[FEATURE] Inner hit of multi vector with filter

[heemin32]

Similar to #1447, I want to see inner hit of multi vector with filter.

Create KNN index

PUT /my-knn-index-1
{
  "settings": {
    "index": {
      "knn": true,
      "knn.algo_param.ef_search": 100
    }
  },
  "mappings": {
    "properties": {
      "nested_field": {
        "type": "nested",
        "properties": {
          "my_vector": {
            "type": "knn_vector",
            "dimension": 3,
            "method": {
              "name": "hnsw",
              "space_type": "l2",
              "engine": "lucene",
              "parameters": {
                "ef_construction": 100,
                "m": 16
              }
            }
          }
        }
      }
    }
  }
}

Insert Data

PUT /_bulk?refresh=true
{ "index": { "_index": "my-knn-index-1", "_id": "1" } }
{"parking": false, "nested_field":[{"my_vector":[1,1,1], "color": "yellow"},{"my_vector":[2,2,2]},{"my_vector":[3,3,3]}]}
{ "index": { "_index": "my-knn-index-1", "_id": "2" } }
{"parking": true, "nested_field":[{"my_vector":[10,10,10], "color": "blue"},{"my_vector":[20,20,20], "color": "blue"},{"my_vector":[30,30,30]}]}
{ "index": { "_index": "my-knn-index-1", "_id": "3" } }
{"parking": true, "nested_field":[{"my_vector":[100,100,100], "color": "blue"},{"my_vector":[200,200,200]},{"my_vector":[300,300,300]}]}

Query

GET /my-knn-index-1/_search
{
  "query": {
    "nested": {
      "path": "nested_field",
      "query": {
        "knn": {
          "nested_field.my_vector": {
            "vector": [
              10,
              11,
              12
            ],
            "k": 1,
            "filter": {
              "term": {
                "parking": true
              }
            }
          }
        }
      },
      "inner_hits": {
        "name": "nested_hits"
      }
    }
  }
}

Response

It return empty result for inner hit. The filter is getting applied to nested field item for inner hit even though it is for parent doc. Therefore, every nested items are filtered out.

{
  "took": 20,
  "timed_out": false,
  "_shards": {
    "total": 1,
    "successful": 1,
    "skipped": 0,
    "failed": 0
  },
  "hits": {
    "total": {
      "value": 1,
      "relation": "eq"
    },
    "max_score": 0.16666667,
    "hits": [
      {
        "_index": "my-knn-index-1",
        "_id": "2",
        "_score": 0.16666667,
        "_source": {
          "parking": true,
          "nested_field": [
            {
              "my_vector": [
                10,
                10,
                10
              ],
              "color": "blue"
            },
            {
              "my_vector": [
                20,
                20,
                20
              ],
              "color": "blue"
            },
            {
              "my_vector": [
                30,
                30,
                30
              ]
            }
          ]
        },
        "inner_hits": {
          "nested_hits": {
            "hits": {
              "total": {
                "value": 0,
                "relation": "eq"
              },
              "max_score": null,
              "hits": []
            }
          }
        }
      }
    ]
  }
}

Query with filter on nested item

When filter is applied to nested item attribute only, it returns non-empty result for inner hit. The result is not correct due to #1447 though.

{
  "query": {
    "nested": {
      "path": "nested_field",
      "query": {
        "knn": {
          "nested_field.my_vector": {
            "vector": [
              10,
              11,
              12
            ],
            "k": 2,
            "filter": {
              "term": {
                "nested_field.color": "blue"
              }
            }
          }
        }
      },
      "inner_hits": {}
    }
  }
}

Response

{
  "took": 2,
  "timed_out": false,
  "_shards": {
    "total": 1,
    "successful": 1,
    "skipped": 0,
    "failed": 0
  },
  "hits": {
    "total": {
      "value": 2,
      "relation": "eq"
    },
    "max_score": 0.16666667,
    "hits": [
      {
        "_index": "my-knn-index-1",
        "_id": "2",
        "_score": 0.16666667,
        "_source": {
          "parking": true,
          "nested_field": [
            {
              "my_vector": [
                10,
                10,
                10
              ],
              "color": "blue"
            },
            {
              "my_vector": [
                20,
                20,
                20
              ],
              "color": "blue"
            },
            {
              "my_vector": [
                30,
                30,
                30
              ]
            }
          ]
        },
        "inner_hits": {
          "nested_field": {
            "hits": {
              "total": {
                "value": 2,
                "relation": "eq"
              },
              "max_score": 0.16666667,
              "hits": [
                {
                  "_index": "my-knn-index-1",
                  "_id": "2",
                  "_nested": {
                    "field": "nested_field",
                    "offset": 0
                  },
                  "_score": 0.16666667,
                  "_source": {
                    "color": "blue",
                    "my_vector": [
                      10,
                      10,
                      10
                    ]
                  }
                },
                {
                  "_index": "my-knn-index-1",
                  "_id": "2",
                  "_nested": {
                    "field": "nested_field",
                    "offset": 1
                  },
                  "_score": 0.0040650405,
                  "_source": {
                    "color": "blue",
                    "my_vector": [
                      20,
                      20,
                      20
                    ]
                  }
                }
              ]
            }
          }
        }
      },
      {
        "_index": "my-knn-index-1",
        "_id": "3",
        "_score": 4.2076917E-5,
        "_source": {
          "parking": true,
          "nested_field": [
            {
              "my_vector": [
                100,
                100,
                100
              ],
              "color": "blue"
            },
            {
              "my_vector": [
                200,
                200,
                200
              ]
            },
            {
              "my_vector": [
                300,
                300,
                300
              ]
            }
          ]
        },
        "inner_hits": {
          "nested_field": {
            "hits": {
              "total": {
                "value": 0,
                "relation": "eq"
              },
              "max_score": null,
              "hits": []
            }
          }
        }
      }
    ]
  }
}

[heemin32]

Closing in favor of opensearch-project/OpenSearch#13903