向量索引 Index 用於儲存和管理向量Bucket中的向量資料。每個 Index 定義了向量的維度、距離度量方法和中繼資料結構,為向量檢索提供基礎。
建立向量 Index
單個向量 Bucket 中最多可建立 100 張向量索引,若需調大該配額,可以聯絡支援人員申請調整。
建立向量索引的 API (PutVectorIndex)請求頻率限制為最多每秒 5 次。
控制台
在向量Bucket頁面,單擊已建立的向量Bucket。
在索引列表頁面,單擊建立索引表。
配置索引參數:
索引表名稱:1~63字元,由英文字母、數字組成,首字母必須為英文字母,在向量 Bucket 內全域唯一。
向量資料類型:系統預設float32:浮點型。
向量維度:1~4096維,添加到此索引中的所有向量必須具有完全相同的數值個數。
距離度量函數:根據業務情境選擇距離計算方式。
歐氏距離:空間中兩點間的直線距離,適用于衡量數值差異。
餘弦距離:衡量兩個向量在方向上的差異,適用於文本、映像等高維語義相似性計算。
中繼資料配置:配置非過濾中繼資料欄位,用於儲存不參與搜尋過濾的附加資訊,作為向量資料的描述資訊。非過濾中繼資料配置。有如下輸入限制:
中繼資料個數為 1~10 個
每個中繼資料主鍵名稱的長度 1~63 位元組
單擊確定完成建立。
ossutil
為儲存桶 examplebucket 建立 index 索引:向量維度 512,資料類型float32,距離度量 euclidean。
ossutil vectors-api put-vector-index --bucket examplebucket --index-name index --data-type float32 --dimension 512 --distance-metric euclideanSDK
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(®ion, "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
調用PutVectorIndex介面以建立向量Index。
擷取向量 Index 資訊
控制台
在向量Bucket頁面,單擊已建立的向量Bucket,在索引列表頁面,可以查看向量 Index 資訊。
ossutil
擷取向量儲存空間examplebucket中向量名字為index的向量索引屬性。
ossutil vectors-api get-vector-index --bucket examplebucket --index-name indexSDK
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(®ion, "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
調用GetVectorIndex介面以擷取向量Index資訊。
列舉向量 Index 資訊
列舉向量索引的 API (ListVectorIndexes)每次最大返回索引個數 500,可分頁擷取下一批索引列表,並發度上限 16 個。
控制台
在向量Bucket頁面,單擊已建立的向量Bucket,進入索引列表頁面,即列舉出當前向量Bucket下所有索引。
ossutil
列舉名為examplebucket的向量Bucket中所有向量索引。
ossutil vectors-api list-vector-indexes --bucket examplebucketSDK
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(®ion, "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
調用ListVectorIndexes介面以列舉向量Bucket中的所有向量索引。
刪除向量 Index
刪除索引會同時刪除索引中的所有向量資料。刪除操作無法復原,請謹慎操作,確保已備份重要資料。
控制台
在向量Bucket頁面,單擊已建立的向量Bucket,進入索引列表頁面,選擇需要刪除的索引完成刪除操作。
ossutil
刪除向量儲存空間examplebucket中名為index的向量索引。
ossutil vectors-api delete-vector-index --bucket examplebucket --index-name indexSDK
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(®ion, "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, accountId 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
調用DeleteVectorIndex介面以刪除向量索引。