On Packagist:
WikibaseEntityStore is a small library that provides an unified interface to interact with Wikibase entities.
It currently has two backends:
- An API based backend, slow but very easy to deploy
- A MongoDB based backend, faster but requires to maintain a local copy of Wikibase content
Use one of the below methods:
1 - Use composer to install the library and all its dependencies using the master branch:
composer require "ppp/wikibase-entity-store":dev-master"
2 - Create a composer.json file that just defines a dependency on version 1.0 of this package, and run 'composer install' in the directory:
{
"require": {
"ppp/wikibase-entity-store": "~1.0"
}
}The entity storage system is based on the abstract class EntityStore that provides different services to manipulate entities.
These services are:
$storeBuilder = new EntityStoreFromConfigurationBuilder();
$store = $storeBuilder->buildEntityStore( 'MY_CONFIG_FILE.json' ); //See backend section for examples of configuration file
//Retrieves the item Q1
try {
$item = $store->getItemLookup()->getItemForId( new ItemId( 'Q1' ) );
} catch( ItemNotFoundException $e ) {
//Item not found
}
//Retrieves the property P1
try {
$item = $store->getPropertyLookup()->getPropertyForId( new PropertyId( 'P1' ) );
} catch( PropertyNotFoundException $e ) {
//Property not found
}
//Retrieves the item Q1 as EntityDocument
try {
$item = $store->getEntityLookup()->getEntityDocumentForId( new ItemId( 'Q1' ) );
} catch( EntityNotFoundException $e ) {
//Property not found
}
//Retrieves the item Q1 and the property P1 as EntityDocuments
$entities = $store->getEntityLookup()->getEntityDocumentsForIds( array( new ItemId( 'Q1' ), new PropertyId( 'P1' ) ) );
//Retrieves the ids of the items that have as label or alias the term "Nyan Cat" in English (with a case insensitive compare)
$itemIds = $store->getItemIdForTermLookup()->getItemIdsForTerm( new Term( 'en', 'Nyan Cat' ) );
//Retrieves the ids of the properties that have as label or alias the term "foo" in French (with a case insensitive compare)
$propertyIds = $store->getPropertyIdForTermLookup()->getPropertyIdsForTerm( new Term( 'fr', 'Foo' ) );
//Do a query on items using the Ask query language: retrieves the first 10 items with P1: Q1
$itemIds = $store->getItemIdForQueryLookup()->getItemIdsForQuery(
new Query(
new SomeProperty(
new EntityIdValue( new PropertyId( 'P1' ) ),
new ValueDescription( new EntityIdValue( new ItemId( 'Q1' ) ) )
),
array(),
new QueryOptions( 10, 0 )
)
);The API backend is the most easy to use one. It uses the API of a Wikibase instance and WikidataQuery if you use this EntityStore as a backend for Wikidata data and you want query support.
The configuration file looks like:
{
"backend": "api",
"api": {
"url": "http://www.wikidata.org/w/api.php",
"wikidataquery-url": "http://wdq.wmflabs.org/api"
}
}Replace http://www.wikidata.org/w/api.php with the URL of your WediaWiki API if you want to use your store with an other Wikibase instance than Wikidata.
The parameter wikidataquery-url is optional and may be unset if you don't want query support using Wikidata content.
Without configuration file:
$store = new \Wikibase\EntityStore\Api\ApiEntityStore(
new \Mediawiki\Api\MediawikiApi( 'http://www.wikidata.org/w/api.php' ),
new \WikidataQueryApi\WikidataQueryApi( 'http://wdq.wmflabs.org/api' )
);The MongoDB backend uses a MongoDB database. Requires doctrine/mongodb.
The configuration file looks like:
{
"backend": "mongodb",
"mongodb": {
"server": SERVER,
"database": DATABASE
}
}server should be a MongoDB server connection string and database the name of the database to use.
Without configuration file:
//Connect to MongoDB
$connection = new \Doctrine\MongoDB\Connection( MY_CONNECTION_STRING );
if( !$connection->connect() ) {
throw new RuntimeException( 'Fail to connect to the database' );
}
//Gets the database where entities are stored
$database = $connection
->selectDatabase( 'wikibase' );
$store = new \Wikibase\EntityStore\MongoDB\MongDBEntityStore( $database );You can fill the MongoDB database from Wikidata JSON dumps using this script:
php entitystore import-json-dump MY_JSON_DUMP MY_CONFIGURATION_FILE
Or from incremental XML dumps using this script:
php entitystore import-incremental-xml-dump MY_XML_DUMP MY_CONFIGURATION_FILE
Backend based on an array of EntityDocuments. Useful for tests.
$store = new \Wikibase\EntityStore\InMemory\InMemoryEntityStore( array(
new Item( new ItemId( 'Q42' ) )
) );The different backends support a shared set of options. These options are:
- string[]
languages: Allows to filter the set of internationalized values. Default value:null(a.k.a. all languages). - bool
languagefallback: Apply language fallback system to languages defined using languages option. Default value:false.
They can be injected in the configuration:
{
"options": {
"languages": ["en", "fr"],
"languagefallback": true
}
}They can be also passed as last parameter of EntityStore constructors:
$options = new \Wikibase\EntityStore\EntityStoreOptions( array(
EntityStore::OPTION_LANGUAGES => array( 'en', 'fr' ),
EntityStore::OPTION_LANGUAGE_FALLBACK => true
) );
$store = new \Wikibase\EntityStore\Api\ApiEntityStore(
new \Mediawiki\Api\MediawikiApi( 'http://www.wikidata.org/w/api.php' ),
null,
$options
);It is possible, in order to get far better performances, to add a cache layer on top of EntityStore:
Adds to the configuration file a cache section.
Example with a two layers cache. The first one is a PHP array and the second one a Memcached instance on localhost:11211.
{
"cache": {
"array": true,
"memcached": {
"host": "localhost",
"port": 11211
}
}
}Without configuration file:
$store = MY_ENTITY_STORE;
$cache = new \Doctrine\Common\Cache\ArrayCache(); //A very simple cache
$cacheLifeTime = 100000; //Life time of cache in seconds
$cachedStore = new \Wikibase\EntityStore\Cache\CachedEntityStore( $store, $cache, $cacheLifeTime );- Adds support of WikibaseDataModel 5.0 and 6.0 and Number 0.7
- Update to WikibaseDataModel 4.0, WikibaseDataModelServices 3.2 and MediaWikiApiBase 2.0
- Change encoding of label indexes in MongoDB
Initial release with these features:
- Retrieve entities from ids or terms
- Import Wikidata JSON full dumps and incremental XML dumps
- Beginning of support of simple queries
- API and MongoDB backends
- Basic cache system