Struct PgVectorStoreBuilder 

pub struct PgVectorStoreBuilder { /* private fields */ }

Builds a PgVectorStore with connection pooling and optional auto-migration.

Example

use daimon_plugin_pgvector::{PgVectorStoreBuilder, DistanceMetric};

let store = PgVectorStoreBuilder::new("host=localhost dbname=mydb", 1536)
    .table("embeddings")
    .distance_metric(DistanceMetric::Cosine)
    .hnsw_m(16)
    .hnsw_ef_construction(64)
    .auto_migrate(true)
    .build()
    .await?;

Implementations

impl PgVectorStoreBuilder

pub fn new(connection_string: impl Into<String>, dimensions: usize) -> Self

Creates a new builder.

• connection_string: PostgreSQL connection string (e.g. "host=localhost dbname=mydb user=postgres" or "postgresql://user:pass@host/db")
• dimensions: the fixed vector dimension count (must match your embedding model)

pub fn table(self, table: impl Into<String>) -> Self

Sets the table name. Default: "daimon_vectors".

pub fn distance_metric(self, metric: DistanceMetric) -> Self

Sets the distance metric. Default: DistanceMetric::Cosine.

pub fn auto_migrate(self, enabled: bool) -> Self

Enables or disables automatic schema creation on first connection. Default: true.

When disabled, use the SQL from crate::migrations to set up the schema manually.

pub fn hnsw_m(self, m: usize) -> Self

Sets the HNSW m parameter (max connections per layer). None uses the PostgreSQL default (16).

pub fn hnsw_ef_construction(self, ef: usize) -> Self

Sets the HNSW ef_construction parameter (build-time search width). None uses the PostgreSQL default (64).

pub fn pool_size(self, size: usize) -> Self

Sets the maximum number of connections in the pool. Default: 16.

pub async fn build(self) -> Result<PgVectorStore>

Builds the PgVectorStore, optionally running migrations.