使用向量嵌入搜索

本页面介绍如何通过以下方法使用 Cloud Firestore 执行 K 最近邻 (KNN) 向量搜索：

  • 存储向量值
  • 创建和管理 KNN 向量索引
  • 使用一个支持的向量距离衡量方式执行 K 最近邻 (KNN) 查询

存储向量嵌入

您可以使用 Cloud Firestore 数据创建向量值（例如文本嵌入），并将它们存储在 Cloud Firestore 文档中。

通过向量嵌入执行写入操作

以下示例展示了如何将向量嵌入存储在 Cloud Firestore 文档中：

Python
from google.cloud import firestore
from google.cloud.firestore_v1.vector import Vector

firestore_client = firestore.Client()
collection = firestore_client.collection("coffee-beans")
doc = {
    "name": "Kahawa coffee beans",
    "description": "Information about the Kahawa coffee beans.",
    "embedding_field": Vector([1.0, 2.0, 3.0]),
}

collection.add(doc)
vector_search.py
Node.js
import {
  Firestore,
  FieldValue,
} from "@google-cloud/firestore";

const db = new Firestore();
const coll = db.collection('coffee-beans');
await coll.add({
  name: "Kahawa coffee beans",
  description: "Information about the Kahawa coffee beans.",
  embedding_field: FieldValue.vector([1.0 , 2.0, 3.0])
});
Go
import (
	"context"
	"fmt"
	"io"

	"cloud.google.com/go/firestore"
)

type CoffeeBean struct {
	Name           string             `firestore:"name,omitempty"`
	Description    string             `firestore:"description,omitempty"`
	EmbeddingField firestore.Vector32 `firestore:"embedding_field,omitempty"`
	Color          string             `firestore:"color,omitempty"`
}

func storeVectors(w io.Writer, projectID string) error {
	ctx := context.Background()

	// Create client
	client, err := firestore.NewClient(ctx, projectID)
	if err != nil {
		return fmt.Errorf("firestore.NewClient: %w", err)
	}
	defer client.Close()

	// Vector can be represented by Vector32 or Vector64
	doc := CoffeeBean{
		Name:           "Kahawa coffee beans",
		Description:    "Information about the Kahawa coffee beans.",
		EmbeddingField: []float32{1.0, 2.0, 3.0},
		Color:          "red",
	}
	ref := client.Collection("coffee-beans").NewDoc()
	if _, err = ref.Set(ctx, doc); err != nil {
		fmt.Fprintf(w, "failed to upsert: %v", err)
		return err
	}

	return nil
}

vector_store.go
Java
import com.google.cloud.firestore.CollectionReference;
import com.google.cloud.firestore.DocumentReference;
import com.google.cloud.firestore.FieldValue;
import com.google.cloud.firestore.VectorQuery;

CollectionReference coll = firestore.collection("coffee-beans");

Map<String, Object> docData = new HashMap<>();
docData.put("name", "Kahawa coffee beans");
docData.put("description", "Information about the Kahawa coffee beans.");
docData.put("embedding_field", FieldValue.vector(new double[] {1.0, 2.0, 3.0}));

ApiFuture<DocumentReference> future = coll.add(docData);
DocumentReference documentReference = future.get();

使用 Cloud Functions 函数计算向量嵌入

如要在文档每次创建或更新时便计算并存储相应的向量嵌入，您可以设置一个 Cloud Functions 函数：

Python
@functions_framework.cloud_event
def store_embedding(cloud_event) -> None:
  """Triggers by a change to a Firestore document.
  """
  firestore_payload = firestore.DocumentEventData()
  payload = firestore_payload._pb.ParseFromString(cloud_event.data)

  collection_id, doc_id = from_payload(payload)
  # Call a function to calculate the embedding
  embedding = calculate_embedding(payload)
  # Update the document
  doc = firestore_client.collection(collection_id).document(doc_id)
  doc.set({"embedding_field": embedding}, merge=True)
Node.js
/**
 * A vector embedding will be computed from the
 * value of the `content` field. The vector value
 * will be stored in the `embedding` field. The
 * field names `content` and `embedding` are arbitrary
 * field names chosen for this example.
 */
async function storeEmbedding(event: FirestoreEvent<any>): Promise<void> {
  // Get the previous value of the document's `content` field.
  const previousDocumentSnapshot = event.data.before as QueryDocumentSnapshot;
  const previousContent = previousDocumentSnapshot.get("content");

