search_es.py

from config import settings


def search_topN(query_vec, es, no_of_recommendation):
    """
    It takes a vector as input and returns the top 10 movies that are most similar to the input vector.

    :param query_vec: The vector of the query movie
    """

    script_query = {
        "script_score": {
            "query": {"match_all": {}},
            "script": {
                "params": {"query_vector": query_vec},
                "source": "cosineSimilarity(params.query_vector, 'embedding') + 1.0",
            },
        }
    }
    response = es.search(
        index=settings.INDEX_NAME,
        body={
            "size": no_of_recommendation,
            "query": script_query,
            "_source": ["movie_name", "genre"],
        },
    )
    return response["hits"]["hits"]


# def search_movies(query, es, no_of_content):
#     """
#     It takes a query string, connects to the Elasticsearch cluster, and returns the top 10 results

#     :param query: The query string to search for
#     """

#     script_query = {
#             "query" : {
#                 "match" : {
#                     'movie_name': query
#                 },
#             },
#         },
#     response = es.search(
#         index = settings.INDEX_NAME,
#         body = script_query,
#         size = no_of_content,
#     )
#     return response['hits']['hits']