googleapis/google/cloud/aiplatform/v1/index.proto
2025-08-04 09:46:35 -07:00

267 lines
11 KiB
Protocol Buffer

// Copyright 2025 Google LLC
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
syntax = "proto3";
package google.cloud.aiplatform.v1;
import "google/api/field_behavior.proto";
import "google/api/resource.proto";
import "google/cloud/aiplatform/v1/deployed_index_ref.proto";
import "google/cloud/aiplatform/v1/encryption_spec.proto";
import "google/protobuf/struct.proto";
import "google/protobuf/timestamp.proto";
option csharp_namespace = "Google.Cloud.AIPlatform.V1";
option go_package = "cloud.google.com/go/aiplatform/apiv1/aiplatformpb;aiplatformpb";
option java_multiple_files = true;
option java_outer_classname = "IndexProto";
option java_package = "com.google.cloud.aiplatform.v1";
option php_namespace = "Google\\Cloud\\AIPlatform\\V1";
option ruby_package = "Google::Cloud::AIPlatform::V1";
// A representation of a collection of database items organized in a way that
// allows for approximate nearest neighbor (a.k.a ANN) algorithms search.
message Index {
option (google.api.resource) = {
type: "aiplatform.googleapis.com/Index"
pattern: "projects/{project}/locations/{location}/indexes/{index}"
};
// The update method of an Index.
enum IndexUpdateMethod {
// Should not be used.
INDEX_UPDATE_METHOD_UNSPECIFIED = 0;
// BatchUpdate: user can call UpdateIndex with files on Cloud Storage of
// Datapoints to update.
BATCH_UPDATE = 1;
// StreamUpdate: user can call UpsertDatapoints/DeleteDatapoints to update
// the Index and the updates will be applied in corresponding
// DeployedIndexes in nearly real-time.
STREAM_UPDATE = 2;
}
// Output only. The resource name of the Index.
string name = 1 [(google.api.field_behavior) = OUTPUT_ONLY];
// Required. The display name of the Index.
// The name can be up to 128 characters long and can consist of any UTF-8
// characters.
string display_name = 2 [(google.api.field_behavior) = REQUIRED];
// The description of the Index.
string description = 3;
// Immutable. Points to a YAML file stored on Google Cloud Storage describing
// additional information about the Index, that is specific to it. Unset if
// the Index does not have any additional information. The schema is defined
// as an OpenAPI 3.0.2 [Schema
// Object](https://github.com/OAI/OpenAPI-Specification/blob/main/versions/3.0.2.md#schemaObject).
// Note: The URI given on output will be immutable and probably different,
// including the URI scheme, than the one given on input. The output URI will
// point to a location where the user only has a read access.
string metadata_schema_uri = 4 [(google.api.field_behavior) = IMMUTABLE];
// An additional information about the Index; the schema of the metadata can
// be found in
// [metadata_schema][google.cloud.aiplatform.v1.Index.metadata_schema_uri].
google.protobuf.Value metadata = 6;
// Output only. The pointers to DeployedIndexes created from this Index.
// An Index can be only deleted if all its DeployedIndexes had been undeployed
// first.
repeated DeployedIndexRef deployed_indexes = 7
[(google.api.field_behavior) = OUTPUT_ONLY];
// Used to perform consistent read-modify-write updates. If not set, a blind
// "overwrite" update happens.
string etag = 8;
// The labels with user-defined metadata to organize your Indexes.
//
// Label keys and values can be no longer than 64 characters
// (Unicode codepoints), can only contain lowercase letters, numeric
// characters, underscores and dashes. International characters are allowed.
//
// See https://goo.gl/xmQnxf for more information and examples of labels.
map<string, string> labels = 9;
// Output only. Timestamp when this Index was created.
google.protobuf.Timestamp create_time = 10
[(google.api.field_behavior) = OUTPUT_ONLY];
// Output only. Timestamp when this Index was most recently updated.
// This also includes any update to the contents of the Index.
// Note that Operations working on this Index may have their
// [Operations.metadata.generic_metadata.update_time]
// [google.cloud.aiplatform.v1.GenericOperationMetadata.update_time] a little
// after the value of this timestamp, yet that does not mean their results are
// not already reflected in the Index. Result of any successfully completed
// Operation on the Index is reflected in it.
google.protobuf.Timestamp update_time = 11
[(google.api.field_behavior) = OUTPUT_ONLY];
// Output only. Stats of the index resource.
IndexStats index_stats = 14 [(google.api.field_behavior) = OUTPUT_ONLY];
// Immutable. The update method to use with this Index. If not set,
// BATCH_UPDATE will be used by default.
IndexUpdateMethod index_update_method = 16
[(google.api.field_behavior) = IMMUTABLE];
// Immutable. Customer-managed encryption key spec for an Index. If set, this
// Index and all sub-resources of this Index will be secured by this key.
