Find answers from the community

Updated 3 months ago

So the documentation on the newly

So... the documentation on the newly renamed MongoDBAtlasVectorSearch is... not terribly inclusive.
Unless I'm missing something... I do not see anything about what settings should/could be used to optimize the index properly... or whether or not it will create an index called default on it's own (because looking at the source code that's what I'm seeing)... so... how am I supposed to use this feature properly? 🤔

18 comments

LLogan M

What do you mean by optimize?

Tbh the docs on this one are missing a bit, I ran this earlier this week and discovered you have to use their UI to create the "search index" manully :PepeHands:

RRubenator

Meaning, it's got a buncha settings to pick from, and idk which I should be using:

Attachment

RRubenator

Or if it makes it's own if one with the name doesn't exist

LLogan M

ah, you'll want to use the JSON editor

RRubenator

Well, even if I did that... idk what settings to use

LLogan M

Neither do I lol would have to read mongo docs for that

I just copy/paste this from the langchain docs

Plain Text

{
  "mappings": {
    "dynamic": true,
    "fields": {
      "embedding": {
        "dimensions": 1536,
        "similarity": "cosine",
        "type": "knnVector"
      }
    }
  }
}

LLogan M

https://python.langchain.com/docs/modules/data_connection/vectorstores/integrations/mongodb_atlas

LLogan M

seems to work fine

RRubenator

I went to a link in a blog post for that but, I got a 404

RRubenator

I'll try that ty

LLogan M

but tbh, I dislike having to do this manually lol and there doesn't seem to be a way to create the search index from their python client :PepeHands:

LLogan M

which is why this feels so janky