Google AlloyDB for PostgreSQL - `AlloyDBVectorStore`

AlloyDB is a fully managed relational database service that offers high performance, seamless integration, and impressive scalability. AlloyDB is 100% compatible with PostgreSQL. Extend your database application to build AI-powered experiences leveraging AlloyDB’s LlamaIndex integrations.

This notebook goes over how to use AlloyDB for PostgreSQL to store vector embeddings with the AlloyDBVectorStore class.

Learn more about the package on GitHub.

Before you begin

To run this notebook, you will need to do the following:

• Create a Google Cloud Project
• Enable the AlloyDB API
• Create a AlloyDB cluster and instance.
• Create a AlloyDB database.
• Add a User to the database.

🦙 Library Installation

Install the integration library, llama-index-alloydb-pg, and the library for the embedding service, llama-index-embeddings-vertex.

%pip install --upgrade --quiet llama-index-alloydb-pg llama-index-embeddings-vertex llama-index-llms-vertex llama-index

Colab only: Uncomment the following cell to restart the kernel or use the button to restart the kernel. For Vertex AI Workbench you can restart the terminal using the button on top.

# # Automatically restart kernel after installs so that your environment can access the new packages
# import IPython

# app = IPython.Application.instance()
# app.kernel.do_shutdown(True)

🔐 Authentication

Authenticate to Google Cloud as the IAM user logged into this notebook in order to access your Google Cloud Project.

• If you are using Colab to run this notebook, use the cell below and continue.
• If you are using Vertex AI Workbench, check out the setup instructions here.

from google.colab import auth

auth.authenticate_user()

☁ Set Your Google Cloud Project

Set your Google Cloud project so that you can leverage Google Cloud resources within this notebook.

If you don’t know your project ID, try the following:

• Run gcloud config list.
• Run gcloud projects list.
• See the support page: Locate the project ID.

# @markdown Please fill in the value below with your Google Cloud project ID and then run the cell.

PROJECT_ID = "my-project-id"  # @param {type:"string"}

# Set the project id
!gcloud config set project {PROJECT_ID}

Basic Usage

Set AlloyDB database values

Find your database values, in the AlloyDB Instances page.

# @title Set Your Values Here { display-mode: "form" }
REGION = "us-central1"  # @param {type: "string"}
CLUSTER = "my-cluster"  # @param {type: "string"}
INSTANCE = "my-primary"  # @param {type: "string"}
DATABASE = "my-database"  # @param {type: "string"}
TABLE_NAME = "vector_store"  # @param {type: "string"}
USER = "postgres"  # @param {type: "string"}
PASSWORD = "my-password"  # @param {type: "string"}

AlloyDBEngine Connection Pool

One of the requirements and arguments to establish AlloyDB as a vector store is a AlloyDBEngine object. The AlloyDBEngine configures a connection pool to your AlloyDB database, enabling successful connections from your application and following industry best practices.

To create a AlloyDBEngine using AlloyDBEngine.from_instance() you need to provide only 5 things:

1. project_id : Project ID of the Google Cloud Project where the AlloyDB instance is located.
2. region : Region where the AlloyDB instance is located.
3. cluster: The name of the AlloyDB cluster.
4. instance : The name of the AlloyDB instance.
5. database : The name of the database to connect to on the AlloyDB instance.

By default, IAM database authentication will be used as the method of database authentication. This library uses the IAM principal belonging to the Application Default Credentials (ADC) sourced from the environment.

Optionally, built-in database authentication using a username and password to access the AlloyDB database can also be used. Just provide the optional user and password arguments to AlloyDBEngine.from_instance():

• user : Database user to use for built-in database authentication and login
• password : Database password to use for built-in database authentication and login.

Note: This tutorial demonstrates the async interface. All async methods have corresponding sync methods.

from llama_index_alloydb_pg import AlloyDBEngine

engine = await AlloyDBEngine.afrom_instance(
    project_id=PROJECT_ID,
    region=REGION,
    cluster=CLUSTER,
    instance=INSTANCE,
    database=DATABASE,
    user=USER,
    password=PASSWORD,
)

AlloyDBEngine for AlloyDB Omni

To create an AlloyDBEngine for AlloyDB Omni, you will need a connection url. AlloyDBEngine.from_connection_string first creates an async engine and then turns it into an AlloyDBEngine. Here is an example connection with the asyncpg driver:

# Replace with your own AlloyDB Omni info
OMNI_USER = "my-omni-user"
OMNI_PASSWORD = ""
OMNI_HOST = "127.0.0.1"
OMNI_PORT = "5432"
OMNI_DATABASE = "my-omni-db"

connstring = f"postgresql+asyncpg://{OMNI_USER}:{OMNI_PASSWORD}@{OMNI_HOST}:{OMNI_PORT}/{OMNI_DATABASE}"
engine = AlloyDBEngine.from_connection_string(connstring)

Initialize a table

The AlloyDBVectorStore class requires a database table. The AlloyDBEngine engine has a helper method init_vector_store_table() that can be used to create a table with the proper schema for you.

await engine.ainit_vector_store_table(
    table_name=TABLE_NAME,
    vector_size=768,  # Vector size for VertexAI model(textembedding-gecko@latest)
)

Optional Tip: 💡

You can also specify a schema name by passing schema_name wherever you pass table_name.

