基于SpringBoot+ElasticSearch实现文档智能化检索的完整指南

<!-- pom.xml -->
<dependencies>
    <!-- SpringBoot基础 -->
    <dependency>
        <groupId>org.springframework.boot</groupId>
        <artifactId>spring-boot-starter-web</artifactId>
    </dependency>

    <!-- Elasticsearch连接 -->
    <dependency>
        <groupId>org.springframework.boot</groupId>
        <artifactId>spring-boot-starter-data-elasticsearch</artifactId>
    </dependency>

    <!-- 文件处理工具 -->
    <dependency>
        <groupId>org.apache.poi</groupId>
        <artifactId>poi-ooxml</artifactId>
        <version>5.2.3</version>
    </dependency>

    <!-- Jieba分词插件 -->
    <dependency>
        <groupId>com.nlp</groupId>
        <artifactId>elasticsearch-analysis-jieba</artifactId>
        <version>7.17.0</version>
    </dependency>
</dependencies>

# application.yml
spring:
  data:
    elasticsearch:
      cluster-name: my-cluster
      cluster-nodes: localhost:9200
  elasticsearch:
    rest:
      uris: http://localhost:9200
      username: elastic
      password: your_password

# 安装插件（本地ES）
elasticsearch-plugin install ingest-attachment

# 安装插件（Docker容器内）
docker exec -it elasticsearch bin/elasticsearch-plugin install ingest-attachment

elasticsearch-plugin install https://github.com/medcl/elasticsearch-analysis-jieba/releases/download/v7.17.0/elasticsearch-analysis-jieba-7.17.0.zip

PUT _ingest/pipeline/attachment-extract
{
  "description": "Extract attachment content",
  "processors": [
    {
      "attachment": {
        "field": "content",
        "target_field": "attachment",
        "ignore_missing": true
      }
    },
    {
      "remove": {
        "field": "content"
      }
    }
  ]
}

PUT /fileinfo
{
  "mappings": {
    "properties": {
      "id": { "type": "keyword" },
      "fileName": { "type": "text" },
      "fileType": { "type": "keyword" },
      "attachment": {
        "properties": {
          "content": { "type": "text", "analyzer": "jieba" }  // 使用Jieba分词
        }
      }
    }
  },
  "settings": {
    "analysis": {
      "analyzer": {
        "jieba": {
          "type": "custom",
          "tokenizer": "jieba_tokenizer"
        }
      }
    }
  }
}

@RestController
@RequestMapping("/api/files")
public class FileUploadController {

    @Autowired
    private ElasticsearchRestTemplate elasticsearchRestTemplate;

    @PostMapping("/upload")
    public ResponseEntity<String> uploadFile(@RequestParam("file") MultipartFile file) throws IOException {
        // 1. 文件转Base64
        String base64Content = Base64.getEncoder().encodeToString(file.getBytes());

        // 2. 构造文档对象
        Map<String, Object> document = new HashMap<>();
        document.put("id", UUID.randomUUID().toString());
        document.put("fileName", file.getOriginalFilename());
        document.put("fileType", getFileType(file.getOriginalFilename()));
        document.put("content", base64Content);  // 二进制字段

        // 3. 使用Pipeline处理并索引文档
        IndexRequest request = new IndexRequest("fileinfo")
                .setId(document.get("id").toString())
                .setPipeline("attachment-extract")  // 关键：绑定Pipeline
                .setSource(document);

        elasticsearchRestTemplate.index(request);

        return ResponseEntity.ok("文件已成功索引");
    }

    private String getContentType(MultipartFile file) {
        String originalFilename = file.getOriginalFilename();
        if (originalFilename.endsWith(".pdf")) {
            return "application/pdf";
        } else if (originalFilename.endsWith(".docx")) {
            return "application/vnd.openxmlformats-officedocument.wordprocessingml.document";
        }
        return "application/octet-stream";
    }
}

@GetMapping("/search")
public ResponseEntity<Map<String, Object>> searchFiles(@RequestParam String keyword) {
    // 1. 构建查询
    SearchSourceBuilder sourceBuilder = new SearchSourceBuilder();
    sourceBuilder.query(QueryBuilders.matchQuery("attachment.content", keyword)
            .analyzer("jieba")  // 使用Jieba分词
            .fuzziness("AUTO"));

    // 2. 启用高亮
    HighlightBuilder highlightBuilder = new HighlightBuilder();
    highlightBuilder.field("attachment.content").preTags("<mark>").postTags("</mark>");
    sourceBuilder.highlighter(highlightBuilder);

    // 3. 执行搜索
    SearchRequest searchRequest = new SearchRequest("fileinfo");
    searchRequest.source(sourceBuilder);
    SearchResponse response = elasticsearchRestTemplate.search(searchRequest);

    // 4. 提取高亮结果
    List<Map<String, Object>> results = new ArrayList<>();
    for (SearchHit hit : response.getHits().getHits()) {
        Map<String, Object> source = hit.getSourceAsMap();
        Map<String, HighlightField> highlights = hit.getHighlightFields();
        HighlightField contentHighlight = highlights.get("attachment.content");
        if (contentHighlight != null) {
            source.put("highlight", contentHighlight.fragments()[0].string());
        }
        results.add(source);
    }

    return ResponseEntity.ok(Collections.singletonMap("results", results));
}

// 分页示例
sourceBuilder.from(0).size(10);  // 限制每页10条
sourceBuilder.sort(SortBuilders.fieldSort("createTime").order(SortOrder.DESC));  // 按时间排序

@Configuration
public class ElasticsearchConfig {

    @Value("${spring.elasticsearch.rest.uris}")
    private String esUri;

    @Bean
    public RestHighLevelClient elasticsearchClient() {
        return new RestHighLevelClient(
                RestClient.builder(new HttpHost(esUri.split(":")[0], Integer.parseInt(esUri.split(":")[1]), "http")));
    }

    @Bean
    public ElasticsearchRestTemplate elasticsearchRestTemplate(RestHighLevelClient client) {
        return new ElasticsearchRestTemplate(client);
    }
}

{
  "results": [
    {
      "id": "123",
      "fileName": "进口红酒.pdf",
      "fileType": "pdf",
      "attachment": {
        "content": "这款红酒产自法国波尔多地区，口感醇厚..."
      },
      "highlight": "这款红酒产自法国波尔多地区，<mark>口感醇厚</mark>..."
    }
  ]
}