  // Get the current value of the document's `content` field.
  const currentDocumentSnapshot = event.data.after as QueryDocumentSnapshot;
  const currentContent = currentDocumentSnapshot.get("content");

  // Don't update the embedding if the content field did not change
  if (previousContent === currentContent) {
    return;
  }

  // Call a function to calculate the embedding for the value
  // of the `content` field.
  const embeddingVector = calculateEmbedding(currentContent);

  // Update the `embedding` field on the document.
  await currentDocumentSnapshot.ref.update({
    embedding: embeddingVector,
  });
}
Go
  // Not yet supported in the Go client library
Java
  // Not yet supported in the Java client library

创建和管理向量索引

您必须先创建相应的索引，然后才能通过向量嵌入执行最近邻搜索。以下示例展示了如何使用 Google Cloud CLI 创建和管理向量索引。您还可以使用 Firebase CLI 和 Terraform 管理向量索引。

创建矢量索引

在创建向量索引之前，请先升级到最新版本的 Google Cloud CLI：

gcloud components update

如需创建向量索引，请使用 gcloud firestore indexes composite create：

gcloud
gcloud firestore indexes composite create \
--collection-group=collection-group \
--query-scope=COLLECTION \
--field-config field-path=vector-field,vector-config='vector-configuration' \
--database=database-id

其中：

  • collection-group 是集合组的 ID。
  • vector-field 是包含向量嵌入的字段的名称。
  • database-id 是相应数据库的 ID。
  • vector-configuration 包含向量 dimension 和索引类型。dimension 是一个不超过 2,048 的整数。索引类型必须为 flat。按如下方式设置索引配置的格式：{"dimension":"DIMENSION", "flat": "{}"}。

以下示例创建了一个复合索引，其中包含字段 vector-field 的向量索引以及字段 color 的升序索引。您可以在执行最近邻搜索之前使用此类索引预先过滤数据。

gcloud
gcloud firestore indexes composite create \
--collection-group=collection-group \
--query-scope=COLLECTION \
--field-config=order=ASCENDING,field-path="color" \
--field-config field-path=vector-field,vector-config='{"dimension":"1024", "flat": "{}"}' \
--database=database-id

列出所有向量索引

gcloud
gcloud firestore indexes composite list --database=database-id

将 database-id 替换为相应数据库的 ID。

删除矢量索引

gcloud
gcloud firestore indexes composite delete index-id --database=database-id

其中：

  • index-id 是要删除的索引的 ID。可使用 indexes composite list 检索索引 ID。
  • database-id 是相应数据库的 ID。

描述向量索引

gcloud
gcloud firestore indexes composite describe index-id --database=database-id

其中：

  • index-id 是要描述的索引的 ID。可使用 indexes composite list 检索索引 ID。
  • database-id 是相应数据库的 ID。

执行最近邻查询

您可以执行相似度搜索来查找向量嵌入的最近邻。相似度搜索需要使用向量索引。如果没有现成的索引，Cloud Firestore 会使用 gcloud CLI 建议一个可创建的索引。

以下示例会查找查询向量的 10 个最近邻。

Python
from google.cloud.firestore_v1.base_vector_query import DistanceMeasure
from google.cloud.firestore_v1.vector import Vector

collection = db.collection("coffee-beans")

# Requires a single-field vector index
vector_query = collection.find_nearest(
    vector_field="embedding_field",
    query_vector=Vector([3.0, 1.0, 2.0]),
    distance_measure=DistanceMeasure.EUCLIDEAN,
    limit=5,
)
vector_search.py
Node.js
import {
  Firestore,
  FieldValue,
  VectorQuery,
  VectorQuerySnapshot,
} from "@google-cloud/firestore";

// Requires a single-field vector index
const vectorQuery: VectorQuery = coll.findNearest({
  vectorField: 'embedding_field',
  queryVector: [3.0, 1.0, 2.0],
  limit: 10,
  distanceMeasure: 'EUCLIDEAN'
});

const vectorQuerySnapshot: VectorQuerySnapshot = await vectorQuery.get();
Go
import (
	"context"
	"fmt"
	"io"

