vector: S3 Vectors API gateway (P1) (#12103)

* vector: S3 Vectors API gateway over VectorService

weed/s3api/s3vectors serves the AWS S3 Vectors JSON API: bucket CRUD
maps to S3 Tables table buckets, index CRUD to Lance table records plus
the VectorService CreateIndex/DropIndex RPCs, and the data plane
(Put/Get/List/Delete/QueryVectors) calls the shared service. The AWS
filter document translates to a Lance SQL predicate over the metadata
JSON column; unsupported operators fail loudly rather than overmatch.

Generated with [Devin](https://devin.ai)

Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>

* vector: -port.vectors and -vectors.grpc on weed s3

Mirror the Lance server startup: the S3 Vectors API binds its own port
and dials the weed-vector-server gRPC address; both disabled by default.

Generated with [Devin](https://devin.ai)

Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>

* vector: resolve index ARNs, authorize index ops, fix catalog fields

- vector buckets declare FormatLance so a new bucket accepts index
  tables instead of defaulting to ICEBERG.
- ListIndexes decodes the `tables` response field; `tableSummaries`
  hid every index.
- resolveIndexARNs honors indexArn alone; handlers no longer treat
  indexName as the only carrier.
- data-plane RPCs fetch the index record first, so callers authorize
  against the index resource and a missing index 404s before touching
  the engine; QueryVectors reports the index's configured metric.
- IndexArn added to the index-level request types.

Generated with [Devin](https://devin.ai)

Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>

* vector: serve the vectors port over TLS when configured

When -key.file/-cert.file are set, the vectors listener mirrors the
main API and serves TLS rather than a second plaintext endpoint.

Generated with [Devin](https://devin.ai)

Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>

* vector: harden the s3vectors gateway handlers

- $ne no longer shadows the operator: exclusion filters emit <>
  instead of equality.
- a null element in a PutVectors list fails the request instead of
  panicking.
- requests are capped at 32 MiB; distances return via *float32 so an
  exact match still reports 0 when returnDistance is set.
- CreateIndex declares the `vector` namespace first and validates
  dimension (1..=4096) before the catalog write; a failed
  VectorService.CreateIndex rolls the table record back.
- non-filterable metadata keys persist as JSON rather than a
  comma-joined string, and index records carry their createdAt.
- ListVectors honors segmentIndex/segmentCount by slicing the full
  listing in the gateway.
- catalog errors map to AccessDeniedException/ConflictException instead
  of a blanket ValidationException, and malformed ARNs 400 instead of
  slicing.

Generated with [Devin](https://devin.ai)

Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>

* vector: /{Operation} routes and vectors-port TLS, conn cleanup

Path-dispatched clients (POST /QueryVectors) reach the same handlers as
X-Amz-Target calls. A failed listener bind now closes the vector gRPC
connection instead of leaving it open with the endpoint disabled.

Generated with [Devin](https://devin.ai)

Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>

---------

Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
This commit is contained in:
Chris LuandDevin authored and GitHub committed 2026-10-26 14:57:44 -07:00
1 parent 74abcb630e
commit dcc94f818d
4 files changed
+1279

