向量 Index

向量索引 Index 用于存储和管理向量Bucket中的向量数据。每个 Index 定义了向量的维度、距离度量方法和元数据结构,为向量检索提供基础。

创建向量 Index

单个向量 Bucket 中最多可创建 100 张向量索引。创建向量索引的 API (PutVectorIndex)请求频率限制为最多每秒 5 次。

控制台

  1. 向量Bucket页面,单击已创建的向量Bucket。

  2. 索引列表页面,单击创建索引表

  3. 配置索引参数:

    • 索引表名称:1~63字符,由英文字母、数字组成,首字母必须为英文字母,在向量 Bucket 内全局唯一。

    • 向量数据类型:系统默认float32:浮点型。

    • 向量维度:1~4096维,添加到此索引中的所有向量必须具有完全相同的数值个数。

    • 距离度量函数:根据业务场景选择距离计算方式。

      • 欧氏距离:空间中两点间的直线距离,适用于衡量数值差异。

      • 余弦距离:衡量两个向量在方向上的差异,适用于文本、图像等高维语义相似度计算。

    • 元数据配置:配置非过滤元数据字段,用于存储不参与搜索过滤的附加信息,作为向量数据的描述信息。非过滤元数据配置。有如下输入限制:

      • 元数据个数为 1~10 个

      • 每个元数据主键名称的长度 1~63 字节

  4. 单击确定完成创建。

ossutil

为存储桶 examplebucket 创建 index 索引:向量维度 512,数据类型float32,距离度量 euclidean。

ossutil vectors-api put-vector-index --bucket examplebucket --index-name index --data-type float32 --dimension 512 --distance-metric euclidean

SDK

Python

import argparse
import alibabacloud_oss_v2 as oss
import alibabacloud_oss_v2.vectors as oss_vectors

parser = argparse.ArgumentParser(description="vector put vector index sample")
parser.add_argument('--region', help='The region in which the bucket is located.', required=True)
parser.add_argument('--bucket', help='The name of the bucket.', required=True)
parser.add_argument('--endpoint', help='The domain names that other services can use to access OSS')
parser.add_argument('--index_name', help='The name of the vector index.', required=True)
parser.add_argument('--account_id', help='The account id.', required=True)

def main():
    args = parser.parse_args()

    # Loading credentials values from the environment variables
    credentials_provider = oss.credentials.EnvironmentVariableCredentialsProvider()

    # Using the SDK's default configuration
    cfg = oss.config.load_default()
    cfg.credentials_provider = credentials_provider
    cfg.region = args.region
    cfg.account_id = args.account_id
    if args.endpoint is not None:
        cfg.endpoint = args.endpoint

    vector_client = oss_vectors.Client(cfg)

    result = vector_client.put_vector_index(oss_vectors.models.PutVectorIndexRequest(
        bucket=args.bucket,
        index_name=args.index_name,
        dimension=512,
        data_type='float32',
        distance_metric='euclidean',
        metadata={"nonFilterableMetadataKeys": ["key1", "key2"]}
    ))

    print(f'status code: {result.status_code},'
          f' request id: {result.request_id},'
    )

if __name__ == "__main__":
    main()

Go

package main

import (
	"context"
	"flag"
	"log"

	"github.com/aliyun/alibabacloud-oss-go-sdk-v2/oss"
	"github.com/aliyun/alibabacloud-oss-go-sdk-v2/oss/credentials"
	"github.com/aliyun/alibabacloud-oss-go-sdk-v2/oss/vectors"
)

var (
	region     string
	bucketName string
	accountId  string
	indexName  string
)

func init() {
	flag.StringVar(&region, "region", "", "The region in which the vector bucket is located.")
	flag.StringVar(&bucketName, "bucket", "", "The name of the vector bucket.")
	flag.StringVar(&accountId, "account-id", "", "The id of vector account.")
	flag.StringVar(&indexName, "index", "", "The name of vector index.")
}

func main() {
	flag.Parse()
	if len(bucketName) == 0 {
		flag.PrintDefaults()
		log.Fatalf("invalid parameters, bucket name required")
	}

	if len(region) == 0 {
		flag.PrintDefaults()
		log.Fatalf("invalid parameters, region required")
	}