	"cloud.google.com/go/firestore"
)

func vectorSearchBasic(w io.Writer, projectID string) error {
	ctx := context.Background()

	// Create client
	client, err := firestore.NewClient(ctx, projectID)
	if err != nil {
		return fmt.Errorf("firestore.NewClient: %w", err)
	}
	defer client.Close()

	collection := client.Collection("coffee-beans")

	// Requires a vector index
	// https://firebase.google.com/docs/firestore/vector-search#create_and_manage_vector_indexes
	vectorQuery := collection.FindNearest("embedding_field",
		[]float32{3.0, 1.0, 2.0},
		5,
		// More info: https://firebase.google.com/docs/firestore/vector-search#vector_distances
		firestore.DistanceMeasureEuclidean,
		nil)

	docs, err := vectorQuery.Documents(ctx).GetAll()
	if err != nil {
		fmt.Fprintf(w, "failed to get vector query results: %v", err)
		return err
	}

	for _, doc := range docs {
		fmt.Fprintln(w, doc.Data()["name"])
	}
	return nil
}

vector_search_basic.go
Java
import com.google.cloud.firestore.VectorQuery;
import com.google.cloud.firestore.VectorQuerySnapshot;

VectorQuery vectorQuery = coll.findNearest(
        "embedding_field",
        new double[] {3.0, 1.0, 2.0},
        /* limit */ 10,
        VectorQuery.DistanceMeasure.EUCLIDEAN);

ApiFuture<VectorQuerySnapshot> future = vectorQuery.get();
VectorQuerySnapshot vectorQuerySnapshot = future.get();

向量距离

最近邻查询支持下列向量距离选项：

  • EUCLIDEAN：测量向量之间的欧几里得距离。如需了解详情，请参阅欧几里得。
  • COSINE：基于向量之间的角度来比较向量，这样可以测量不依赖于向量大小的相似度。对于单位归一化向量，建议使用 DOT_PRODUCT，而不是余弦距离，虽然两者在数学上是等效的，但前者性能更好。如需了解详情，请参阅余弦相似度。
  • DOT_PRODUCT：与 COSINE 类似，但受向量大小影响。如需了解详情，请参阅点积。

选择距离衡量方式

根据您的所有向量嵌入是否已归一化，您可以确定要使用哪种距离衡量方式来计算距离衡量值。已归一化向量嵌入的大小（长度）正好为 1.0。

此外，如果您知道用于训练模型的距离衡量方式，请使用该距离衡量方式来计算向量嵌入之间的距离。

已归一化数据

如果您的数据集中所有向量嵌入都已归一化，那么这三种距离衡量方式都会提供相同的语义搜索结果。从本质上讲，虽然每种距离衡量方式都会返回不同的值，但这些值的排序方式相同。如果嵌入已归一化，DOT_PRODUCT 通常具有最高的计算效率，但在大多数情况下，差异可以忽略不计。不过，如果您的应用对性能非常敏感，DOT_PRODUCT 可能会有助于进行性能调优。

未归一化数据

如果您的数据集中的向量嵌入未归一化，那么从数学角度来看，使用 DOT_PRODUCT 作为距离衡量方式是不正确的，因为点积无法衡量距离。根据生成嵌入的方式和偏好的搜索类型，COSINE 或 EUCLIDEAN 距离衡量方式生成的搜索结果在主观上优于其他距离衡量方式。您可能需要对 COSINE 或 EUCLIDEAN 进行实验，以确定哪个最适合您的应用场景。

不确定数据是已归一化还是未归一化

如果您不确定数据是否已归一化，但又想使用 DOT_PRODUCT，我们建议您改用 COSINE。COSINE 类似于内置了归一化的 DOT_PRODUCT。使用 COSINE 衡量的距离范围为 0 到 2。结果接近 0 表示向量非常相似。

预先过滤文档

如需在查找最近邻之前预先过滤文档，您可以将相似度搜索与其他查询运算符结合使用。支持 and 和 or 复合过滤条件。如需详细了解支持的字段过滤条件，请参阅查询运算符。

Python
from google.cloud.firestore_v1.base_vector_query import DistanceMeasure
from google.cloud.firestore_v1.vector import Vector

collection = db.collection("coffee-beans")