No files matched your search

+94
View File
@@ -22,17 +22,20 @@ import (
"github.com/seaweedfs/seaweedfs/weed/pb/filer_pb"
"github.com/seaweedfs/seaweedfs/weed/pb/s3_lifecycle_pb"
"github.com/seaweedfs/seaweedfs/weed/pb/s3_pb"
"github.com/seaweedfs/seaweedfs/weed/pb/vector_pb"
"github.com/seaweedfs/seaweedfs/weed/s3api"
"github.com/seaweedfs/seaweedfs/weed/s3api/iceberg"
"github.com/seaweedfs/seaweedfs/weed/s3api/lance"
"github.com/seaweedfs/seaweedfs/weed/s3api/s3err"
"github.com/seaweedfs/seaweedfs/weed/s3api/s3tables"
"github.com/seaweedfs/seaweedfs/weed/s3api/s3vectors"
"github.com/seaweedfs/seaweedfs/weed/security"
stats_collect "github.com/seaweedfs/seaweedfs/weed/stats"
"github.com/seaweedfs/seaweedfs/weed/util"
"github.com/seaweedfs/seaweedfs/weed/util/grace"
util_http "github.com/seaweedfs/seaweedfs/weed/util/http"
"github.com/seaweedfs/seaweedfs/weed/util/version"
"google.golang.org/grpc"
)
var (
@@ -52,6 +55,8 @@ type S3Options struct {
portGrpc *int
portIceberg *int
portLance *int
portVectors *int
vectorsGrpc *string
icebergCredentialRole *string
icebergCredentialDuration *int
config *string
@@ -110,6 +115,8 @@ func init() {
s3StandaloneOptions.portGrpc = cmdS3.Flag.Int("port.grpc", 0, "s3 server grpc listen port")
s3StandaloneOptions.portIceberg = cmdS3.Flag.Int("port.iceberg", 8181, "Iceberg REST Catalog server listen port (0 to disable)")
s3StandaloneOptions.portLance = cmdS3.Flag.Int("port.lance", 9101, "Lance Namespace server listen port (0 to disable); credential vending uses -iceberg.credentialRole")
s3StandaloneOptions.portVectors = cmdS3.Flag.Int("port.vectors", 0, "S3 Vectors API server listen port (0 to disable); requires -vectors.grpc")
s3StandaloneOptions.vectorsGrpc = cmdS3.Flag.String("vectors.grpc", "", "weed-vector-server gRPC address for the S3 Vectors API")
s3StandaloneOptions.icebergCredentialRole = cmdS3.Flag.String("iceberg.credentialRole", "", "IAM role ARN the Iceberg catalog assumes to vend table-scoped credentials (empty disables vending)")
s3StandaloneOptions.icebergCredentialDuration = cmdS3.Flag.Int("iceberg.credentialDurationSeconds", 3600, "lifetime of credentials vended by the Iceberg catalog")
s3StandaloneOptions.domainName = cmdS3.Flag.String("domainName", "", "suffix of the host name in comma separated list, {bucket}.{domainName}")
@@ -433,6 +440,13 @@ func (s3opt *S3Options) startS3Server() bool {
}
}
// Start S3 Vectors API server if enabled
if s3opt.portVectors != nil && *s3opt.portVectors > 0 {
if httpS := s3opt.startVectorsServer(s3ApiServer, grpcDialOption, listenOnLocalhost); httpS != nil {
httpServers = append(httpServers, httpS)
}
}
if runtime.GOOS != "windows" {
localSocket := *s3opt.localSocket
explicit := localSocket != ""
@@ -692,6 +706,86 @@ func (s3opt *S3Options) startLanceServer(s3ApiServer *s3api.S3ApiServer, listenO
return httpS
}
// startVectorsServer starts the S3 Vectors API server on a separate port. It
// fans index and data calls out to the shared VectorService over gRPC; with no
// -vectors.grpc the API is not served.
func (s3opt *S3Options) startVectorsServer(s3ApiServer *s3api.S3ApiServer, grpcDialOption grpc.DialOption, listenOnLocalhost bool) *http.Server {
if *s3opt.vectorsGrpc == "" {
glog.Warningf("port.vectors set without -vectors.grpc; S3 Vectors API disabled")
return nil
}
conn, err := pb.GrpcDial(context.Background(), *s3opt.vectorsGrpc, false, grpcDialOption)
if err != nil {
glog.Errorf("S3 Vectors: dial %s: %v", *s3opt.vectorsGrpc, err)
return nil
}
vectorsRouter := mux.NewRouter().SkipClean(true)
vectorsServer := s3vectors.NewServer(s3ApiServer, s3ApiServer, vector_pb.NewVectorServiceClient(conn))
vectorsServer.SetRegion(s3tables.DefaultRegion)
vectorsServer.SetAccountID(s3tables.DefaultAccountID)
vectorsServer.RegisterRoutes(vectorsRouter)
listenAddress := fmt.Sprintf("%s:%d", *s3opt.bindIp, *s3opt.portVectors)
vectorsListener, vectorsLocalListener, err := util.NewIpAndLocalListeners(
*s3opt.bindIp, *s3opt.portVectors, time.Duration(*s3opt.idleTimeout)*time.Second, listenOnLocalhost)
if err != nil {
glog.Errorf("S3 Vectors listener on %s error: %v", listenAddress, err)
conn.Close()
return nil
}
// Mirror the main API: when TLS is configured the vectors port serves
// HTTPS, not a second plaintext endpoint.
if *s3opt.tlsPrivateKey != "" {
getCert, _, err := security.NewReloadingServerCertificate(*s3opt.tlsCertificate, *s3opt.tlsPrivateKey)
if err != nil {
glog.Errorf("S3 Vectors: load HTTPS certificate: %v", err)
return nil
}
clientAuth := tls.NoClientCert
if *s3opt.tlsVerifyClientCert {
clientAuth = tls.RequireAndVerifyClientCert
}
tlsConfig := &tls.Config{GetCertificate: getCert, ClientAuth: clientAuth}
if err := security.FixTlsConfig(util.GetViper(), tlsConfig); err != nil {
glog.Errorf("S3 Vectors: tls config: %v", err)
return nil
}
httpS := newHttpServer(vectorsRouter, tlsConfig)
if vectorsLocalListener != nil {
go func() {
if err := httpS.ServeTLS(vectorsLocalListener, "", ""); err != nil && err != http.ErrServerClosed {
glog.V(0).Infof("S3 Vectors localhost listener error: %v", err)
}
}()
}
go func() {
if err := httpS.ServeTLS(vectorsListener, "", ""); err != nil && err != http.ErrServerClosed {
glog.Errorf("S3 Vectors API Server Fail to serve: %v", err)
}
}()
return httpS
}
glog.V(0).Infof("Start S3 Vectors API Server at http://%s", listenAddress)
httpS := newHttpServer(vectorsRouter, nil)
if vectorsLocalListener != nil {
go func() {
if err := httpS.Serve(vectorsLocalListener); err != nil && err != http.ErrServerClosed {
glog.V(0).Infof("S3 Vectors localhost listener error: %v", err)
}
}()
}
go func() {
if err := httpS.Serve(vectorsListener); err != nil && err != http.ErrServerClosed {
glog.Errorf("S3 Vectors API Server Fail to serve: %v", err)
}
}()
return httpS
}
// deriveLanceStorageEndpoint picks the endpoint the Lance namespace puts in
