Package org.springframework.ai.vectorstore

Class PgVectorStore

java.lang.Object

org.springframework.ai.vectorstore.PgVectorStore

All Implemented Interfaces:

Consumer<List<Document>>, DocumentWriter, VectorStore, org.springframework.beans.factory.InitializingBean

public class PgVectorStore
extends Object
implements VectorStore, org.springframework.beans.factory.InitializingBean

Uses the "vector_store" table to store the Spring AI vector data. The table and the vector index will be auto-created if not available.

Author:

Christian Tzolov

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static enum	PgVectorStore.PgDistanceType	Defaults to CosineDistance.
static enum	PgVectorStore.PgIndexType	By default, pgvector performs exact nearest neighbor search, which provides perfect recall.

Field Summary

Fields

Modifier and Type	Field	Description
final FilterExpressionConverter	filterExpressionConverter
static final int	INVALID_EMBEDDING_DIMENSION
static final int	OPENAI_EMBEDDING_DIMENSION_SIZE
static final String	VECTOR_TABLE_NAME

Constructor Summary

Constructors

Constructor	Description
PgVectorStore(org.springframework.jdbc.core.JdbcTemplate jdbcTemplate, EmbeddingClient embeddingClient)
PgVectorStore(org.springframework.jdbc.core.JdbcTemplate jdbcTemplate, EmbeddingClient embeddingClient, int dimensions)
PgVectorStore(org.springframework.jdbc.core.JdbcTemplate jdbcTemplate, EmbeddingClient embeddingClient, int dimensions, PgVectorStore.PgDistanceType distanceType, boolean removeExistingVectorStoreTable, PgVectorStore.PgIndexType createIndexMethod)

Method Summary

Modifier and Type	Method	Description
void	add(List<Document> documents)	Adds list of Documents to the vector store.
void	afterPropertiesSet()
Optional<Boolean>	delete(List<String> idList)	Deletes documents from the vector store.
List<Double>	embeddingDistance(String query)
PgVectorStore.PgDistanceType	getDistanceType()
List<Document>	similaritySearch(SearchRequest request)	Retrieves documents by query embedding similarity and metadata filters to retrieve exactly the number of nearest-neighbor results that match the request criteria.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Methods inherited from interface java.util.function.Consumer

andThen

Methods inherited from interface org.springframework.ai.vectorstore.VectorStore

accept, similaritySearch

Field Details

OPENAI_EMBEDDING_DIMENSION_SIZE

public static final int OPENAI_EMBEDDING_DIMENSION_SIZE

See Also:

• Constant Field Values

INVALID_EMBEDDING_DIMENSION

public static final int INVALID_EMBEDDING_DIMENSION

See Also:

• Constant Field Values

VECTOR_TABLE_NAME

public static final String VECTOR_TABLE_NAME

See Also:

• Constant Field Values

filterExpressionConverter

public final FilterExpressionConverter filterExpressionConverter

Constructor Details

PgVectorStore

public PgVectorStore(org.springframework.jdbc.core.JdbcTemplate jdbcTemplate,
 EmbeddingClient embeddingClient)

PgVectorStore

public PgVectorStore(org.springframework.jdbc.core.JdbcTemplate jdbcTemplate,
 EmbeddingClient embeddingClient,
 int dimensions)

PgVectorStore

public PgVectorStore(org.springframework.jdbc.core.JdbcTemplate jdbcTemplate,
 EmbeddingClient embeddingClient,
 int dimensions,
 PgVectorStore.PgDistanceType distanceType,
 boolean removeExistingVectorStoreTable,
 PgVectorStore.PgIndexType createIndexMethod)

Method Details

getDistanceType

public PgVectorStore.PgDistanceType getDistanceType()

add

public void add(List<Document> documents)

Description copied from interface: VectorStore

Adds list of Documents to the vector store.

Specified by:

add in interface VectorStore

Parameters:

documents - the list of documents to store. Throws an exception if the underlying provider checks for duplicate IDs.

delete

public Optional<Boolean> delete(List<String> idList)

Description copied from interface: VectorStore

Deletes documents from the vector store.

Specified by:

delete in interface VectorStore

Parameters:

idList - list of document ids for which documents will be removed.

Returns:

similaritySearch

public List<Document> similaritySearch(SearchRequest request)

Description copied from interface: VectorStore

Retrieves documents by query embedding similarity and metadata filters to retrieve exactly the number of nearest-neighbor results that match the request criteria.

Specified by:

similaritySearch in interface VectorStore

Parameters:

request - Search request for set search parameters, such as the query text, topK, similarity threshold and metadata filter expressions.

Returns:

Returns documents th match the query request conditions.

embeddingDistance

public List<Double> embeddingDistance(String query)

afterPropertiesSet

public void afterPropertiesSet()
                        throws Exception

Specified by:

afterPropertiesSet in interface org.springframework.beans.factory.InitializingBean

Throws:

Exception