	if len(accountId) == 0 {
		flag.PrintDefaults()
		log.Fatalf("invalid parameters, accounId required")
	}

	if len(indexName) == 0 {
		flag.PrintDefaults()
		log.Fatalf("invalid parameters, index required")
	}

	cfg := oss.LoadDefaultConfig().
		WithCredentialsProvider(credentials.NewEnvironmentVariableCredentialsProvider()).
		WithRegion(region).WithAccountId(accountId)

	client := vectors.NewVectorsClient(cfg)

	request := &vectors.PutVectorIndexRequest{
		Bucket:         oss.Ptr(bucketName),
		DataType:       oss.Ptr("float32"),
		Dimension:      oss.Ptr(128),
		DistanceMetric: oss.Ptr("cosine"),
		IndexName:      oss.Ptr(indexName),
		Metadata: map[string]any{
			"nonFilterableMetadataKeys": []string{"foo", "bar"},
		},
	}
	result, err := client.PutVectorIndex(context.TODO(), request)
	if err != nil {
		log.Fatalf("failed to put vector index%v", err)
	}
	log.Printf("put vector index result:%#v\n", result)
}

API

调用PutVectorIndex接口以创建向量Index。

获取向量 Index 信息

控制台

向量Bucket页面,单击已创建的向量Bucket,在索引列表页面,可以查看向量 Index 信息。

ossutil

获取向量存储空间examplebucket中向量名字为index的向量索引属性。

ossutil vectors-api get-vector-index --bucket examplebucket --index-name index

SDK

Python

import argparse
import alibabacloud_oss_v2 as oss
import alibabacloud_oss_v2.vectors as oss_vectors

parser = argparse.ArgumentParser(description="vector get vector index sample")
parser.add_argument('--region', help='The region in which the bucket is located.', required=True)
parser.add_argument('--bucket', help='The name of the bucket.', required=True)
parser.add_argument('--endpoint', help='The domain names that other services can use to access OSS')
parser.add_argument('--index_name', help='The name of the vector index.', required=True)
parser.add_argument('--account_id', help='The account id.', required=True)

def main():
    args = parser.parse_args()

    # Loading credentials values from the environment variables
    credentials_provider = oss.credentials.EnvironmentVariableCredentialsProvider()

    # Using the SDK's default configuration
    cfg = oss.config.load_default()
    cfg.credentials_provider = credentials_provider
    cfg.region = args.region
    cfg.account_id = args.account_id
    if args.endpoint is not None:
        cfg.endpoint = args.endpoint

    vector_client = oss_vectors.Client(cfg)

    result = vector_client.get_vector_index(oss_vectors.models.GetVectorIndexRequest(
        bucket=args.bucket,
        index_name=args.index_name,
    ))

    print(f'status code: {result.status_code},'
          f' request id: {result.request_id},'
          )

    if result.index:
        print(f'index name: {result.index}')

if __name__ == "__main__":
    main()

Go

package main

import (
	"context"
	"flag"
	"log"

	"github.com/aliyun/alibabacloud-oss-go-sdk-v2/oss"
	"github.com/aliyun/alibabacloud-oss-go-sdk-v2/oss/credentials"
	"github.com/aliyun/alibabacloud-oss-go-sdk-v2/oss/vectors"
)

var (
	region     string
	bucketName string
	indexName  string
	accountId  string
)

func init() {
	flag.StringVar(&region, "region", "", "The region in which the vector bucket is located.")
	flag.StringVar(&bucketName, "bucket", "", "The name of the vector bucket.")
	flag.StringVar(&indexName, "index", "", "The name of the vector index.")
	flag.StringVar(&accountId, "account-id", "", "The id of vector account.")
}

func main() {
	flag.Parse()
	if len(bucketName) == 0 {
		flag.PrintDefaults()
		log.Fatalf("invalid parameters, bucket name required")
	}

	if len(region) == 0 {
		flag.PrintDefaults()
		log.Fatalf("invalid parameters, region required")
	}

	if len(indexName) == 0 {
		flag.PrintDefaults()
		log.Fatalf("invalid parameters, index required")
	}

	if len(accountId) == 0 {
		flag.PrintDefaults()
		log.Fatalf("invalid parameters, accounId required")
	}

	cfg := oss.LoadDefaultConfig().
		WithCredentialsProvider(credentials.NewEnvironmentVariableCredentialsProvider()).
		WithRegion(region).WithAccountId(accountId)

	client := vectors.NewVectorsClient(cfg)

	request := &vectors.GetVectorIndexRequest{
		Bucket:    oss.Ptr(bucketName),
		IndexName: oss.Ptr(indexName),
	}
	result, err := client.GetVectorIndex(context.TODO(), request)
	if err != nil {
		log.Fatalf("failed to get vector index %v", err)
	}
	log.Printf("get vector index result:%#v\n", result)
}