// storage_options. It falls back to the advertised -ip where the Iceberg
// derivation gives up, because the two clients are not in the same position: a
+840
View File
@@ -0,0 +1,840 @@
package s3vectors
import (
"context"
"encoding/json"
"fmt"
"net/http"
"strings"
"time"
"github.com/seaweedfs/seaweedfs/weed/pb/vector_pb"
"github.com/seaweedfs/seaweedfs/weed/s3api/s3tables"
"google.golang.org/protobuf/types/known/structpb"
)
var handlers = map[string]func(*Server, http.ResponseWriter, *http.Request){
"CreateVectorBucket": (*Server).handleCreateVectorBucket,
"GetVectorBucket": (*Server).handleGetVectorBucket,
"ListVectorBuckets": (*Server).handleListVectorBuckets,
"DeleteVectorBucket": (*Server).handleDeleteVectorBucket,
"CreateIndex": (*Server).handleCreateIndex,
"GetIndex": (*Server).handleGetIndex,
"ListIndexes": (*Server).handleListIndexes,
"DeleteIndex": (*Server).handleDeleteIndex,
"PutVectors": (*Server).handlePutVectors,
"GetVectors": (*Server).handleGetVectors,
"ListVectors": (*Server).handleListVectors,
"DeleteVectors": (*Server).handleDeleteVectors,
"QueryVectors": (*Server).handleQueryVectors,
}
// lance table ids put indexes under a "vector" namespace inside their bucket.
const indexNamespace = "vector"
// Vector bucket records are s3tables table buckets; the S3 Vectors ARN form is
// arn:aws:s3vectors:{region}:{account}:bucket/{name}[/index/{name}].
func (s *Server) vectorBucketArn(bucket string) string {
return fmt.Sprintf("arn:aws:s3vectors:%s:%s:bucket/%s", s.region, s.accountID, bucket)
}
func (s *Server) indexArn(bucket, index string) string {
return fmt.Sprintf("%s/index/%s", s.vectorBucketArn(bucket), index)
}
// s3tablesArn renders the same bucket in the table-bucket ARN form the S3
// Tables catalog keys records by.
func (s *Server) s3tablesArn(bucket string) string {
return fmt.Sprintf("arn:aws:s3tables:%s:%s:bucket/%s", s.region, s.accountID, bucket)
}
func arnParts(v *vectorBucketARNs, name, arn string) error {
if arn == "" {
v.bucket = name
return nil
}
if !strings.HasPrefix(arn, "arn:") {
v.bucket = arn
return nil
}
i := strings.Index(arn, "bucket/")
if i < 0 {
return fmt.Errorf("malformed ARN %q", arn)
}
rest := arn[i+len("bucket/"):]
parts := strings.Split(rest, "/index/")
v.bucket = parts[0]
if len(parts) == 2 {
v.index = parts[1]
}
return nil
}
// resolveIndexARNs fills bucket and index from any combination of plain names,
// a bucket ARN, and a full index ARN — callers may supply only an indexArn.
func resolveIndexARNs(bucketName, bucketArn, indexName, indexArn string) (vectorBucketARNs, error) {
var v vectorBucketARNs
if strings.HasPrefix(indexArn, "arn:") {
if err := arnParts(&v, "", indexArn); err != nil {
return v, err
}
} else if indexArn != "" {
return v, fmt.Errorf("indexArn must be an ARN, got %q", indexArn)
}
if strings.HasPrefix(bucketArn, "arn:") {
if err := arnParts(&v, "", bucketArn); err != nil {
return v, err
}
}
if v.bucket == "" {
if bucketArn != "" {
v.bucket = bucketArn
} else {
v.bucket = bucketName
}
}
if v.index == "" {
v.index = indexName
}
if v.bucket == "" || v.index == "" {
return v, fmt.Errorf("vectorBucketName and indexName (or indexArn) are required")
}
return v, nil
}
// indexRecord fetches the index's catalog record; the call also authorizes
// the request identity against the index resource.
func (s *Server) indexRecord(ctx context.Context, r *http.Request, bucket, index string) (map[string]interface{}, error) {
var resp map[string]interface{}
err := s.catalog(ctx, "GetTable", &s3tables.GetTableRequest{
TableBucketARN: s.s3tablesArn(bucket),
Namespace: []string{indexNamespace},
Name: index,
}, &resp, r)
return resp, err
}
func (s *Server) catalog(ctx context.Context, operation string, req, resp interface{}, r *http.Request) error {
return s.tables.Execute(ctx, s.filerClient, operation, req, resp, identityOf(r))
}
func isNotFound(err error) bool {
return err != nil && (strings.Contains(err.Error(), "NoSuchTableBucket") ||
strings.Contains(err.Error(), "NoSuchTable") ||
strings.Contains(err.Error(), "NotFoundException") ||
strings.Contains(err.Error(), "NoSuchBucket"))
}
func catalogErr(w http.ResponseWriter, err error) {
msg := err.Error()
switch {
case isNotFound(err):
writeError(w, http.StatusNotFound, "NotFoundException", msg)
case strings.Contains(msg, "AccessDenied") || strings.Contains(msg, "PermissionDenied"):
writeError(w, http.StatusForbidden, "AccessDeniedException", "access denied")
case strings.Contains(msg, "AlreadyExists") || strings.Contains(msg, "Conflict"):
writeError(w, http.StatusConflict, "ConflictException", msg)
default:
writeError(w, http.StatusBadRequest, "ValidationException", msg)
}
}
func (s *Server) handleCreateVectorBucket(w http.ResponseWriter, r *http.Request) {
var req CreateVectorBucketRequest
if err := readJSON(r, &req); err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
if req.VectorBucketName == "" {
writeError(w, http.StatusBadRequest, "ValidationException", "vectorBucketName is required")
return
}
var resp map[string]interface{}
err := s.catalog(r.Context(), "CreateTableBucket",
&s3tables.CreateTableBucketRequest{Name: req.VectorBucketName, Format: s3tables.FormatLance}, &resp, r)
if err != nil {
catalogErr(w, err)
return
}
writeJSON(w, http.StatusOK, &CreateVectorBucketResponse{})
}
func (s *Server) handleGetVectorBucket(w http.ResponseWriter, r *http.Request) {
var req GetVectorBucketRequest
if err := readJSON(r, &req); err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
var arns vectorBucketARNs
if err := arnParts(&arns, req.VectorBucketName, req.VectorBucketArn); err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
var resp map[string]interface{}
err := s.catalog(r.Context(), "GetTableBucket",
&s3tables.GetTableBucketRequest{TableBucketARN: s.s3tablesArn(arns.bucket)}, &resp, r)
