OpenSearch:Sparse vector-based query

URL

/vector-service/query

• The sample URL omits information such as parameters in request headers and the encoding method.

• The sample URL also omits the endpoint that is used to connect to an OpenSearch instance.

• For more information about the definitions, usage, and example values of all the request parameters that are concatenated in the preceding URL, see the "Request parameters" section of this topic.

Protocol

HTTP

Request method

POST

Supported format

JSON

Request signature

You can calculate the request signature by using the parameters described in the following table. The request signature is stored in the authorization header.

import com.aliyun.darabonba.encode.Encoder;
import com.aliyun.darabonbastring.Client;

public class GenerateAuthorization {

    public static void main(String[] args) throws Exception {
        String accessUserName = "username";
        String accessPassWord = "password";
        String realmStr = "" + accessUserName + ":" + accessPassWord + "";
        String authorization = Encoder.base64EncodeToString(Client.toBytes(realmStr, "UTF-8"));
        System.out.println(authorization);
    }
}

Valid format for the value of the authorization header:

cm9vdDp******mdhbA==

You must add the Basic prefix when you specify the authorization header in an HTTP request.

Example:

authorization: Basic cm9vdDp******mdhbA==

Request parameters

• SearchRequest

• SparseData

In OpenSearch Vector Search Edition, two independent multi-value fields are used to represent sparse vector subscripts and values. The number of values must be the same in the two fields. The positions of the values in the two fields have a one-to-one correspondence.

For example, the sparse vector (0, 0.5), (100, 0.9), (40, 0.3), (50, 0.7), (20, 0.6) exists. The following code shows the multi-value fields of the sparse vector:

{
  "indices": [0, 100, 40, 50, 20],
  "values": [0.5, 0.9, 0.3, 0.7, 0.6]
}

When you write or query data, subscripts must be arranged in ascending order, and values need to be adjusted based on the subscripts. The following code shows the adjusted multi-value fields:

{
  "indices": [0, 20, 40, 50, 100],
  "values": [0.5, 0.6, 0.3, 0.7, 0.9]
}

Query weight

In a hybrid query, the final score of the same document is the sum of the distance of the dense vector and the distance of the sparse vector. If you want to configure different weights for the sparse vector and dense vector, you can perform configurations based on the following code:

{
    "vector": [v * weight for v in dense_vector],
    "sparseData": {
        "indices": sparse_data["indices"],
        "values": [v * (1 - weight) for v in sparse_data["values"]]
    }
}

Sample request

{
    "tableName": "in0",
    "indexName": "vector",
    "vector": [
        0.1,
        0.2,
        0.3,
        0.4,
        0.5
    ],
    "sparseData": {
        "indices": [
            0,
            2
        ],
        "values": [
            1.2,
            2.4
        ]
    },
    "topK": 2,
    "order": "DESC"
}

Response parameters

Item

Example

{
    "totalCount": 50,
    "result": [
        {
            "id": 99,
            "score": 0.18588095903396607,
            "__source__": 1,
            "fields": {
                "content": "Zhonghao Cat 6 Network Cable RJ45 Cat 6 Gigabit 8-Core Twisted Pair Jumper Broadband Monitoring Home Network Cable Finished Cable"
            }
        }
    ],
    "totalTime": 4.057
}