EncryptionSpec encryption_spec = 17 [(google.api.field_behavior) = IMMUTABLE];
// Output only. Reserved for future use.
bool satisfies_pzs = 18 [(google.api.field_behavior) = OUTPUT_ONLY];
// Output only. Reserved for future use.
bool satisfies_pzi = 19 [(google.api.field_behavior) = OUTPUT_ONLY];
}
// A datapoint of Index.
message IndexDatapoint {
// Feature embedding vector for sparse index. An array of numbers whose values
// are located in the specified dimensions.
message SparseEmbedding {
// Required. The list of embedding values of the sparse vector.
repeated float values = 1 [(google.api.field_behavior) = REQUIRED];
// Required. The list of indexes for the embedding values of the sparse
// vector.
repeated int64 dimensions = 2 [(google.api.field_behavior) = REQUIRED];
}
// Restriction of a datapoint which describe its attributes(tokens) from each
// of several attribute categories(namespaces).
message Restriction {
// The namespace of this restriction. e.g.: color.
string namespace = 1;
// The attributes to allow in this namespace. e.g.: 'red'
repeated string allow_list = 2;
// The attributes to deny in this namespace. e.g.: 'blue'
repeated string deny_list = 3;
}
// This field allows restricts to be based on numeric comparisons rather
// than categorical tokens.
message NumericRestriction {
// Which comparison operator to use. Should be specified for queries only;
// specifying this for a datapoint is an error.
//
// Datapoints for which Operator is true relative to the query's Value
// field will be allowlisted.
enum Operator {
// Default value of the enum.
OPERATOR_UNSPECIFIED = 0;
// Datapoints are eligible iff their value is < the query's.
LESS = 1;
// Datapoints are eligible iff their value is <= the query's.
LESS_EQUAL = 2;
// Datapoints are eligible iff their value is == the query's.
EQUAL = 3;
// Datapoints are eligible iff their value is >= the query's.
GREATER_EQUAL = 4;
// Datapoints are eligible iff their value is > the query's.
GREATER = 5;
// Datapoints are eligible iff their value is != the query's.
NOT_EQUAL = 6;
}
// The type of Value must be consistent for all datapoints with a given
// namespace name. This is verified at runtime.
oneof Value {
// Represents 64 bit integer.
int64 value_int = 2;
// Represents 32 bit float.
float value_float = 3;
// Represents 64 bit float.
double value_double = 4;
}
// The namespace of this restriction. e.g.: cost.
string namespace = 1;
// This MUST be specified for queries and must NOT be specified for
// datapoints.
Operator op = 5;
}
// Crowding tag is a constraint on a neighbor list produced by nearest
// neighbor search requiring that no more than some value k' of the k
// neighbors returned have the same value of crowding_attribute.
message CrowdingTag {
// The attribute value used for crowding. The maximum number of neighbors
// to return per crowding attribute value
// (per_crowding_attribute_num_neighbors) is configured per-query. This
// field is ignored if per_crowding_attribute_num_neighbors is larger than
// the total number of neighbors to return for a given query.
string crowding_attribute = 1;
}
// Required. Unique identifier of the datapoint.
string datapoint_id = 1 [(google.api.field_behavior) = REQUIRED];
// Required. Feature embedding vector for dense index. An array of numbers
// with the length of [NearestNeighborSearchConfig.dimensions].
repeated float feature_vector = 2 [(google.api.field_behavior) = REQUIRED];
// Optional. Feature embedding vector for sparse index.
SparseEmbedding sparse_embedding = 7 [(google.api.field_behavior) = OPTIONAL];
// Optional. List of Restrict of the datapoint, used to perform "restricted
// searches" where boolean rule are used to filter the subset of the database
// eligible for matching. This uses categorical tokens. See:
// https://cloud.google.com/vertex-ai/docs/matching-engine/filtering
repeated Restriction restricts = 4 [(google.api.field_behavior) = OPTIONAL];
// Optional. List of Restrict of the datapoint, used to perform "restricted
// searches" where boolean rule are used to filter the subset of the database
// eligible for matching. This uses numeric comparisons.
repeated NumericRestriction numeric_restricts = 6
[(google.api.field_behavior) = OPTIONAL];
// Optional. CrowdingTag of the datapoint, the number of neighbors to return
// in each crowding can be configured during query.
CrowdingTag crowding_tag = 5 [(google.api.field_behavior) = OPTIONAL];
// Optional. The key-value map of additional metadata for the datapoint.
google.protobuf.Struct embedding_metadata = 8
[(google.api.field_behavior) = OPTIONAL];
}
// Stats of the Index.
message IndexStats {
// Output only. The number of dense vectors in the Index.
int64 vectors_count = 1 [(google.api.field_behavior) = OUTPUT_ONLY];
// Output only. The number of sparse vectors in the Index.
int64 sparse_vectors_count = 3 [(google.api.field_behavior) = OUTPUT_ONLY];
// Output only. The number of shards in the Index.
int32 shards_count = 2 [(google.api.field_behavior) = OUTPUT_ONLY];
}