if err != nil {
catalogErr(w, err)
return
}
created, _ := time.Parse(time.RFC3339, fmt.Sprint(resp["createdAt"]))
writeJSON(w, http.StatusOK, &GetVectorBucketResponse{VectorBucket: &VectorBucket{
VectorBucketArn: s.vectorBucketArn(arns.bucket),
VectorBucketName: arns.bucket,
CreationTime: created,
}})
}
func (s *Server) handleListVectorBuckets(w http.ResponseWriter, r *http.Request) {
var req ListVectorBucketsRequest
if err := readJSON(r, &req); err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
var resp struct {
TableBuckets []struct {
Arn string `json:"arn"`
Name string `json:"name"`
CreatedAt string `json:"createdAt"`
} `json:"tableBuckets"`
ContinuationToken string `json:"continuationToken"`
}
err := s.catalog(r.Context(), "ListTableBuckets",
&s3tables.ListTableBucketsRequest{
MaxBuckets: req.MaxResults, ContinuationToken: req.NextToken, Prefix: req.Prefix,
}, &resp, r)
if err != nil {
catalogErr(w, err)
return
}
out := &ListVectorBucketsResponse{NextToken: resp.ContinuationToken}
for _, b := range resp.TableBuckets {
created, _ := time.Parse(time.RFC3339, b.CreatedAt)
out.VectorBuckets = append(out.VectorBuckets, &VectorBucket{
VectorBucketArn: s.vectorBucketArn(b.Name),
VectorBucketName: b.Name,
CreationTime: created,
})
}
writeJSON(w, http.StatusOK, out)
}
func (s *Server) handleDeleteVectorBucket(w http.ResponseWriter, r *http.Request) {
var req DeleteVectorBucketRequest
if err := readJSON(r, &req); err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
var arns vectorBucketARNs
if err := arnParts(&arns, req.VectorBucketName, req.VectorBucketArn); err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
var resp map[string]interface{}
err := s.catalog(r.Context(), "DeleteTableBucket",
&s3tables.DeleteTableBucketRequest{TableBucketARN: s.s3tablesArn(arns.bucket)}, &resp, r)
if err != nil {
catalogErr(w, err)
return
}
writeJSON(w, http.StatusOK, &DeleteVectorBucketResponse{})
}
func distanceMetric(metric string) (vector_pb.DistanceMetric, bool) {
switch strings.ToLower(metric) {
case "", "euclidean":
return vector_pb.DistanceMetric_L2, true
case "cosine":
return vector_pb.DistanceMetric_COSINE, true
case "dotproduct":
return vector_pb.DistanceMetric_DOT_PRODUCT, true
}
return vector_pb.DistanceMetric_L2, false
}
func tableID(bucket, index string) *vector_pb.Table {
return &vector_pb.Table{Id: []string{bucket, indexNamespace, index}}
}
func (s *Server) handleCreateIndex(w http.ResponseWriter, r *http.Request) {
var req CreateIndexRequest
if err := readJSON(r, &req); err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
if strings.ToLower(req.DataType) != "float32" {
writeError(w, http.StatusBadRequest, "ValidationException", "only float32 vectors are supported")
return
}
metric, ok := distanceMetric(req.DistanceMetric)
if !ok {
writeError(w, http.StatusBadRequest, "ValidationException", "unknown distanceMetric "+req.DistanceMetric)
return
}
arns, err := resolveIndexARNs(req.VectorBucketName, req.VectorBucketArn, req.IndexName, req.IndexArn)
if err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
if req.Dimension <= 0 || req.Dimension > 4096 {
writeError(w, http.StatusBadRequest, "ValidationException", "dimension must be 1..=4096")
return
}
metadata := map[string]string{
"dimension": fmt.Sprint(req.Dimension),
"distanceMetric": req.DistanceMetric,
}
if keys, err := json.Marshal(req.MetadataConfiguration.NonFilterableMetadataKeys); req.MetadataConfiguration != nil && err == nil {
metadata["nonFilterableMetadataKeys"] = string(keys)
}
metadataJSON, _ := json.Marshal(metadata)
var nsResp map[string]interface{}
// The index namespace is implicit on AWS; declare it, tolerating an
// already-existing namespace.
if err := s.catalog(r.Context(), "CreateNamespace", &s3tables.CreateNamespaceRequest{
TableBucketARN: s.s3tablesArn(arns.bucket),
Namespace: []string{indexNamespace},
}, &nsResp, r); err != nil && !strings.Contains(err.Error(), "AlreadyExists") && !strings.Contains(err.Error(), "Conflict") {
catalogErr(w, err)
return
}
var resp map[string]interface{}
err = s.catalog(r.Context(), "CreateTable", &s3tables.CreateTableRequest{
TableBucketARN: s.s3tablesArn(arns.bucket),
Namespace: []string{indexNamespace},
Name: arns.index,
Format: s3tables.FormatLance,
Metadata: &s3tables.TableMetadata{FullMetadata: metadataJSON},
}, &resp, r)
if err != nil {
catalogErr(w, err)
return
}
var nonFilterable []string
if req.MetadataConfiguration != nil {
nonFilterable = req.MetadataConfiguration.NonFilterableMetadataKeys
}
_, err = s.vectors.CreateIndex(r.Context(), &vector_pb.CreateIndexRequest{
Table: tableID(arns.bucket, arns.index),
Dimension: uint32(req.Dimension),
DistanceMetric: metric,
NonFilterableMetadataKeys: nonFilterable,
})
if err != nil {
// Roll the index record back so a half-created index does not linger.
var delResp map[string]interface{}
_ = s.catalog(r.Context(), "DeleteTable", &s3tables.DeleteTableRequest{
TableBucketARN: s.s3tablesArn(arns.bucket),
Namespace: []string{indexNamespace},
Name: arns.index,
}, &delResp, r)
writeError(w, http.StatusInternalServerError, "InternalFailure", err.Error())
return
}
writeJSON(w, http.StatusOK, &CreateIndexResponse{})
}
func indexFromRecord(s *Server, bucket, name string, m map[string]interface{}) *Index {
idx := &Index{
IndexArn: s.indexArn(bucket, name),
IndexName: name,
VectorBucketName: bucket,
DataType: "float32",
}
if created, ok := m["createdAt"].(string); ok {
idx.CreationTime, _ = time.Parse(time.RFC3339, created)
}
if meta, ok := m["metadata"].(map[string]interface{}); ok {
if fm, ok := meta["fullMetadata"].(map[string]interface{}); ok {
var dim float64
fmt.Sscan(fmt.Sprint(fm["dimension"]), &dim)
idx.Dimension = int32(dim)
idx.DistanceMetric = fmt.Sprint(fm["distanceMetric"])
