java - Elasticsearch Java 中的 MultiTermVectors

Question

我正在使用以下函数来获取某些 ID 集的术语 vector 。

public static void builtTermVectorRequest(Client client, String index, Map<String, String> postIDs) {
    TermVectorsRequest termVectorsRequest = new TermVectorsRequest();
    termVectorsRequest.index(index).type("post");
    for (Map.Entry<String, String> entry : postIDs.entrySet()) {
      String currentPostId = entry.getKey();
      String currentParentID = entry.getValue();
      termVectorsRequest
              .id(currentPostId)
              .parent(currentParentID)
              .termStatistics(true)
              .selectedFields("content");
    }

    MultiTermVectorsRequestBuilder mtbuilder = client.prepareMultiTermVectors();
    mtbuilder.add(termVectorsRequest);

    MultiTermVectorsResponse response = mtbuilder.execute().actionGet();
    XContentBuilder builder;
    try {
      builder = XContentFactory.jsonBuilder().startObject();
      response.toXContent(builder, ToXContent.EMPTY_PARAMS);
      builder.endObject();
      System.out.println(builder.prettyPrint().string());
    } catch (IOException e) {}
  }

这里我有一些文档 ID 及其父 ID，因为这些文档是子文档。

我发现这些文档即使存在也没有找到。

为了确认我在 Python 中尝试了同样的事情，使用:

body = dict(docs=map(lambda x:
                     {
                         "fields": ["content"],
                         "_id": x["_id"],
                         "_routing": x["_routing"],
                         "term_statistics": "true"
                     }, result["hits"]["hits"]))

es_client = elasticsearch.Elasticsearch([{'host': '192.168.111.12', 'port': 9200}])

all_term_vectors = es_client.mtermvectors(
    index="prf_test",
    doc_type="post",
    body=body
)

我得到了结果。

Java 代码有什么问题？

Answer 1

我尝试了更多关于如何将 TermVectorsRequest 与 MultiTermVectorsRequestBuilder 结合使用的组合，最终得出以下有效的解决方案:

/**
 * Prints term-vectors for child documents given their parent ids
 *
 * @param client    Es client
 * @param index     Index name
 * @param postIDs   Map of child document ID to its _parent/_routing ID
 */
public static void builtTermVectorRequest(Client client, String index, Map<String, String> postIDs) {
  /**
   * Initialize the MultiTermVectorsRequestBuilder first
   */
  MultiTermVectorsRequestBuilder multiTermVectorsRequestBuilder = client.prepareMultiTermVectors();

  /**
   * For every document ID, create a different TermVectorsRequest and 
   * add it to the MultiTermVectorsRequestBuilder created above
   */
  for (Map.Entry<String, String> entry : postIDs.entrySet()) {
    String currentPostId = entry.getKey();
    String currentRoutingID = entry.getValue();
    TermVectorsRequest termVectorsRequest = new TermVectorsRequest()
            .index(index)
            .type("doc_type")
            .id(currentPostId)
            .parent(currentRoutingID) // You can use .routing(currentRoutingID) also
            .selectedFields("some_field")
            .termStatistics(true);
    multiTermVectorsRequestBuilder.add(termVectorsRequest);
  }

  /**
   * Finally execute the MultiTermVectorsRequestBuilder
   */
  MultiTermVectorsResponse response = multiTermVectorsRequestBuilder.execute().actionGet();

  XContentBuilder builder;
  try {
    builder = XContentFactory.jsonBuilder().startObject();
    response.toXContent(builder, ToXContent.EMPTY_PARAMS);
    builder.endObject();
    System.out.println(builder.prettyPrint().string());
  } catch (IOException e) {
  }
}