# Similarity search with pre-filter
# Requires a composite vector index
vector_query = collection.where("color", "==", "red").find_nearest(
    vector_field="embedding_field",
    query_vector=Vector([3.0, 1.0, 2.0]),
    distance_measure=DistanceMeasure.EUCLIDEAN,
    limit=5,
)
vector_search.py
Node.js
// Similarity search with pre-filter
// Requires composite vector index
const preFilteredVectorQuery: VectorQuery = coll
    .where("color", "==", "red")
    .findNearest({
      vectorField: "embedding_field",
      queryVector: [3.0, 1.0, 2.0],
      limit: 5,
      distanceMeasure: "EUCLIDEAN",
    });

const vectorQueryResults = await preFilteredVectorQuery.get();
Go
import (
	"context"
	"fmt"
	"io"

	"cloud.google.com/go/firestore"
)

func vectorSearchPrefilter(w io.Writer, projectID string) error {
	ctx := context.Background()

	// Create client
	client, err := firestore.NewClient(ctx, projectID)
	if err != nil {
		return fmt.Errorf("firestore.NewClient: %w", err)
	}
	defer client.Close()

	collection := client.Collection("coffee-beans")

	// Similarity search with pre-filter
	// Requires a composite vector index
	vectorQuery := collection.Where("color", "==", "red").
		FindNearest("embedding_field",
			[]float32{3.0, 1.0, 2.0},
			5,
			// More info: https://firebase.google.com/docs/firestore/vector-search#vector_distances
			firestore.DistanceMeasureEuclidean,
			nil)

	docs, err := vectorQuery.Documents(ctx).GetAll()
	if err != nil {
		fmt.Fprintf(w, "failed to get vector query results: %v", err)
		return err
	}

	for _, doc := range docs {
		fmt.Fprintln(w, doc.Data()["name"])
	}
	return nil
}

vector_search_prefilter.go
Java
import com.google.cloud.firestore.VectorQuery;
import com.google.cloud.firestore.VectorQuerySnapshot;

VectorQuery preFilteredVectorQuery = coll
        .whereEqualTo("color", "red")
        .findNearest(
                "embedding_field",
                new double[] {3.0, 1.0, 2.0},
                /* limit */ 10,
                VectorQuery.DistanceMeasure.EUCLIDEAN);

ApiFuture<VectorQuerySnapshot> future = preFilteredVectorQuery.get();
VectorQuerySnapshot vectorQuerySnapshot = future.get();

检索计算出的向量距离

您可以通过对 FindNearest 查询分配 distance_result_field 输出属性名称来检索计算出的向量距离，如以下示例所示：

Python
from google.cloud.firestore_v1.base_vector_query import DistanceMeasure
from google.cloud.firestore_v1.vector import Vector

collection = db.collection("coffee-beans")

vector_query = collection.find_nearest(
    vector_field="embedding_field",
    query_vector=Vector([3.0, 1.0, 2.0]),
    distance_measure=DistanceMeasure.EUCLIDEAN,
    limit=10,
    distance_result_field="vector_distance",
)

docs = vector_query.stream()

for doc in docs:
    print(f"{doc.id}, Distance: {doc.get('vector_distance')}")
vector_search.py
Node.js
const vectorQuery: VectorQuery = coll.findNearest(
    {
      vectorField: 'embedding_field',
      queryVector: [3.0, 1.0, 2.0],
      limit: 10,
      distanceMeasure: 'EUCLIDEAN',
      distanceResultField: 'vector_distance'
    });

const snapshot: VectorQuerySnapshot = await vectorQuery.get();

snapshot.forEach((doc) => {
  console.log(doc.id, ' Distance: ', doc.get('vector_distance'));
});
Go
import (
	"context"
	"fmt"
	"io"

	"cloud.google.com/go/firestore"
)

func vectorSearchDistanceResultField(w io.Writer, projectID string) error {
	ctx := context.Background()

	client, err := firestore.NewClient(ctx, projectID)
	if err != nil {
		return fmt.Errorf("firestore.NewClient: %w", err)
	}
	defer client.Close()

	collection := client.Collection("coffee-beans")

