elasticsearch product recommendation

product_recommender.query

POST user_purchases/_search?
{
  
    "size": 0,
   "query": { 
    "bool": { 
      
      "filter": [ 
        { "term":  { "product_id.raw": "polaroid x1566" }}
        ]
    }
  },
    "aggs" : {
        "recommendations" : {
            "significant_terms" : { "field" : "product_id.raw" }
        }
    }
}


############ RESPONSE ##############

....

"aggregations": {
    "recommendations": {
      "doc_count": 35,
      "buckets": [
        {
          "key": "polaroid x1566",
          "doc_count": 35,
          "score": 10852.971428571427,
          "bg_count": 35
        },
        {
          "key": "polaroid 2 x 3 film",
          "doc_count": 5,
          "score": 1107.405247813411,
          "bg_count": 7
        },
        {
          "key": "101 tips for great photographs",
          "doc_count": 10,
          "score": 267.0534834623504,
          "bg_count": 116
        },
        {
          "key": "My love affair with instagram",
          "doc_count": 4,
          "score": 206.62802721088437,
          "bg_count": 24
        },...