SCHEMA_NAME = "my_schema"

await engine.ainit_vector_store_table(
    table_name=TABLE_NAME,
    schema_name=SCHEMA_NAME,
    vector_size=768,
)

Create an embedding class instance

You can use any Llama Index embeddings model. You may need to enable Vertex AI API to use VertexTextEmbeddings. We recommend setting the embedding model’s version for production, learn more about the Text embeddings models.

# enable Vertex AI API
!gcloud services enable aiplatform.googleapis.com

from llama_index.core import Settings
from llama_index.embeddings.vertex import VertexTextEmbedding
from llama_index.llms.vertex import Vertex
import google.auth

credentials, project_id = google.auth.default()
Settings.embed_model = VertexTextEmbedding(
    model_name="textembedding-gecko@003",
    project=PROJECT_ID,
    credentials=credentials,
)

Settings.llm = Vertex(model="gemini-1.5-flash-002", project=PROJECT_ID)

Initialize a default AlloyDBVectorStore

from llama_index_alloydb_pg import AlloyDBVectorStore

vector_store = await AlloyDBVectorStore.create(
    engine=engine,
    table_name=TABLE_NAME,
    # schema_name=SCHEMA_NAME
)

Download data

!mkdir -p 'data/paul_graham/'
!wget 'https://raw.githubusercontent.com/run-llama/llama_index/main/docs/docs/examples/data/paul_graham/paul_graham_essay.txt' -O 'data/paul_graham/paul_graham_essay.txt'

Load documents

from llama_index.core import SimpleDirectoryReader

documents = SimpleDirectoryReader("./data/paul_graham").load_data()
print("Document ID:", documents[0].doc_id)

Use with VectorStoreIndex

Create an index from the vector store by using VectorStoreIndex.

Initialize Vector Store with documents

The simplest way to use a Vector Store is to load a set of documents and build an index from them using from_documents.

from llama_index.core import StorageContext, VectorStoreIndex

storage_context = StorageContext.from_defaults(vector_store=vector_store)
index = VectorStoreIndex.from_documents(
    documents, storage_context=storage_context, show_progress=True
)

Query the index

query_engine = index.as_query_engine()
response = query_engine.query("What did the author do?")
print(response)

Create a custom Vector Store

A Vector Store can take advantage of relational data to filter similarity searches.

Create a new table with custom metadata columns. You can also re-use an existing table which already has custom columns for a Document’s id, content, embedding, and/or metadata.

from llama_index_alloydb_pg import Column

# Set table name
TABLE_NAME = "vectorstore_custom"
# SCHEMA_NAME = "my_schema"

await engine.ainit_vector_store_table(
    table_name=TABLE_NAME,
    # schema_name=SCHEMA_NAME,
    vector_size=768,  # VertexAI model: textembedding-gecko@003
    metadata_columns=[Column("len", "INTEGER")],
)


# Initialize AlloyDBVectorStore
custom_store = await AlloyDBVectorStore.create(
    engine=engine,
    table_name=TABLE_NAME,
    # schema_name=SCHEMA_NAME,
    metadata_columns=["len"],
)

Add documents with metadata

Document metadata can provide the LLM and retrieval process with more information. Learn more about different approaches for extracting and adding metadata.

from llama_index.core import Document

fruits = ["apple", "pear", "orange", "strawberry", "banana", "kiwi"]
documents = [
    Document(text=fruit, metadata={"len": len(fruit)}) for fruit in fruits
]

storage_context = StorageContext.from_defaults(vector_store=custom_store)
custom_doc_index = VectorStoreIndex.from_documents(
    documents, storage_context=storage_context, show_progress=True
)

Search for documents with metadata filter

You can apply pre-filtering to the search results by specifying a filters argument

from llama_index.core.vector_stores.types import (
    MetadataFilter,
    MetadataFilters,
    FilterOperator,
)

filters = MetadataFilters(
    filters=[
        MetadataFilter(key="len", operator=FilterOperator.GT, value="5"),
    ],
)

query_engine = custom_doc_index.as_query_engine(filters=filters)
res = query_engine.query("List some fruits")
print(str(res.source_nodes[0].text))

Add a Index

Speed up vector search queries by applying a vector index. Learn more about vector indexes.

from llama_index_alloydb_pg.indexes import IVFFlatIndex

index = IVFFlatIndex()
await vector_store.aapply_vector_index(index)

The ScaNN index creation (only available in AlloyDB Omni) requires sufficient maintenance work memory. You need to set the database flag maintenance_work_mem by calling set_maintenance_work_mem before applying the index.

from llama_index_alloydb_pg.indexes import ScaNNIndex

VECTOR_SIZE = 768  # Replace with the vector size of your embedding model
index = ScaNNIndex(name="my_scann_index")
await vector_store.aset_maintenance_work_mem(index.num_leaves, VECTOR_SIZE)
await vector_store.aapply_vector_index(index)

Re-index

await vector_store.areindex()  # Re-index using default index name

Remove an index

await vector_store.adrop_vector_index()  # Delete index using default name