if keys, ok := fm["nonFilterableMetadataKeys"].([]interface{}); ok {
var list []string
for _, k := range keys {
list = append(list, fmt.Sprint(k))
}
idx.MetadataConfiguration = &MetadataConfiguration{NonFilterableMetadataKeys: list}
}
}
}
if idx.DistanceMetric == "" || idx.DistanceMetric == "<nil>" {
idx.DistanceMetric = "euclidean"
}
return idx
}
func (s *Server) handleGetIndex(w http.ResponseWriter, r *http.Request) {
var req GetIndexRequest
if err := readJSON(r, &req); err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
arns, err := resolveIndexARNs(req.VectorBucketName, req.VectorBucketArn, req.IndexName, req.IndexArn)
if err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
var resp map[string]interface{}
err = s.catalog(r.Context(), "GetTable", &s3tables.GetTableRequest{
TableBucketARN: s.s3tablesArn(arns.bucket),
Namespace: []string{indexNamespace},
Name: arns.index,
}, &resp, r)
if err != nil {
catalogErr(w, err)
return
}
writeJSON(w, http.StatusOK, &GetIndexResponse{Index: indexFromRecord(s, arns.bucket, req.IndexName, resp)})
}
func (s *Server) handleListIndexes(w http.ResponseWriter, r *http.Request) {
var req ListIndexesRequest
if err := readJSON(r, &req); err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
var arns vectorBucketARNs
if err := arnParts(&arns, req.VectorBucketName, req.VectorBucketArn); err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
var resp struct {
TableSummaries []struct {
Name string `json:"name"`
CreatedAt string `json:"createdAt"`
} `json:"tables"`
ContinuationToken string `json:"continuationToken"`
}
err := s.catalog(r.Context(), "ListTables", &s3tables.ListTablesRequest{
TableBucketARN: s.s3tablesArn(arns.bucket),
Namespace: []string{indexNamespace},
MaxTables: req.MaxResults,
ContinuationToken: req.NextToken,
Prefix: req.Prefix,
}, &resp, r)
if err != nil {
catalogErr(w, err)
return
}
out := &ListIndexesResponse{NextToken: resp.ContinuationToken}
for _, t := range resp.TableSummaries {
out.Indexes = append(out.Indexes, indexFromRecord(s, arns.bucket, t.Name, nil))
}
writeJSON(w, http.StatusOK, out)
}
func (s *Server) handleDeleteIndex(w http.ResponseWriter, r *http.Request) {
var req DeleteIndexRequest
if err := readJSON(r, &req); err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
arns, err := resolveIndexARNs(req.VectorBucketName, req.VectorBucketArn, req.IndexName, req.IndexArn)
if err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
_, _ = s.vectors.DropIndex(r.Context(), &vector_pb.DropIndexRequest{
Table: tableID(arns.bucket, arns.index),
})
var resp map[string]interface{}
err = s.catalog(r.Context(), "DeleteTable", &s3tables.DeleteTableRequest{
TableBucketARN: s.s3tablesArn(arns.bucket),
Namespace: []string{indexNamespace},
Name: arns.index,
}, &resp, r)
if err != nil {
catalogErr(w, err)
return
}
writeJSON(w, http.StatusOK, &DeleteIndexResponse{})
}
func metadataStruct(m map[string]interface{}) (*structpb.Struct, error) {
if m == nil {
return nil, nil
}
return structpb.NewStruct(m)
}
func structToMap(s *structpb.Struct) map[string]interface{} {
if s == nil {
return nil
}
return s.AsMap()
}
func (s *Server) handlePutVectors(w http.ResponseWriter, r *http.Request) {
var req PutVectorsRequest
if err := readJSON(r, &req); err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
arns, err := resolveIndexARNs(req.VectorBucketName, req.VectorBucketArn, req.IndexName, req.IndexArn)
if err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
if _, err := s.indexRecord(r.Context(), r, arns.bucket, arns.index); err != nil {
catalogErr(w, err)
return
}
vectors := make([]*vector_pb.Vector, 0, len(req.Vectors))
for _, v := range req.Vectors {
if v == nil {
writeError(w, http.StatusBadRequest, "ValidationException", "vectors contains a null element")
return
}
metadata, err := metadataStruct(v.Metadata)
if err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
var data []float32
if v.Data != nil {
data = v.Data.Float32
}
vectors = append(vectors, &vector_pb.Vector{Key: v.Key, Data: data, Metadata: metadata})
}
_, err = s.vectors.PutVectors(r.Context(), &vector_pb.PutVectorsRequest{
Table: tableID(arns.bucket, arns.index),
Vectors: vectors,
})
if err != nil {
writeError(w, http.StatusInternalServerError, "InternalFailure", err.Error())
return
}
writeJSON(w, http.StatusOK, &PutVectorsResponse{})
}
func toOutputVectors(vectors []*vector_pb.Vector, returnData, returnMetadata bool) []*OutputVector {
out := make([]*OutputVector, 0, len(vectors))
for _, v := range vectors {
o := &OutputVector{Key: v.Key}
if returnData {
o.Data = &VectorData{Float32: v.Data}
}
if returnMetadata {
o.Metadata = structToMap(v.Metadata)
}
out = append(out, o)
}
return out
}
func (s *Server) handleGetVectors(w http.ResponseWriter, r *http.Request) {
var req GetVectorsRequest
if err := readJSON(r, &req); err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
arns, err := resolveIndexARNs(req.VectorBucketName, req.VectorBucketArn, req.IndexName, req.IndexArn)
if err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
if _, err := s.indexRecord(r.Context(), r, arns.bucket, arns.index); err != nil {
catalogErr(w, err)
return
}
resp, err := s.vectors.GetVectors(r.Context(), &vector_pb.GetVectorsRequest{
Table: tableID(arns.bucket, arns.index),
Keys: req.Keys,
ReturnData: req.ReturnData,
ReturnMetadata: req.ReturnMetadata,
})
if err != nil {
writeError(w, http.StatusInternalServerError, "InternalFailure", err.Error())
return
}
writeJSON(w, http.StatusOK, &GetVectorsResponse{
Vectors: toOutputVectors(resp.Vectors, req.ReturnData, req.ReturnMetadata),
})
}
func (s *Server) handleListVectors(w http.ResponseWriter, r *http.Request) {
var req ListVectorsRequest
if err := readJSON(r, &req); err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
arns, err := resolveIndexARNs(req.VectorBucketName, req.VectorBucketArn, req.IndexName, req.IndexArn)
