Databricks Vector Search

Databricks Vector Search is a serverless similarity search engine that allows you to store a vector representation of your data, including metadata, in a vector database. With Vector Search, you can create auto-updating vector search indexes from Delta tables managed by Unity Catalog and query them with a simple API to return the most similar vectors.

In the walkthrough, we’ll demo the SelfQueryRetriever with a Databricks Vector Search.

create Databricks vector store index

First we’ll want to create a databricks vector store index and seed it with some data. We’ve created a small demo set of documents that contain summaries of movies.

Note: The self-query retriever requires you to have lark installed (pip install lark) along with integration-specific requirements.

%pip install -qU  langchain-core databricks-vectorsearch langchain-openai tiktoken

Note: you may need to restart the kernel to use updated packages.

We want to use OpenAIEmbeddings so we have to get the OpenAI API Key.

import getpass
import os

if "OPENAI_API_KEY" not in os.environ:
    os.environ["OPENAI_API_KEY"] = getpass.getpass("OpenAI API Key:")
databricks_host = getpass.getpass("Databricks host:")
databricks_token = getpass.getpass("Databricks token:")

OpenAI API Key: ········
Databricks host: ········
Databricks token: ········

from databricks.vector_search.client import VectorSearchClient
from langchain_openai import OpenAIEmbeddings

embeddings = OpenAIEmbeddings()
emb_dim = len(embeddings.embed_query("hello"))

vector_search_endpoint_name = "vector_search_demo_endpoint"


vsc = VectorSearchClient(
    workspace_url=databricks_host, personal_access_token=databricks_token
)
vsc.create_endpoint(name=vector_search_endpoint_name, endpoint_type="STANDARD")

[NOTICE] Using a Personal Authentication Token (PAT). Recommended for development only. For improved performance, please use Service Principal based authentication. To disable this message, pass disable_notice=True to VectorSearchClient().

index_name = "udhay_demo.10x.demo_index"

index = vsc.create_direct_access_index(
    endpoint_name=vector_search_endpoint_name,
    index_name=index_name,
    primary_key="id",
    embedding_dimension=emb_dim,
    embedding_vector_column="text_vector",
    schema={
        "id": "string",
        "page_content": "string",
        "year": "int",
        "rating": "float",
        "genre": "string",
        "text_vector": "array<float>",
    },
)

index.describe()

index = vsc.get_index(endpoint_name=vector_search_endpoint_name, index_name=index_name)

index.describe()

from langchain_core.documents import Document

docs = [
    Document(
        page_content="A bunch of scientists bring back dinosaurs and mayhem breaks loose",
        metadata={"id": 1, "year": 1993, "rating": 7.7, "genre": "action"},
    ),
    Document(
        page_content="Leo DiCaprio gets lost in a dream within a dream within a dream within a ...",
        metadata={"id": 2, "year": 2010, "genre": "thriller", "rating": 8.2},
    ),
    Document(
        page_content="A bunch of normal-sized women are supremely wholesome and some men pine after them",
        metadata={"id": 3, "year": 2019, "rating": 8.3, "genre": "drama"},
    ),
    Document(
        page_content="Three men walk into the Zone, three men walk out of the Zone",
        metadata={"id": 4, "year": 1979, "rating": 9.9, "genre": "science fiction"},
    ),
    Document(
        page_content="A psychologist / detective gets lost in a series of dreams within dreams within dreams and Inception reused the idea",
        metadata={"id": 5, "year": 2006, "genre": "thriller", "rating": 9.0},
    ),
    Document(
        page_content="Toys come alive and have a blast doing so",
        metadata={"id": 6, "year": 1995, "genre": "animated", "rating": 9.3},
    ),
]

from langchain_community.vectorstores import DatabricksVectorSearch

vector_store = DatabricksVectorSearch(
    index,
    text_column="page_content",
    embedding=embeddings,
    columns=["year", "rating", "genre"],
)

vector_store.add_documents(docs)

Creating our self-querying retriever

Now we can instantiate our retriever. To do this we’ll need to provide some information upfront about the metadata fields that our documents support and a short description of the document contents.

from langchain.chains.query_constructor.schema import AttributeInfo
from langchain.retrievers.self_query.base import SelfQueryRetriever
from langchain_openai import OpenAI

metadata_field_info = [
    AttributeInfo(
        name="genre",
        description="The genre of the movie",
        type="string",
    ),
    AttributeInfo(
        name="year",
        description="The year the movie was released",
        type="integer",
    ),
    AttributeInfo(
        name="rating", description="A 1-10 rating for the movie", type="float"
    ),
]
document_content_description = "Brief summary of a movie"
llm = OpenAI(temperature=0)
retriever = SelfQueryRetriever.from_llm(
    llm, vector_store, document_content_description, metadata_field_info, verbose=True
)

Test it out

And now we can try actually using our retriever!

# This example only specifies a relevant query
retriever.invoke("What are some movies about dinosaurs")

[Document(page_content='A bunch of scientists bring back dinosaurs and mayhem breaks loose', metadata={'year': 1993.0, 'rating': 7.7, 'genre': 'action', 'id': 1.0}),
 Document(page_content='Toys come alive and have a blast doing so', metadata={'year': 1995.0, 'rating': 9.3, 'genre': 'animated', 'id': 6.0}),
 Document(page_content='Three men walk into the Zone, three men walk out of the Zone', metadata={'year': 1979.0, 'rating': 9.9, 'genre': 'science fiction', 'id': 4.0}),
 Document(page_content='A psychologist / detective gets lost in a series of dreams within dreams within dreams and Inception reused the idea', metadata={'year': 2006.0, 'rating': 9.0, 'genre': 'thriller', 'id': 5.0})]

# This example specifies a filter
retriever.invoke("What are some highly rated movies (above 9)?")

[Document(page_content='Toys come alive and have a blast doing so', metadata={'year': 1995.0, 'rating': 9.3, 'genre': 'animated', 'id': 6.0}),
 Document(page_content='Three men walk into the Zone, three men walk out of the Zone', metadata={'year': 1979.0, 'rating': 9.9, 'genre': 'science fiction', 'id': 4.0})]

# This example specifies both a relevant query and a filter
retriever.invoke("What are the thriller movies that are highly rated?")

[Document(page_content='A psychologist / detective gets lost in a series of dreams within dreams within dreams and Inception reused the idea', metadata={'year': 2006.0, 'rating': 9.0, 'genre': 'thriller', 'id': 5.0}),
 Document(page_content='Leo DiCaprio gets lost in a dream within a dream within a dream within a ...', metadata={'year': 2010.0, 'rating': 8.2, 'genre': 'thriller', 'id': 2.0})]

# This example specifies a query and composite filter
retriever.invoke(
    "What's a movie after 1990 but before 2005 that's all about dinosaurs, \
    and preferably has a lot of action"
)

[Document(page_content='A bunch of scientists bring back dinosaurs and mayhem breaks loose', metadata={'year': 1993.0, 'rating': 7.7, 'genre': 'action', 'id': 1.0})]

Filter k

We can also use the self query retriever to specify k: the number of documents to fetch.

We can do this by passing enable_limit=True to the constructor.

Filter k

We can also use the self query retriever to specify k: the number of documents to fetch.

We can do this by passing enable_limit=True to the constructor.

retriever = SelfQueryRetriever.from_llm(
    llm,
    vector_store,
    document_content_description,
    metadata_field_info,
    verbose=True,
    enable_limit=True,
)

retriever.invoke("What are two movies about dinosaurs?")