Azure Cosmos DB create IVF index error cosmosSearchOptions code 197 InvalidIndexSpecificationOption

I am following this tutorial about taking advantage of Azure Cosmos DB for Mongo DB vCore's vector similarity search functionality. To do so, I created a Cosmos DB resource using "Try Azure Cosmos DB" with a resource group located in East US.

I connected to the database using this connection string:

import urllib 
import pymongo
COSMOS_MONGO_USER = 'cosmosrgeastus3xxxxxxxxxxxxxxxxxxxxxb'
COSMOS_MONGO_PWD = 'zxxxxxxxxxxxxxxxxxxxxxxxxxxxxx='
COSMOS_MONGO_SERVER = 'cosmosrgeastus318282c5-ac03-48af-82f4db.mongo.cosmos.azure.com'
COSMOS_MONGO_PORT = '10255'

mongo_conn = "mongodb://"+urllib.parse.quote(COSMOS_MONGO_USER)+":"+urllib.parse.quote(COSMOS_MONGO_PWD)+"@"+COSMOS_MONGO_SERVER+':'+COSMOS_MONGO_PORT+"?ssl=true&replicaSet=globaldb&retrywrites=false&maxIdleTimeMS=120000&appName=@cosmosrgeastus318282c5-ac03-48af-82f4db@"

mongo_client = pymongo.MongoClient(mongo_conn)

Despite a warning ("You appear to be connected to a CosmosDB cluster"), the client seems to be created successfully.

Note: According to the tutorial, the connection string is supposed to be

mongo_conn = "mongodb+srv://"+urllib.parse.quote(COSMOS_MONGO_USER)+":"+urllib.parse.quote(COSMOS_MONGO_PWD)+"@"+COSMOS_MONGO_SERVER+"?tls=true&authMechanism=SCRAM-SHA-256&retrywrites=false&maxIdleTimeMS=120000"

However, using that raises an exception "ConfigurationError: The DNS query name does not exist: _mongodb._tcp.cosmosrgeastus318282c5-ac03-48af-82f4db.mongo.cosmos.azure.com." That is why I changed it to the actual connection string provided by the Azure CosmosDB resource alongside the user, password and server values.

Then I created a database and a collection

# create a database called TutorialDB
db = mongo_client['TutorialDB']

# Create collection if it doesn't exist
COLLECTION_NAME = "CarrierManualCollection"

collection = db[COLLECTION_NAME]

if COLLECTION_NAME not in db.list_collection_names():
    # Creates a unsharded collection that uses the DBs shared throughput
    db.create_collection(COLLECTION_NAME)
    print("Created collection '{}'.\n".format(COLLECTION_NAME))
else:
    print("Using collection: '{}'.\n".format(COLLECTION_NAME))

Which results as expected printing Created collection 'CarrierManualCollection'.

Then, I try to create an IVF index, since "IVF is supported on all cluster tiers, including the free tier".

db.command({
  'createIndexes': COLLECTION_NAME,
  'indexes': [
    {
      'name': 'VectorSearchIndex',
      'key': {
        "contentVector": "cosmosSearch"
      },
      'cosmosSearchOptions': {
        'kind': 'vector-ivf',
        'numLists': 1,
        'similarity': 'COS',
        'dimensions': 1536
      }
    }
  ]
})

But I got this error message:

OperationFailure: cosmosSearchOptions, full error: {'ok': 0.0, 'errmsg': 'cosmosSearchOptions', 'code': 197, 'codeName': 'InvalidIndexSpecificationOption'}

The expected behavior is to get a success message that allows me to continue with the tutorial adding data to the collection.

What am I missing?

microsoft / AzureDataRetrievalAugmentedGenerationSamples

Azure Cosmos DB create IVF index error cosmosSearchOptions code 197 InvalidIndexSpecificationOption #40