if err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
if _, err := s.indexRecord(r.Context(), r, arns.bucket, arns.index); err != nil {
catalogErr(w, err)
return
}
// AWS exposes a static split: segmentIndex/segmentCount divides the full
// listing. The engine service does not know AWS segments, so page through
// the whole index here and slice the caller's share — O(index) per call,
// acceptable for the MVP.
if req.SegmentCount > 0 {
var all []*vector_pb.Vector
token := ""
for {
page, err := s.vectors.ListVectors(r.Context(), &vector_pb.ListVectorsRequest{
Table: tableID(arns.bucket, arns.index),
MaxResults: 1000,
NextToken: token,
ReturnData: req.ReturnData,
ReturnMetadata: req.ReturnMetadata,
})
if err != nil {
writeError(w, http.StatusInternalServerError, "InternalFailure", err.Error())
return
}
all = append(all, page.Vectors...)
if page.NextToken == "" {
break
}
token = page.NextToken
}
var out []*vector_pb.Vector
for i, v := range all {
if i%int(req.SegmentCount) == int(req.SegmentIndex) {
out = append(out, v)
}
}
writeJSON(w, http.StatusOK, &ListVectorsResponse{
Vectors: toOutputVectors(out, req.ReturnData, req.ReturnMetadata),
})
return
}
if req.MaxResults < 0 {
writeError(w, http.StatusBadRequest, "ValidationException", "maxResults must be non-negative")
return
}
resp, err := s.vectors.ListVectors(r.Context(), &vector_pb.ListVectorsRequest{
Table: tableID(arns.bucket, arns.index),
MaxResults: uint32(req.MaxResults),
NextToken: req.NextToken,
ReturnData: req.ReturnData,
ReturnMetadata: req.ReturnMetadata,
})
if err != nil {
writeError(w, http.StatusInternalServerError, "InternalFailure", err.Error())
return
}
writeJSON(w, http.StatusOK, &ListVectorsResponse{
NextToken: resp.NextToken,
Vectors: toOutputVectors(resp.Vectors, req.ReturnData, req.ReturnMetadata),
})
}
func (s *Server) handleDeleteVectors(w http.ResponseWriter, r *http.Request) {
var req DeleteVectorsRequest
if err := readJSON(r, &req); err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
arns, err := resolveIndexARNs(req.VectorBucketName, req.VectorBucketArn, req.IndexName, req.IndexArn)
if err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
if _, err := s.indexRecord(r.Context(), r, arns.bucket, arns.index); err != nil {
catalogErr(w, err)
return
}
_, err = s.vectors.DeleteVectors(r.Context(), &vector_pb.DeleteVectorsRequest{
Table: tableID(arns.bucket, arns.index),
Keys: req.Keys,
})
if err != nil {
writeError(w, http.StatusInternalServerError, "InternalFailure", err.Error())
return
}
writeJSON(w, http.StatusOK, &DeleteVectorsResponse{})
}
// translateFilter renders an AWS S3 Vectors filter document as a Lance SQL
// predicate over the metadata JSON column. Only the documented operators are
// translated; anything else fails loudly instead of silently overmatching.
func translateFilter(filter map[string]interface{}) (string, error) {
if len(filter) == 0 {
return "", nil
}
return filterExpr(filter)
}
func filterExpr(filter map[string]interface{}) (string, error) {
parts := make([]string, 0, len(filter))
for key, cond := range filter {
if strings.HasPrefix(key, "$") {
sub, ok := cond.([]interface{})
if !ok {
return "", fmt.Errorf("%s expects a list", key)
}
exprs := make([]string, 0, len(sub))
for _, item := range sub {
m, ok := item.(map[string]interface{})
if !ok {
return "", fmt.Errorf("%s expects objects", key)
}
e, err := filterExpr(m)
if err != nil {
return "", err
}
exprs = append(exprs, e)
}
joiner := " AND "
if key == "$or" {
joiner = " OR "
}
if key != "$and" && key != "$or" {
return "", fmt.Errorf("unsupported operator %s", key)
}
parts = append(parts, "("+strings.Join(exprs, joiner)+")")
continue
}
clause, err := fieldClause(key, cond)
if err != nil {
return "", err
}
parts = append(parts, clause)
}
return strings.Join(parts, " AND "), nil
}
func fieldClause(field string, cond interface{}) (string, error) {
col := fmt.Sprintf("json_get_str(`metadata`, '%s')", strings.ReplaceAll(field, "'", "''"))
switch c := cond.(type) {
case map[string]interface{}:
parts := []string{}
for op, operand := range c {
switch op {
case "$exists":
if operand == true {
parts = append(parts, fmt.Sprintf("json_exists(`metadata`, '%s')", strings.ReplaceAll(field, "'", "''")))
} else {
parts = append(parts, fmt.Sprintf("NOT json_exists(`metadata`, '%s')", strings.ReplaceAll(field, "'", "''")))
}
case "$eq", "$ne":
lit, err := sqlLiteral(operand)
if err != nil {
return "", err
}
sym := "="
if op == "$ne" {
sym = "<>"
}
parts = append(parts, fmt.Sprintf("%s %s %s", col, sym, lit))
case "$in", "$nin":
list, ok := operand.([]interface{})
if !ok {
return "", fmt.Errorf("%s expects a list", op)
}
lits := make([]string, 0, len(list))
for _, item := range list {
lit, err := sqlLiteral(item)
if err != nil {
return "", err
}
lits = append(lits, lit)
}
neg := ""
if op == "$nin" {
neg = "NOT "
}
parts = append(parts, fmt.Sprintf("%s%s in (%s)", neg, col, strings.Join(lits, ", ")))
case "$lt", "$lte", "$gt", "$gte":
lit, err := sqlLiteral(operand)
if err != nil {
return "", err
}
sym := map[string]string{"$lt": "<", "$lte": "<=", "$gt": ">", "$gte": ">="}[op]
numCol := fmt.Sprintf("json_get_float(`metadata`, '%s')", strings.ReplaceAll(field, "'", "''"))
parts = append(parts, fmt.Sprintf("%s %s %s", numCol, sym, lit))
default:
return "", fmt.Errorf("unsupported operator %s", op)
}
}
return strings.Join(parts, " AND "), nil
default:
lit, err := sqlLiteral(cond)
if err != nil {
return "", err
}
return fmt.Sprintf("%s = %s", col, lit), nil
}
}
func sqlLiteral(v interface{}) (string, error) {
switch t := v.(type) {
case string:
return "'" + strings.ReplaceAll(t, "'", "''") + "'", nil
case bool:
if t {
return "true", nil
}
return "false", nil
case float64:
return fmt.Sprintf("%v", t), nil
}
