21 +164,6 @@
}
/**
- * Get a lookup that always returns a pt label and description suffixed
by the entity ID
- *
- * @return LabelDescriptionLookup
- */
- private function getMockLabelDescriptionLookup() {
- $mock = $this->getMockBuilder( LabelDescriptionLookup::class )
- ->disableOriginalConstructor()
- ->getMock();
- $mock->expects( $this->any() )
- ->method( 'getDescription' )
- ->will( $this->returnValue( new Term( 'en',
'DESCRIPTION' ) ) );
- return $mock;
- }
-
- /**
* @dataProvider termResultsProvider
* @param $languages
* @param $resultData
@@ -187,7 +172,6 @@
public function testTransformResult( $languages, $resultData, $expected
) {
$res = new ElasticTermResult(
new BasicEntityIdParser(),
- $this->getMockLabelDescriptionLookup(),
$languages
);
diff --git
a/repo/tests/phpunit/includes/Search/Elastic/EntitySearchElasticTest.php
b/repo/tests/phpunit/includes/Search/Elastic/EntitySearchElasticTest.php
index e860c24..c035026 100644
--- a/repo/tests/phpunit/includes/Search/Elastic/EntitySearchElasticTest.php
+++ b/repo/tests/phpunit/includes/Search/Elastic/EntitySearchElasticTest.php
@@ -14,28 +14,12 @@
*/
class EntitySearchElasticTest extends MediaWikiTestCase {
- /**
- * Get a lookup that always returns a fixed label and description
- *
- * @return LabelDescriptionLookup
- */
- private function getMockLabelDescriptionLookup() {
- $mock = $this->getMockBuilder( LabelDescriptionLookup::class )
- ->disableOriginalConstructor()
- ->getMock();
- $mock->expects( $this->any() )
- ->method( 'getDescription' )
- ->will( $this->returnValue( new Term( 'en',
'DESCRIPTION' ) ) );
- return $mock;
- }
-
private function newEntitySearch() {
$repo = \Wikibase\Repo\WikibaseRepo::getDefaultInstance();
return new EntitySearchElastic(
$repo->getLanguageFallbackChainFactory(),
new BasicEntityIdParser(),
- $this->getMockLabelDescriptionLookup(),
$repo->getContentModelMappings(),
$this->getMockRequest(),
$repo->getSettings()->getSetting( 'entitySearch' )
--
To view, visit https://gerrit.wikimedia.org/r/336911
To unsubscribe, visit https://gerrit.wikimedia.org/r/settings
Gerrit-MessageType: newchange
Gerrit-Change-Id: I459c619daf8f7e6c47c1a486428f0ca38f7c8faf
Gerrit-PatchSet: 1
Gerrit-Project: mediawiki/extensions/Wikibase
Gerrit-Branch: master
Gerrit-Owner: Smalyshev <[email protected]>
_______________________________________________
MediaWiki-commits mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/mediawiki-commits
