使用Spring Data Elasticsearch按最高得分百分比过滤结果

import org.springframework.data.domain.PageRequest;
import org.springframework.data.domain.Pageable;
import org.springframework.data.elasticsearch.core.ElasticsearchOperations;
import org.springframework.data.elasticsearch.core.SearchHit;
import org.springframework.data.elasticsearch.core.SearchHits;
import org.springframework.data.elasticsearch.core.query.Criteria;
import org.springframework.data.elasticsearch.core.query.CriteriaQuery;
import org.springframework.stereotype.Service;

import java.util.ArrayList;
import java.util.List;
import java.util.Objects;

@Service
public class DocumentSearchService {

    private final ElasticsearchOperations elasticsearchOperations;

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

    /**
     * 根据最高得分的百分比过滤Elasticsearch文档。
     *
     * @param queryText 查询字符串
     * @param scorePercentageThreshold 得分百分比阈值 (例如 0.8 代表 80%)
     * @param maxResults 期望返回的最大结果数
     * @param <T> 文档类型
     * @return 过滤后的文档列表
     */
    public <T> List<T> searchAndFilterByMaxScorePercentage(
            String queryText, double scorePercentageThreshold, int maxResults, Class<T> entityClass) {

        List<T> filteredDocuments = new ArrayList<>();
        int pageSize = Math.min(maxResults, 20); // 初始页面大小，可根据实际情况调整
        int pageNumber = 0;
        double maxScore = -1.0;
        double scoreThreshold = -1.0;
        boolean stopFetching = false;

        while (!stopFetching && filteredDocuments.size() < maxResults) {
            Pageable pageable = PageRequest.of(pageNumber, pageSize);
            CriteriaQuery query = new CriteriaQuery(new Criteria("content").contains(queryText), pageable); // 示例查询

            SearchHits<T> searchHits = elasticsearchOperations.search(query, entityClass);

            if (searchHits.getTotalHits() == 0) {
                break; // 没有结果
            }

            if (maxScore == -1.0) { // 首次获取，确定 maxScore 和阈值
                maxScore = searchHits.getMaxScore();
                if (maxScore <= 0) {
                    // 如果maxScore为0或负数，意味着所有文档得分都非常低，
                    // 此时可以根据业务逻辑决定是返回所有匹配文档，还是返回空列表，
                    // 或者设置一个非常小的正阈值。这里我们直接停止。
                    System.out.println("Max score is zero or negative, stopping filtering.");
                    break;
                }
                scoreThreshold = maxScore * scorePercentageThreshold;
                // 确保阈值不会过低导致不必要的过滤，或过高导致无结果
                if (scoreThreshold < 0.0001 && maxScore > 0) { // 避免因浮点数精度导致阈值过低
                    scoreThreshold = 0.0001;
                }
            }

            for (SearchHit<T> hit : searchHits) {
                if (hit.getScore() >= scoreThreshold) {
                    filteredDocuments.add(hit.getContent());
                    if (filteredDocuments.size() >= maxResults) {
                        stopFetching = true; // 达到最大结果数
                        break;
                    }
                } else {
                    // 遇到第一个得分低于阈值的文档，停止继续获取
                    stopFetching = true;
                    break;
                }
            }

            if (searchHits.hasMore() && !stopFetching) {
                pageNumber++;
            } else {
                stopFetching = true; // 没有更多页或已决定停止
            }
        }
        return filteredDocuments;
    }
}

// 假设 MyDocument 是你的 Elasticsearch 实体类
// @Document(indexName = "my_documents")
// public class MyDocument { ... }

@Autowired
private DocumentSearchService documentSearchService;

public List<MyDocument> getTopRelevantDocuments(String query) {
    // 获取得分在最高得分80%以上，且最多100个结果的文档
    return documentSearchService.searchAndFilterByMaxScorePercentage(query, 0.8, 100, MyDocument.class);
}