	// Requires a vector index
	// https://firebase.google.com/docs/firestore/vector-search#create_and_manage_vector_indexes
	vectorQuery := collection.FindNearest("embedding_field",
		[]float32{3.0, 1.0, 2.0},
		10,
		firestore.DistanceMeasureEuclidean,
		&firestore.FindNearestOptions{
			DistanceResultField: "vector_distance",
		})

	docs, err := vectorQuery.Documents(ctx).GetAll()
	if err != nil {
		fmt.Fprintf(w, "failed to get vector query results: %v", err)
		return err
	}

	for _, doc := range docs {
		fmt.Fprintf(w, "%v, Distance: %v\n", doc.Data()["name"], doc.Data()["vector_distance"])
	}
	return nil
}

vector_search_result_field.go
Java
import com.google.cloud.firestore.VectorQuery;
import com.google.cloud.firestore.VectorQueryOptions;
import com.google.cloud.firestore.VectorQuerySnapshot;

VectorQuery vectorQuery = coll.findNearest(
        "embedding_field",
        new double[] {3.0, 1.0, 2.0},
        /* limit */ 10,
        VectorQuery.DistanceMeasure.EUCLIDEAN,
        VectorQueryOptions.newBuilder().setDistanceResultField("vector_distance").build());

ApiFuture<VectorQuerySnapshot> future = vectorQuery.get();
VectorQuerySnapshot vectorQuerySnapshot = future.get();

for (DocumentSnapshot document : vectorQuerySnapshot.getDocuments()) {
    System.out.println(document.getId() + " Distance: " + document.get("vector_distance"));
}

如果您想使用字段掩码返回部分文档字段以及 distanceResultField，则还必须在字段掩码中添加 distanceResultField 的值，如以下示例所示：

Python
vector_query = collection.select(["color", "vector_distance"]).find_nearest(
    vector_field="embedding_field",
    query_vector=Vector([3.0, 1.0, 2.0]),
    distance_measure=DistanceMeasure.EUCLIDEAN,
    limit=10,
    distance_result_field="vector_distance",
)
vector_search.py
Node.js
const vectorQuery: VectorQuery = coll
    .select('name', 'description', 'vector_distance')
    .findNearest({
      vectorField: 'embedding_field',
      queryVector: [3.0, 1.0, 2.0],
      limit: 10,
      distanceMeasure: 'EUCLIDEAN',
      distanceResultField: 'vector_distance'
    });
Go
import (
	"context"
	"fmt"
	"io"

	"cloud.google.com/go/firestore"
)

func vectorSearchDistanceResultFieldMasked(w io.Writer, projectID string) error {
	ctx := context.Background()

	client, err := firestore.NewClient(ctx, projectID)
	if err != nil {
		return fmt.Errorf("firestore.NewClient: %w", err)
	}
	defer client.Close()

	collection := client.Collection("coffee-beans")

	// Requires a vector index
	// https://firebase.google.com/docs/firestore/vector-search#create_and_manage_vector_indexes
	vectorQuery := collection.Select("color", "vector_distance").
		FindNearest("embedding_field",
			[]float32{3.0, 1.0, 2.0},
			10,
			firestore.DistanceMeasureEuclidean,
			&firestore.FindNearestOptions{
				DistanceResultField: "vector_distance",
			})

	docs, err := vectorQuery.Documents(ctx).GetAll()
	if err != nil {
		fmt.Fprintf(w, "failed to get vector query results: %v", err)
		return err
	}

	for _, doc := range docs {
		fmt.Fprintf(w, "%v, Distance: %v\n", doc.Data()["color"], doc.Data()["vector_distance"])
	}
	return nil
}

vector_search_result_field_masked.go
Java
import com.google.cloud.firestore.VectorQuery;
import com.google.cloud.firestore.VectorQueryOptions;
import com.google.cloud.firestore.VectorQuerySnapshot;

VectorQuery vectorQuery = coll
        .select("name", "description", "vector_distance")
        .findNearest(
          "embedding_field",
          new double[] {3.0, 1.0, 2.0},
          /* limit */ 10,
          VectorQuery.DistanceMeasure.EUCLIDEAN,
          VectorQueryOptions.newBuilder()
            .setDistanceResultField("vector_distance")
            .build());

