详解Java的Hibernate框架中的搜索工具的运用

hibernate提供了全文索引功能，非常棒，这里简要介绍下它的用法，

1. 在pom.xml引入包依赖

<dependency> <groupId>org.hibernate</groupId> <artifactId>hibernate-search-orm</artifactId> <version>${hibernate-search.version}</version> </dependency> <dependency> <groupId>org.apache.lucene</groupId> <artifactId>lucene-analyzers-smartcn</artifactId> <version>${lucene.version}</version> </dependency> <dependency> <groupId>org.apache.lucene</groupId> <artifactId>lucene-queryparser</artifactId> <version>${lucene.version}</version> </dependency> <dependency> <groupId>org.apache.lucene</groupId> <artifactId>lucene-analyzers-phonetic</artifactId> <version>${lucene.version}</version> </dependency>

hibernate配置 search index保存路径

<bean id="sessionFactory" destroy-method="destroy"> <property name="dataSource" ref="poolingDataSource" /> <property name="configLocation"> <value> classpath:hibernate.cfg.xml </value> </property> <property name="hibernateProperties"> <props> <prop key="hibernate.dialect">${hibernate.dialect}</prop>  <prop key="hibernate.query.substitutions">true 'Y', false 'N'</prop> <> <> <prop key="hibernate.cache.use_second_level_cache">true</prop> <> <prop key="hibernate.cache.region.factory_class">org.hibernate.cache.ehcache.EhCacheRegionFactory</prop>  <prop key="hibernate.cache.use_query_cache">true</prop> <prop key="net.sf.ehcache.configurationResourceName">ehcache-hibernate.xml</prop> <> ***<prop key="hibernate.search.default.indexBase">indexes/</prop>*** </props> </property> </bean>

对需要搜索的类加上Indexed Annotation，然后对类中可以被搜索的字段加上@Field Annotation，通常Enum字段不需要Analyzer进行词法分析，其他字段则需要，对于不需要Projection（返回部分字段）的情况下，不需要在index中存储实际数据。可以通过AnalyzerDef来定义不同的词法分析器以及对于的特殊词过滤器

@Indexed @AnalyzerDef( name="enTopicAnalyzer", charFilters={ @CharFilterDef(factory=HTMLStripCharFilterFactory.class) }, tokenizer=@TokenizerDef(factory=StandardTokenizerFactory.class), filters={ @TokenFilterDef(factory=StandardFilterFactory.class), @TokenFilterDef(factory=StopFilterFactory.class), @TokenFilterDef(factory=PhoneticFilterFactory.class, params = { @Parameter(name="encoder", value="DoubleMetaphone") }), @TokenFilterDef(factory=SnowballPorterFilterFactory.class, params = { @Parameter(name="language", value="English") }) } ) public class Topic { ...... @Field(index=Index.YES, analyze=Analyze.YES, store=Store.NO) @Analyzer(definition = "enTopicAnalyzer") private String title; ...... @Field(index=Index.YES, analyze=Analyze.YES, store=Store.NO) @Analyzer(definition = "enTopicAnalyzer") private String content; ...... @Enumerated(EnumType.STRING) @Field(index=Index.YES, analyze=Analyze.NO, store=Store.NO, bridge=@FieldBridge(impl=EnumBridge.class)) private TopicStatus status; ... }

通过代码对已有数据创建index

ApplicationContext context = new ClassPathXmlApplicationContext("spring-resources.xml"); SessionFactory sessionFactory = (SessionFactory) context.getBean("sessionFactory"); Session sess = sessionFactory.openSession(); FullTextSession fullTextSession = Search.getFullTextSession(sess); try { fullTextSession.createIndexer().startAndWait(); } catch (InterruptedException e) { LOG.error(e.getMessage(), e); } finally { fullTextSession.close(); } ((AbstractApplicationContext)context).close();