API

调用GetVectorIndex接口以获取向量Index信息。

列举向量 Index 信息

列举向量索引的 API (ListVectorIndexes)每次最大返回索引个数 500,可分页获取下一批索引列表,并发度上限 16 个。

控制台

向量Bucket页面,单击已创建的向量Bucket,进入索引列表页面,即列举出当前向量Bucket下所有索引。

ossutil

列举名为examplebucket的向量Bucket中所有向量索引。

ossutil vectors-api list-vector-indexes --bucket examplebucket

SDK

Python

import argparse
import alibabacloud_oss_v2 as oss
import alibabacloud_oss_v2.vectors as oss_vectors

parser = argparse.ArgumentParser(description="list vector indexes sample")

parser.add_argument('--region', help='The region in which the bucket is located.', required=True)
parser.add_argument('--endpoint', help='The domain names that other services can use to access OSS')
parser.add_argument('--account_id', help='The account id.', required=True)
parser.add_argument('--bucket', help='The name of the bucket.', required=True)

def main():

    args = parser.parse_args()

    # Loading credentials values from the environment variables
    credentials_provider = oss.credentials.EnvironmentVariableCredentialsProvider()

    # Using the SDK's default configuration
    cfg = oss.config.load_default()
    cfg.credentials_provider = credentials_provider
    cfg.region = args.region
    cfg.account_id = args.account_id
    if args.endpoint is not None:
        cfg.endpoint = args.endpoint

    client = oss_vectors.Client(cfg)

    # Create the Paginator for the ListVectorIndex operation
    paginator = client.list_vector_indexes_paginator()

    # Iterate through the vector index pages
    for page in paginator.iter_page(oss_vectors.models.ListVectorIndexesRequest(
        bucket=args.bucket
        )
    ):
        for o in page.indexes:
            print(f'Index: {o.get("indexName")}, {o.get("dataType")}, {o.get("dimension")}, {o.get("status")}')

if __name__ == "__main__":
    main()

Go

package main

import (
	"context"
	"flag"
	"log"

	"github.com/aliyun/alibabacloud-oss-go-sdk-v2/oss"
	"github.com/aliyun/alibabacloud-oss-go-sdk-v2/oss/credentials"
	"github.com/aliyun/alibabacloud-oss-go-sdk-v2/oss/vectors"
)

var (
	region     string
	bucketName string
	accountId  string
)

func init() {
	flag.StringVar(&region, "region", "", "The region in which the vector bucket is located.")
	flag.StringVar(&bucketName, "bucket", "", "The name of the vector bucket.")
	flag.StringVar(&accountId, "account-id", "", "The id of vector account.")
}

func main() {
	flag.Parse()
	if len(region) == 0 {
		flag.PrintDefaults()
		log.Fatalf("invalid parameters, region required")
	}

	if len(bucketName) == 0 {
		flag.PrintDefaults()
		log.Fatalf("invalid parameters, bucket required")
	}

	if len(accountId) == 0 {
		flag.PrintDefaults()
		log.Fatalf("invalid parameters, accounId required")
	}

	cfg := oss.LoadDefaultConfig().
		WithCredentialsProvider(credentials.NewEnvironmentVariableCredentialsProvider()).
		WithRegion(region).WithAccountId(accountId)

	client := vectors.NewVectorsClient(cfg)

	request := &vectors.ListVectorIndexesRequest{
		Bucket: oss.Ptr(bucketName),
	}
	p := client.NewListVectorIndexesPaginator(request)

	var i int
	log.Println("Vector Indexes:")
	for p.HasNext() {
		i++

		page, err := p.NextPage(context.TODO())
		if err != nil {
			log.Fatalf("failed to get page %v, %v", i, err)
		}

