How to ensure Hybrid query definitely matches the inner queries

OpenSearch
1

Versions (relevant - OpenSearch/Dashboard/Server OS/Browser):
2.15 / 2.18

Describe the issue:
I have a Hybrid query which is having a multi_match and a neural query. The results individually match these queries. I need results which definitely matches the query string mentioned in the multi_match query. In my example I want to search for Open Tickets about some issue. The issue is mentioned in the neural query against which I want to do a semantic search. For the multi_match , I wangt to perform the query with string as Open. The results should ensure that whicever Service Requests match the semantic meaning ALSO have status(or some other field..hence multi_match) as Open. I do not want CLosed Service Requests in my results even though some of these might have very high relevance score from the neural search.
I have tried to nest it bool query with ‘must’. But the results do not change.

Configuration:
original request where I want Open Service Requests with clinical information

{
“_source”: {
“exclude”: [
“embeddingVector”
]
},
“query”: {
“hybrid”: {
“queries”: [
{
“multi_match”: {
“query”: “Open”
}
},
{
“neural”: {
“embeddingVector”: {
“query_text”: “clinical”,
“model_id”: “AVixxxj”,
“k”: 3
}
}
}
]
}
}
}

Request payload where Bool query is wrapped around hybrid search:
{
“_source”: {
“exclude”: [
“embeddingVector”
]
},
“query”: {
“bool” : {
“must”: [ {
“hybrid”: {
“queries”: [
{
“multi_match”: {
“query”: “Open”
}
},
{
“neural”: {
“embeddingVector”: {
“query_text”: “clinical”,
“model_id”: “AVxxxxxj”,
“k”: 3
}
}
}
]
}
}]
}
}
}

Relevant Logs or Screenshots:

2

@SuryaM Great first post! I would recommend to surround your config in code block for easier readability.

Have you tried the following, not sure if this will work for your use case but maybe will point you in the right direction:

GET my-index/_search
{
  "_source": {
    "exclude": ["passage_embedding"]
  },
  "query": {
    "bool": {
      "filter": [
        {
          "match": {
            "status": "Open"
          }
        }
      ],
      "must": [
        {
          "neural": {
            "passage_embedding": {
              "query_text": "clinical",
              "model_id": "xxxxx",
              "k": 10
            }
          }
        }
      ]
    }
  }
}