创建查询fulltextsession，按照query条件获取结果

FullTextSession fullTextSession = Search .getFullTextSession(getSession()); QueryBuilder queryBuilder = fullTextSession.getSearchFactory() .buildQueryBuilder().forEntity(Show.class).get(); org.apache.lucene.search.Query luceneQuery = null; luceneQuery = queryBuilder.keyword()// .wildcard() .onFields("title", "content").matching(query.getKeyword()) // .matching("*" + query.getKeyword() + "*") .createQuery(); FullTextQuery hibernateQuery = fullTextSession.createFullTextQuery( luceneQuery, Show.class); return hibernateQuery.list();

note：

1. 在一次测试过程中，修改了value object，添加了新的index，忘记了rebuildIndex，结果unit test没问题，生成环境就出错了。

2. 搜索还不是很强大，比如搜索测，含有测试的结果可能就搜索不出来

中文词法分析

hibernate search底层使用Lucene，所以Lucene可以使用的中文分词，hibernate search都可以用来支持中文词法分析，比较常用的词法分析器包括paoding，IKAnalyzer，mmseg4j 等等。具体可以参考分词分析最近分析。hibernate search默认的分词器是org.apache.lucene.analysis.standard.StandardAnalyzer，中文按字分词，显然不符合我们的需求。

这里介绍一下如何在hibernate中配置中文分词，选择的是Lucene自带的中文分词–。使用可以通过3种方式，一种是在hibernate的配置文件设置词法分析方法，另外一种是在每个需要被搜索的类中定义分词方法，最后一种是对单个字段配置。这里介绍下前2种的配置方式。

hibernate配置方式：

<property name="hibernate.search.analyzer"> org.apache.lucene.analysis.cn.smart.SmartChineseAnalyzer</property>

被搜索类配置中文分词：

@Indexed @Analyzer(impl=SmartChineseAnalyzer.class)

同时需要在maven中引入相关包依赖

<dependency> <groupId>org.apache.lucene</groupId> <artifactId>lucene-analyzers-smartcn</artifactId> <version>${lucene.version}</version> </dependency>

多条件查询

hibernate search可以通过多组合条件来实现多条件查询，这里简单介绍一下多条件查询的一个实践。

如果只是单个条件查询，那么这个查询就可以很简单

luceneQuery = queryBuilder.keyword().onFields("title", "content").matching(query.getKeyword()).createQuery()

如果是多条件并查询，那么就需要使用到Must Join，如果是多条件或查询，就需要使用should Join，这里举个Must Join的例子

//must true MustJunction term = queryBuilder.bool().must(queryBuilder.keyword() .onFields("title", "content") .matching(query.getKeyword()).createQuery()); //must false term.must(queryBuilder.keyword() .onField("status") .matching(query.getExcludeStatus()).createQuery()).not();

完整例子：

private FullTextQuery findByKeywordQuery(TopicQuery query) { FullTextSession fullTextSession = Search .getFullTextSession(getSession()); QueryBuilder queryBuilder = fullTextSession.getSearchFactory() .buildQueryBuilder().forEntity(Topic.class).get(); org.apache.lucene.search.Query luceneQuery = null; if (null == query.getStatus() && null == query.getUsername() && null == query.getExcludeStatus()) { luceneQuery = queryBuilder.keyword()// .wildcard() .onFields("title", "content").matching(query.getKeyword()) // .matching("*" + query.getKeyword() + "*") .createQuery(); if(LOG.isDebugEnabled()){ LOG.debug("create clean keyword search query: " + luceneQuery.toString()); } } else { MustJunction term = queryBuilder.bool().must(queryBuilder.keyword() .onFields("title", "content") .matching(query.getKeyword()).createQuery()); if(null != query.getStatus()){ term.must(queryBuilder.keyword() // .wildcard() .onField("status") .matching(query.getStatus()).createQuery()); } if(null != query.getExcludeStatus()){ term.must(queryBuilder.keyword() .onField("status") .matching(query.getExcludeStatus()).createQuery()).not(); } if(null != query.getUsername()){ term.must(queryBuilder.keyword() // .wildcard() .onField("owner.username") .ignoreFieldBridge() .matching(query.getUsername()).createQuery()); } luceneQuery =term.createQuery(); if(LOG.isDebugEnabled()){ LOG.debug("create complicated keyword search query: " + luceneQuery.toString()); } } // BooleanQuery FullTextQuery hibernateQuery = fullTextSession.createFullTextQuery( luceneQuery, Topic.class); return hibernateQuery; }