How do we get the similarity score along with results in Atlas Vector Search?

MongoDB Atlas Vector Search
, ,
1

Hi,

I am planning to utilize Atlas Vector Search for my product, and I believe it could greatly benefit my project. However, I may require some assistance to ensure it aligns with my goals.
I referred to this documentation and I am able to get the results.
However, just getting the semantically similar results would not suffice my goals. I would also need the similarity score along with each matching result.
Do I need to make changes in my python script as well as the vector search index?

Any guidance or support you could provide in this regard would be greatly appreciated.

2

Hi @Jayakrishna_Manokaran - welcome to the MongoDB community forums!

You need to add an extra stage to your aggregation pipeline, to add in a field containing the similarity score:

db.embedded_movies.aggregate([
  {
    "$search": {
     #  ...
      }
    }
  },
  {
    "$project": {
      "_id": 0,
      "plot": 1,
      "title": 1,
      "score": { $meta: "searchScore" }
    }
  }
])

I took this example from the MongoDB Vector Search documentation in case you find it useful.

Hope this helps!

Mark

3 Likes
3

@Mark_Smith
this approach did not work

db.embedded_movies.aggregate([
  {
    "$search": {
     #  ...
      }
    }
  },
  {
    "$project": {
      "_id": 0,
      "plot": 1,
      "title": 1,
      "score": { $meta: "searchScore" }
    }
  }
])