ApiFuture<VectorQuerySnapshot> future = vectorQuery.get();
VectorQuerySnapshot vectorQuerySnapshot = future.get();

for (DocumentSnapshot document : vectorQuerySnapshot.getDocuments()) {
    System.out.println(document.getId() + " Distance: " + document.get("vector_distance"));
}

指定距离阈值

您可以指定相似度阈值，以便仅返回相似度在阈值范围内的文档。阈值字段的行为取决于您选择的距离衡量方式：

  • EUCLIDEAN 和 COSINE 距离会将阈值限制为距离小于或等于指定阈值的文档。随着向量的相似度增加，这些距离衡量值会减小。
  • DOT_PRODUCT 距离会将阈值限制为距离大于或等于指定阈值的文档。随着向量的相似度增加，点积距离会增加。

以下示例展示了如何指定距离阈值，以使用 EUCLIDEAN 距离指标返回最多 10 个距离不超过 4.5 个单位的最近文档：

Python
from google.cloud.firestore_v1.base_vector_query import DistanceMeasure
from google.cloud.firestore_v1.vector import Vector

collection = db.collection("coffee-beans")

vector_query = collection.find_nearest(
    vector_field="embedding_field",
    query_vector=Vector([3.0, 1.0, 2.0]),
    distance_measure=DistanceMeasure.EUCLIDEAN,
    limit=10,
    distance_threshold=4.5,
)

docs = vector_query.stream()

for doc in docs:
    print(f"{doc.id}")
vector_search.py
Node.js
const vectorQuery: VectorQuery = coll.findNearest({
  vectorField: 'embedding_field',
  queryVector: [3.0, 1.0, 2.0],
  limit: 10,
  distanceMeasure: 'EUCLIDEAN',
  distanceThreshold: 4.5
});

const snapshot: VectorQuerySnapshot = await vectorQuery.get();

snapshot.forEach((doc) => {
  console.log(doc.id);
});
Go
import (
	"context"
	"fmt"
	"io"

	"cloud.google.com/go/firestore"
)

func vectorSearchDistanceThreshold(w io.Writer, projectID string) error {
	ctx := context.Background()

	client, err := firestore.NewClient(ctx, projectID)
	if err != nil {
		return fmt.Errorf("firestore.NewClient: %w", err)
	}
	defer client.Close()

	collection := client.Collection("coffee-beans")

	// Requires a vector index
	// https://firebase.google.com/docs/firestore/vector-search#create_and_manage_vector_indexes
	vectorQuery := collection.FindNearest("embedding_field",
		[]float32{3.0, 1.0, 2.0},
		10,
		firestore.DistanceMeasureEuclidean,
		&firestore.FindNearestOptions{
			DistanceThreshold: firestore.Ptr[float64](4.5),
		})

	docs, err := vectorQuery.Documents(ctx).GetAll()
	if err != nil {
		fmt.Fprintf(w, "failed to get vector query results: %v", err)
		return err
	}

	for _, doc := range docs {
		fmt.Fprintln(w, doc.Data()["name"])
	}
	return nil
}

vector_search_distance_threshold.go
Java
import com.google.cloud.firestore.VectorQuery;
import com.google.cloud.firestore.VectorQueryOptions;
import com.google.cloud.firestore.VectorQuerySnapshot;

VectorQuery vectorQuery = coll.findNearest(
        "embedding_field",
        new double[] {3.0, 1.0, 2.0},
        /* limit */ 10,
        VectorQuery.DistanceMeasure.EUCLIDEAN,
        VectorQueryOptions.newBuilder()
          .setDistanceThreshold(4.5)
          .build());

ApiFuture<VectorQuerySnapshot> future = vectorQuery.get();
VectorQuerySnapshot vectorQuerySnapshot = future.get();

for (DocumentSnapshot document : vectorQuerySnapshot.getDocuments()) {
    System.out.println(document.getId());
}

限制

请注意，在使用向量嵌入时，有以下限制：

  • 支持的嵌入维度上限为 2,048。如要存储更大的索引，可使用降维。
  • 通过最近邻查询返回的文档数量上限为 1,000。
  • 向量搜索不支持实时快照监听器。
  • 只有 Python、Node.js、Go 和 Java 客户端库支持向量搜索。

后续步骤