PgVectorStore

PostgreSQL pgvector database store implementation.

This class provides an interface for storing and searching vectors using PostgreSQL with the pgvector extension. It implements the VDBStoreBase interface to provide a unified API for vector storage operations.

In PostgreSQL with pgvector, we use a table to store the vectors and metadata, including the document ID, chunk ID, content, and custom payload as JSONB.

This implementation uses the pgvector JDBC driver (com.pgvector:pgvector). To use this class, add the following dependencies:

`<dependency>
<groupId>org.postgresql</groupId>
<artifactId>postgresql</artifactId>
<version>42.7.4</version>
</dependency>
<dependency>
<groupId>com.pgvector</groupId>
<artifactId>pgvector</artifactId>
<version>0.1.6</version>
</dependency>
`

Prerequisites:

• PostgreSQL 11+ with pgvector extension installed

• Run: CREATE EXTENSION IF NOT EXISTS vector;

Example usage:

`// Using builder with full configuration (recommended)
try (PgVectorStore store = PgVectorStore.builder()
.jdbcUrl("jdbc:postgresql://localhost:5432/mydb")
.username("postgres")
.password("password")
.schema("my_schema") // optional, defaults to "public"
.tableName("embeddings")
.dimensions(1024)
.distanceType(PgVectorStore.DistanceType.COSINE)
.build()) {
store.add(documents).block();
List<Document> results = store.search(searchDto).block();
}

// Using static factory method (simpler, uses defaults)
try (PgVectorStore store = PgVectorStore.create(
"jdbc:postgresql://localhost:5432/mydb",
"postgres", "password", "embeddings", 1024)) {
store.add(documents).block();
List<Document> results = store.search(searchDto).block();
}
`

Note: The JDBC URL should follow PostgreSQL format:

• jdbc:postgresql://host:port/database

• jdbc:postgresql://localhost:5432/vectordb