使用Spring Data Elasticsearch Repository时如何指定max_analyzed_offset？

由于@HighlightParameters注解未提供max_analyzed_offset的配置属性，你可以通过自定义Repository实现或直接使用ElasticsearchOperations构建编程式查询来解决这个问题，无需修改索引的全局highlight.max_analyzed_offset配置。

方法一：自定义Repository实现 #

这种方式保持Repository接口的调用习惯，同时实现灵活的查询配置：

1. 定义自定义Repository接口

public interface CustomEsRepository {
    List<SearchHit<SearchResult>> queryAllFieldsByQueryStringHighlighting(String queryTerm, Pageable pageable);
}

2. 修改原有Repository接口，继承自定义接口

@Repository
public interface EsRepository extends ElasticsearchRepository<SearchResult, String>, CustomEsRepository {
}

3. 实现自定义Repository逻辑

public class EsRepositoryImpl implements CustomEsRepository {

    private final ElasticsearchOperations elasticsearchOperations;

    // 通过构造函数注入ElasticsearchOperations
    public EsRepositoryImpl(ElasticsearchOperations elasticsearchOperations) {
        this.elasticsearchOperations = elasticsearchOperations;
    }

    @Override
    public List<SearchHit<SearchResult>> queryAllFieldsByQueryStringHighlighting(String queryTerm, Pageable pageable) {
        // 构建原生查询，直接使用Elasticsearch原生HighlightBuilder配置高亮参数
        NativeSearchQuery searchQuery = new NativeSearchQueryBuilder()
                .withQuery(QueryBuilders.queryStringQuery(queryTerm))
                .withPageable(pageable)
                .withHighlightBuilder(new HighlightBuilder()
                        .field("attachment.content")
                        .preTags("<strong>")
                        .postTags("</strong>")
                        .fragmentSize(300)
                        .numberOfFragments(1)
                        // 直接设置max_analyzed_offset参数
                        .setMaxAnalyzedOffset(2000000)) // 根据需求调整数值
                .build();

        // 执行查询并返回带高亮的结果
        return elasticsearchOperations.search(searchQuery, SearchResult.class).getSearchHits();
    }
}

方法二：在业务层直接使用ElasticsearchOperations #

如果不需要封装到Repository，也可以在Service中直接注入ElasticsearchOperations构建查询：

@Service
public class SearchService {

    private final ElasticsearchOperations elasticsearchOperations;

    public SearchService(ElasticsearchOperations elasticsearchOperations) {
        this.elasticsearchOperations = elasticsearchOperations;
    }

    public List<SearchHit<SearchResult>> searchWithHighlight(String queryTerm, Pageable pageable) {
        NativeSearchQuery searchQuery = new NativeSearchQueryBuilder()
                .withQuery(QueryBuilders.queryStringQuery(queryTerm))
                .withPageable(pageable)
                .withHighlightBuilder(new HighlightBuilder()
                        .field("attachment.content")
                        .preTags("<strong>")
                        .postTags("</strong>")
                        .fragmentSize(300)
                        .numberOfFragments(1)
                        .setMaxAnalyzedOffset(2000000))
                .build();

        return elasticsearchOperations.search(searchQuery, SearchResult.class).getSearchHits();
    }
}

原理说明 #

通过使用Elasticsearch原生的HighlightBuilder，我们可以直接访问所有高亮相关参数（包括max_analyzed_offset），绕开Spring Data Elasticsearch注解的限制，实现查询级别的参数配置。

内容的提问来源于stack exchange，提问作者coder764323