return "", fmt.Errorf("unsupported filter operand %v", v)
}
func (s *Server) handleQueryVectors(w http.ResponseWriter, r *http.Request) {
var req QueryVectorsRequest
if err := readJSON(r, &req); err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
arns, err := resolveIndexARNs(req.VectorBucketName, req.VectorBucketArn, req.IndexName, req.IndexArn)
if err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
var queryVector []float32
if req.QueryVector != nil {
queryVector = req.QueryVector.Float32
}
if len(queryVector) == 0 {
writeError(w, http.StatusBadRequest, "ValidationException", "queryVector is required")
return
}
record, err := s.indexRecord(r.Context(), r, arns.bucket, arns.index)
if err != nil {
catalogErr(w, err)
return
}
metric := indexFromRecord(s, arns.bucket, arns.index, record).DistanceMetric
filter, err := translateFilter(req.Filter)
if err != nil {
writeError(w, http.StatusBadRequest, "ValidationException", err.Error())
return
}
if req.TopK < 0 {
writeError(w, http.StatusBadRequest, "ValidationException", "topK must be non-negative")
return
}
resp, err := s.vectors.QueryVectors(r.Context(), &vector_pb.QueryVectorsRequest{
Table: tableID(arns.bucket, arns.index),
QueryVector: queryVector,
TopK: uint32(req.TopK),
Filter: filter,
ReturnMetadata: req.ReturnMetadata,
ReturnDistance: req.ReturnDistance,
})
if err != nil {
writeError(w, http.StatusInternalServerError, "InternalFailure", err.Error())
return
}
out := &QueryVectorsResponse{DistanceMetric: metric}
for _, v := range resp.Vectors {
o := &QueryOutputVector{Key: v.Key}
if req.ReturnDistance {
o.Distance = &v.Distance
}
if req.ReturnMetadata {
o.Metadata = structToMap(v.Metadata)
}
out.Vectors = append(out.Vectors, o)
}
writeJSON(w, http.StatusOK, out)
}
+94
View File
@@ -0,0 +1,94 @@
// Package s3vectors serves the AWS S3 Vectors API on top of the shared
// VectorService gRPC boundary and the S3 Tables catalog.
package s3vectors
import (
"net/http"
"strings"
"github.com/gorilla/mux"
"github.com/seaweedfs/seaweedfs/weed/pb/filer_pb"
"github.com/seaweedfs/seaweedfs/weed/pb/vector_pb"
"github.com/seaweedfs/seaweedfs/weed/s3api/s3err"
"github.com/seaweedfs/seaweedfs/weed/s3api/s3tables"
)
// FilerClient provides access to the filer for storage operations.
type FilerClient interface {
WithFilerClient(streamingMode bool, fn func(client filer_pb.SeaweedFilerClient) error) error
}
// S3Authenticator authenticates a request, returning the caller identity.
type S3Authenticator interface {
AuthenticateRequest(r *http.Request) (string, interface{}, s3err.ErrorCode)
DefaultAllow() bool
}
// Server implements the AWS S3 Vectors JSON API.
type Server struct {
filerClient FilerClient
tables *s3tables.Manager
vectors vector_pb.VectorServiceClient
authenticator S3Authenticator
region string
accountID string
}
// NewServer creates an S3 Vectors server over the given filer and catalog.
func NewServer(filerClient FilerClient, authenticator S3Authenticator, vectors vector_pb.VectorServiceClient) *Server {
manager := s3tables.NewManager()
if authenticator != nil {
manager.SetDefaultAllow(authenticator.DefaultAllow())
}
return &Server{
filerClient: filerClient,
tables: manager,
vectors: vectors,
authenticator: authenticator,
}
}
// SetRegion configures the region embedded in the ARNs this server mints.
func (s *Server) SetRegion(region string) { s.region = region }
// SetAccountID configures the account id embedded in the ARNs this server mints.
func (s *Server) SetAccountID(id string) { s.accountID = id }
// RegisterRoutes mounts POST / dispatching on the X-Amz-Target header and
// POST /{Operation} for clients that route per-path.
func (s *Server) RegisterRoutes(router *mux.Router) {
router.HandleFunc("/", s.Auth(s.dispatch)).Methods(http.MethodPost)
router.HandleFunc("/{op}", s.Auth(s.dispatch)).Methods(http.MethodPost)
}
func (s *Server) dispatch(w http.ResponseWriter, r *http.Request) {
// Clients carry the operation in the X-Amz-Target header or in the path —
// POST /QueryVectors and POST / with a target mean the same call here.
op := strings.Trim(r.URL.Path, "/")
if i := strings.LastIndex(op, "/"); i >= 0 {
op = op[i+1:]
}
if op == "" {
target := r.Header.Get("X-Amz-Target")
op = target[strings.LastIndex(target, ".")+1:]
}
handler, ok := handlers[op]
if !ok {
writeError(w, http.StatusBadRequest, "ValidationException", "unknown operation "+op)
return
}
handler(s, w, r)
}
// Auth authenticates the caller and puts the identity in the request context.
func (s *Server) Auth(handler http.HandlerFunc) http.HandlerFunc {
return func(w http.ResponseWriter, r *http.Request) {
identity, _, errCode := s.authenticator.AuthenticateRequest(r)
if errCode != s3err.ErrNone {
writeError(w, http.StatusForbidden, "AccessDenied", "authentication failed")
return
}
r = r.WithContext(contextWithIdentity(r.Context(), identity))
handler(w, r)
}
}
+251
View File
@@ -0,0 +1,251 @@
package s3vectors
import (
"context"
"encoding/json"
"io"
"net/http"
"time"
)
type contextKey string
const identityKey contextKey = "s3vectors-identity"
func contextWithIdentity(ctx context.Context, identity string) context.Context {
return context.WithValue(ctx, identityKey, identity)
}
func identityOf(r *http.Request) string {
v, _ := r.Context().Value(identityKey).(string)
return v
}
func writeJSON(w http.ResponseWriter, status int, payload interface{}) {
w.Header().Set("Content-Type", "application/x-amz-json-1.1")
w.WriteHeader(status)
if payload != nil {
_ = json.NewEncoder(w).Encode(payload)
}
}
func writeError(w http.ResponseWriter, status int, code, message string) {
writeJSON(w, status, map[string]string{
"__type": code,
"code": code,
"message": message,
"requestId": "local",
})
}
const maxRequestBodyBytes = 32 << 20
func readJSON(r *http.Request, req interface{}) error {
