LangChain Community: VectorStores: Azure Cosmos DB Mongo vCore with DiskANN

[fatmelon]

Description

Add a new vector index type diskann to Azure Cosmos DB Mongo vCore vector store. Paper of DiskANN can be found here DiskANN: Fast Accurate Billion-point Nearest Neighbor Search on a Single Node.

Sample Usage

from pymongo import MongoClient

# INDEX_NAME = "izzy-test-index-2"
# NAMESPACE = "izzy_test_db.izzy_test_collection"
# DB_NAME, COLLECTION_NAME = NAMESPACE.split(".")

client: MongoClient = MongoClient(CONNECTION_STRING)
collection = client[DB_NAME][COLLECTION_NAME]

model_deployment = os.getenv(
    "OPENAI_EMBEDDINGS_DEPLOYMENT", "smart-agent-embedding-ada"
)
model_name = os.getenv("OPENAI_EMBEDDINGS_MODEL_NAME", "text-embedding-ada-002")

vectorstore = AzureCosmosDBVectorSearch.from_documents(
    docs,
    openai_embeddings,
    collection=collection,
    index_name=INDEX_NAME,
)

# Read more about these variables in detail here. https://learn.microsoft.com/en-us/azure/cosmos-db/mongodb/vcore/vector-search
maxDegree = 40
dimensions = 1536
similarity_algorithm = CosmosDBSimilarityType.COS
kind = CosmosDBVectorSearchType.VECTOR_DISKANN
lBuild = 20

vectorstore.create_index(
            dimensions=dimensions,
            similarity=similarity_algorithm,
            kind=kind ,
            max_degree=maxDegree,
            l_build=lBuild,
        )

Dependencies

No additional dependencies were added

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
langchain	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Dec 12, 2024 1:54am

[fatmelon]

@baskaryan could you help review this PR? We have some users waiting for this feature, thanks :)

[fatmelon]

@efriis could you help review this PR? We have some users waiting for this feature, thanks :)

[fatmelon]

@eyurtsev could you help review this PR? We have some users waiting for this feature, thanks :)

[fatmelon]

hi @ccurme , @vbarda , @hwchase17 could anyone review this PR? We have some users waiting for this feature, thanks :)

[fatmelon]

@isahers1 could you review this PR? I @-mention the recommended people, but they didn't come to see it ....

[fatmelon]

@baskaryan, @efriis, @eyurtsev, @ccurme, @vbarda, @hwchase17, could you help review this PR? We have some users waiting for this feature, thanks :)

[aayush3011]

Friendly ping @baskaryan for a review here.

[efriis]

Also, if you're interested in maintaining this integration without us in the loop, we'd love to get an integration package out! Future PRs against langchain would just be {docs updates, as well as registering your package in libs/packages.yml, deprecating this community integration in favor of your integration package}

Here's the guide, and if you have questions, feel free to leave them in the comments on those pages so others can see them! https://python.langchain.com/docs/contributing/how_to/integrations/

[efriis]

You should reach out to @marlenezw for help getting set up in the https://github.com/langchain-ai/langchain-azure repo if you want to host out of there!