AlloyDBDocumentStore

Store and retrieve documents using AlloyDB for PostgreSQL, Google Cloud's fully managed, PostgreSQL-compatible database with built-in vector search capabilities.

Key Features

• Stores documents in AlloyDB with full-text and vector search support.
• Supports exact nearest neighbor (ENN) and HNSW approximate nearest neighbor (ANN) vector search strategies.
• Supports multiple vector similarity functions: cosine similarity, inner product, and L2 distance.
• Supports IAM authentication for secure, keyless access.
• Configurable connection via public, private, or PSC IP types.

Configuration

Add Workspace-Level Integration

1. Click your profile icon and choose Settings.
2. Go to Workspace>Integrations.
3. Find the provider you want to connect and click Connect next to them.
4. Enter the API key and any other required details.
5. Click Connect. You can use this integration in pipelines and indexes in the current workspace.

Add Organization-Level Integration

1. Click your profile icon and choose Settings.
2. Go to Organization>Integrations.
3. Find the provider you want to connect and click Connect next to them.
4. Enter the API key and any other required details.
5. Click Connect. You can use this integration in pipelines and indexes in all workspaces in the current organization.

1. Set up the AlloyDB instance and create a database.
2. Set the ALLOYDB_INSTANCE_URI, ALLOYDB_USER, and ALLOYDB_PASSWORD environment variables. For instructions, see Create Secrets.
3. Configure the AlloyDBDocumentStore in your pipeline YAML with the appropriate connection parameters and embedding dimension.

Source Code

To check this component's source code, open document_store.py in the Haystack Core Integrations repository.

Usage Examples

Basic Configuration

  AlloyDBDocumentStore:
    type: haystack_integrations.document_stores.alloydb.document_store.AlloyDBDocumentStore
    init_parameters:
      instance_uri:
        type: env_var
        env_vars:
        - ALLOYDB_INSTANCE_URI
        strict: true
      user:
        type: env_var
        env_vars:
        - ALLOYDB_USER
        strict: true
      password:
        type: env_var
        env_vars:
        - ALLOYDB_PASSWORD
        strict: false
      db: my_database
      embedding_dimension: 768

Using HNSW Index for Approximate Nearest Neighbor Search

  AlloyDBDocumentStore:
    type: haystack_integrations.document_stores.alloydb.document_store.AlloyDBDocumentStore
    init_parameters:
      instance_uri:
        type: env_var
        env_vars:
        - ALLOYDB_INSTANCE_URI
        strict: true
      user:
        type: env_var
        env_vars:
        - ALLOYDB_USER
        strict: true
      password:
        type: env_var
        env_vars:
        - ALLOYDB_PASSWORD
        strict: false
      db: my_database
      embedding_dimension: 768
      search_strategy: hnsw
      vector_function: cosine_similarity

Parameters

Init Parameters

These are the parameters you can configure in Pipeline Builder:

Parameter	Type	Default	Description
instance_uri	Secret	Secret.from_env_var("ALLOYDB_INSTANCE_URI")	The AlloyDB instance URI.
user	Secret	Secret.from_env_var("ALLOYDB_USER")	The database user.
password	Secret	Secret.from_env_var("ALLOYDB_PASSWORD", strict=False)	The database password. Not required when using IAM authentication.
db	str	postgres	The name of the database to connect to.
enable_iam_auth	bool	False	Whether to use IAM authentication.
ip_type	str	PRIVATE	The IP type for the connection. One of PRIVATE, PUBLIC, or PSC.
create_extension	bool	True	Whether to create the vector extension if it doesn't exist.
schema_name	str	public	The schema name where the document table is stored.
table_name	str	haystack_documents	The name of the table to store documents in.
language	str	english	The language used for full-text search indexing.
embedding_dimension	int	768	The dimensionality of document embeddings. Must match your embedding model's output.
vector_function	str	cosine_similarity	The vector similarity function. One of cosine_similarity, inner_product, or l2_distance.
recreate_table	bool	False	Whether to drop and recreate the document table on initialization.
search_strategy	str	exact_nearest_neighbor	The vector search strategy. One of exact_nearest_neighbor or hnsw.
hnsw_recreate_index_if_exists	bool	False	Whether to recreate the HNSW index if it already exists.
hnsw_index_creation_kwargs	Optional[Dict[str, int]]	None	Additional keyword arguments for HNSW index creation (for example, m and ef_construction).
hnsw_index_name	str	haystack_hnsw_index	The name of the HNSW index.
hnsw_ef_search	Optional[int]	None	The ef_search parameter for HNSW queries. Higher values improve recall at the cost of speed.
keyword_index_name	str	haystack_keyword_index	The name of the full-text search index.

Related Information

• AlloyDBEmbeddingRetriever
• AlloyDBKeywordRetriever
• Pipeline Components