		// Log the objects found
		for _, index := range page.Indexes {
			log.Printf("index:%v, %v, %v, %v\n", oss.ToString(index.IndexName), oss.ToTime(index.CreateTime), oss.ToString(index.DataType), oss.ToString(index.Status))
		}
	}
}

API

调用ListVectorIndexes接口以列举向量Bucket中的所有向量索引。

删除向量 Index

删除索引会同时删除索引中的所有向量数据。删除操作不可逆,请谨慎操作,确保已备份重要数据。

控制台

向量Bucket页面,单击已创建的向量Bucket,进入索引列表页面,选择需要删除的索引完成删除操作。

ossutil

删除向量存储空间examplebucket中名为index的向量索引。

ossutil vectors-api delete-vector-index --bucket examplebucket --index-name index

SDK

Python

import argparse
import alibabacloud_oss_v2 as oss
import alibabacloud_oss_v2.vectors as oss_vectors

parser = argparse.ArgumentParser(description="vector delete vector index sample")
parser.add_argument('--region', help='The region in which the bucket is located.', required=True)
parser.add_argument('--bucket', help='The name of the bucket.', required=True)
parser.add_argument('--endpoint', help='The domain names that other services can use to access OSS')
parser.add_argument('--index_name', help='The name of the vector index.', required=True)
parser.add_argument('--account_id', help='The account id.', required=True)

def main():
    args = parser.parse_args()

    # Loading credentials values from the environment variables
    credentials_provider = oss.credentials.EnvironmentVariableCredentialsProvider()

    # Using the SDK's default configuration
    cfg = oss.config.load_default()
    cfg.credentials_provider = credentials_provider
    cfg.region = args.region
    cfg.account_id = args.account_id
    if args.endpoint is not None:
        cfg.endpoint = args.endpoint

    vector_client = oss_vectors.Client(cfg)

    result = vector_client.delete_vector_index(oss_vectors.models.DeleteVectorIndexRequest(
        bucket=args.bucket,
        index_name=args.index_name,
    ))

    print(f'status code: {result.status_code},'
          f' request id: {result.request_id},'
    )

if __name__ == "__main__":
    main()

Go

package main

import (
	"context"
	"flag"
	"log"

	"github.com/aliyun/alibabacloud-oss-go-sdk-v2/oss"
	"github.com/aliyun/alibabacloud-oss-go-sdk-v2/oss/credentials"
	"github.com/aliyun/alibabacloud-oss-go-sdk-v2/oss/vectors"
)

var (
	region     string
	bucketName string
	indexName  string
	accountId  string
)

func init() {
	flag.StringVar(&region, "region", "", "The region in which the vector bucket is located.")
	flag.StringVar(&bucketName, "bucket", "", "The name of the vector bucket.")
	flag.StringVar(&indexName, "index", "", "The name of the vector index.")
	flag.StringVar(&accountId, "account-id", "", "The id of vector account.")
}

func main() {
	flag.Parse()
	if len(bucketName) == 0 {
		flag.PrintDefaults()
		log.Fatalf("invalid parameters, bucket name required")
	}

	if len(region) == 0 {
		flag.PrintDefaults()
		log.Fatalf("invalid parameters, region required")
	}

	if len(indexName) == 0 {
		flag.PrintDefaults()
		log.Fatalf("invalid parameters, index required")
	}

	if len(accountId) == 0 {
		flag.PrintDefaults()
		log.Fatalf("invalid parameters, accounId required")
	}

	cfg := oss.LoadDefaultConfig().
		WithCredentialsProvider(credentials.NewEnvironmentVariableCredentialsProvider()).
		WithRegion(region).WithAccountId(accountId)

	client := vectors.NewVectorsClient(cfg)

	request := &vectors.DeleteVectorIndexRequest{
		Bucket:    oss.Ptr(bucketName),
		IndexName: oss.Ptr(indexName),
	}
	result, err := client.DeleteVectorIndex(context.TODO(), request)
	if err != nil {
		log.Fatalf("failed to delete vector index %v", err)
	}

	log.Printf("delete vector index result:%#v\n", result)
}

API

调用DeleteVectorIndex接口以删除向量索引。