return json.NewDecoder(io.LimitReader(r.Body, maxRequestBodyBytes)).Decode(req)
}
// Request/response shapes for the AWS S3 Vectors API (awsJson1_1).
type vectorBucketARNs struct {
bucket string
index string
}
type CreateVectorBucketRequest struct {
VectorBucketName string `json:"vectorBucketName"`
EncryptionConfiguration map[string]interface{} `json:"encryptionConfiguration,omitempty"`
MetadataResourceTags map[string]string `json:"-"`
}
type CreateVectorBucketResponse struct{}
type GetVectorBucketRequest struct {
VectorBucketName string `json:"vectorBucketName"`
VectorBucketArn string `json:"vectorBucketArn"`
}
type GetVectorBucketResponse struct {
VectorBucket *VectorBucket `json:"vectorBucket"`
}
type VectorBucket struct {
VectorBucketArn string `json:"vectorBucketArn"`
VectorBucketName string `json:"vectorBucketName"`
CreationTime time.Time `json:"creationTime"`
}
type ListVectorBucketsRequest struct {
MaxResults int `json:"maxResults"`
NextToken string `json:"nextToken"`
Prefix string `json:"prefix"`
}
type ListVectorBucketsResponse struct {
NextToken string `json:"nextToken,omitempty"`
VectorBuckets []*VectorBucket `json:"vectorBuckets"`
}
type DeleteVectorBucketRequest struct {
VectorBucketName string `json:"vectorBucketName"`
VectorBucketArn string `json:"vectorBucketArn"`
}
type DeleteVectorBucketResponse struct{}
type CreateIndexRequest struct {
DataType string `json:"dataType"`
Dimension int32 `json:"dimension"`
DistanceMetric string `json:"distanceMetric"`
IndexArn string `json:"indexArn"`
IndexName string `json:"indexName"`
MetadataConfiguration *MetadataConfiguration `json:"metadataConfiguration,omitempty"`
VectorBucketName string `json:"vectorBucketName"`
VectorBucketArn string `json:"vectorBucketArn"`
}
type MetadataConfiguration struct {
NonFilterableMetadataKeys []string `json:"nonFilterableMetadataKeys"`
}
type CreateIndexResponse struct{}
type GetIndexRequest struct {
IndexArn string `json:"indexArn"`
IndexName string `json:"indexName"`
VectorBucketName string `json:"vectorBucketName"`
VectorBucketArn string `json:"vectorBucketArn"`
}
type Index struct {
CreationTime time.Time `json:"creationTime"`
DataType string `json:"dataType"`
Dimension int32 `json:"dimension"`
DistanceMetric string `json:"distanceMetric"`
IndexArn string `json:"indexArn"`
IndexName string `json:"indexName"`
MetadataConfiguration *MetadataConfiguration `json:"metadataConfiguration,omitempty"`
VectorBucketName string `json:"vectorBucketName"`
}
type GetIndexResponse struct {
Index *Index `json:"index"`
}
type ListIndexesRequest struct {
MaxResults int `json:"maxResults"`
NextToken string `json:"nextToken"`
Prefix string `json:"prefix"`
VectorBucketName string `json:"vectorBucketName"`
VectorBucketArn string `json:"vectorBucketArn"`
}
type ListIndexesResponse struct {
Indexes []*Index `json:"indexes"`
NextToken string `json:"nextToken,omitempty"`
}
type DeleteIndexRequest struct {
IndexArn string `json:"indexArn"`
IndexName string `json:"indexName"`
VectorBucketName string `json:"vectorBucketName"`
VectorBucketArn string `json:"vectorBucketArn"`
}
type DeleteIndexResponse struct{}
type PutInputVector struct {
Data *VectorData `json:"data"`
Key string `json:"key"`
Metadata map[string]interface{} `json:"metadata,omitempty"`
}
type VectorData struct {
Float32 []float32 `json:"float32"`
}
type PutVectorsRequest struct {
VectorBucketName string `json:"vectorBucketName"`
VectorBucketArn string `json:"vectorBucketArn"`
IndexArn string `json:"indexArn"`
IndexName string `json:"indexName"`
Vectors []*PutInputVector `json:"vectors"`
}
type PutVectorsResponse struct{}
type GetVectorsRequest struct {
IndexArn string `json:"indexArn"`
IndexName string `json:"indexName"`
VectorBucketName string `json:"vectorBucketName"`
VectorBucketArn string `json:"vectorBucketArn"`
Keys []string `json:"keys"`
ReturnData bool `json:"returnData"`
ReturnMetadata bool `json:"returnMetadata"`
}
type OutputVector struct {
Data *VectorData `json:"data,omitempty"`
Key string `json:"key"`
Metadata map[string]interface{} `json:"metadata,omitempty"`
}
type GetVectorsResponse struct {
Vectors []*OutputVector `json:"vectors"`
}
type ListVectorsRequest struct {
IndexArn string `json:"indexArn"`
IndexName string `json:"indexName"`
VectorBucketName string `json:"vectorBucketName"`
VectorBucketArn string `json:"vectorBucketArn"`
MaxResults int `json:"maxResults"`
NextToken string `json:"nextToken"`
ReturnData bool `json:"returnData"`
ReturnMetadata bool `json:"returnMetadata"`
SegmentCount int `json:"segmentCount"`
SegmentIndex int `json:"segmentIndex"`
}
type ListVectorsResponse struct {
NextToken string `json:"nextToken,omitempty"`
Vectors []*OutputVector `json:"vectors"`
}
type DeleteVectorsRequest struct {
IndexArn string `json:"indexArn"`
IndexName string `json:"indexName"`
VectorBucketName string `json:"vectorBucketName"`
VectorBucketArn string `json:"vectorBucketArn"`
Keys []string `json:"keys"`
}
type DeleteVectorsResponse struct{}
type QueryVector struct {
Float32 []float32 `json:"float32"`
}
type QueryVectorsRequest struct {
DistanceMetric string `json:"distanceMetric"`
Filter map[string]interface{} `json:"filter"`
IndexArn string `json:"indexArn"`
IndexName string `json:"indexName"`
QueryVector *QueryVector `json:"queryVector"`
ReturnDistance bool `json:"returnDistance"`
ReturnMetadata bool `json:"returnMetadata"`
TopK int32 `json:"topK"`
VectorBucketName string `json:"vectorBucketName"`
VectorBucketArn string `json:"vectorBucketArn"`
}
type QueryOutputVector struct {
Data *VectorData `json:"data,omitempty"`
Distance *float32 `json:"distance,omitempty"`
Key string `json:"key"`
Metadata map[string]interface{} `json:"metadata,omitempty"`
}
type QueryVectorsResponse struct {
DistanceMetric string `json:"distanceMetric"`
Vectors []*QueryOutputVector `json:"vectors"`
}