Can a local mongodb instance be used when working with langchain MongoDBAtlasVectorSearch?

Generative AI Vector Search for LLMs
,
1

I know the docs are all about using an Atlas cluster, but, just checking whether that’s actually a technical requirement. Can this be used with a local instance instead, or, do I have to have an atlas instance? TIA

2

Yes it is a requirement. Vector Search is only supported on Atlas Clusters.

That said, if you have a requirement for a Local Development solution you can use our Atlas CLI to deploy something locally, but this is exclusively for development and testing purposes.

3

Thanks for the prompt reply.

4

So, following along with the docs here: https://www.mongodb.com/docs/atlas/cli/stable/atlas-cli-deploy-docker/ with docker-compose, I’m not able to persist anything between runs. I tried adding another volume that maps the db data in a typical mongo container, but that doesn’t work here. Now, I can easily accomplish this with mongosh and scripts, but, in a docker compose environment, I have other services which are dependent on mongo, and this isn’t working. What’s the solution for persisting, if there is one? Or failing that, a work around suggestion?

5

And if you want to service which depends on other service , you could use the depends on creiteria , but as of my knowledge , it only influnces the order in which your containers are started but it doesn’t mean than your second image is started after the completion of mongo , but i’m sure that the persisting will work

6

mongo:
image: mongo
ports:
- “27017:27017”
# environment:
# MONGO_INITDB_ROOT_USERNAME: root
# MONGO_INITDB_ROOT_PASSWORD: root
volumes:
- ./your_chache_dir:/data/db