CODEWITHSUNDEEP
pythonjsonelasticsearchelastic-stack
Shyam
Shyam

Reputation: 11

Elasticsearch field boosting with knn search

I want to implement a knn search in elastic on a particular field (on an ipython notebook), and my data also contains ratings (from 1 to 5), which I also want to be taken into consideration, and hence i am implementing a functional boost as follows:

query = {
    "field" : "DescriptionVector",
    "query_vector": vector_of_input_keyword,
    "k": 2,
    "num_candidates": 500,
    "boosts": [
        {
            "type": "functional",
            "operation": "multiply",
            "factor": 0.1,
            "function": {
                "field": "Ratings"
            }
        }
    ]
}

res = es.search(index="all_products", knn=query, _source=["ProductName", "Description"])

but, the above code gives an error: BadRequestError(400, 'x_content_parse_exception', '[1:16260] [knn] unknown field [boosts] did you mean [boost]?')

Now I know there is an error in syntax but cannot figure out what it is, could someone help me out, thanks!

Upvotes: 1

Views: 444

Answers (1)

Iulia Feroli
Iulia Feroli

Reputation: 46

You can combine knn with regular queries for hybrid search, but you would separate the two. See the docs here; and this example:

"query": {
  "match": {
    "title": {
    "query": "mountain lake",
    "boost": 0.9
    }
  }
},
"knn": {
  "field": "image-vector",
  "query_vector": [54, 10, -2],
  "k": 5,
  "num_candidates": 50,
  "boost": 0.1
},
"size": 10

Hope this helps!

Upvotes: 1

Related Questions