WHY?

• Complex Queries (for e.g. Get all the articles for sports by New York Times combined with those from CNN)

​

• Handling user generated random queries (for e.g User wants movies from Brad Pitt with IMDb score >7)

​

• Apply different kind of analyzing for different kind of fields in documents

​

• Realtime addition, updation and deletion of documents

WHAT?

• ​Based on ElasticSearch
• Supports mappings and settings
• Realtime Distributed
• Supports querying documents vai Query DSL similar to ElasticSearch

{
  "movie": {
    "properties": {
      "director_name": {
        "type": "text",
        "analyzer": "n_gram",
        "search_analyzer": "standard"
      },
      "actor_names": {
        "type": "text",
        "analyzer": "n_gram",
        "search_analyzer": "standard"
      },
      "genres": {
        "type": "text"
      },
      "movie_title": {
        "type": "text",
        "analyzer": "n_gram",
        "search_analyzer": "standard"
      },
      "plot_keywords": {
        "type": "text"
      },
      "movie_imdb_link": {
        "type": "keyword",
        "index": false
      },
      "imdb_score": {
        "type": "float"
      }
    }
  }
}

Movie Recommendation System

Movie Recommendation System

• Built using React
• SPA, uses axios to make AJAX requests
• Provides simple UI to search for your favorite movie and then get recommendations

Architecture

Search Query

{"query": 
  {"bool": 
    {"should":
       [{"match": {"director_name": "Shane Black"}},
       {"match": {"plot_keywords": {"query": "human bomb", "boost": 2.0}}},
       {"match": {"actor_names": "Robert Downey Jr."}},
       {"term": {"actor_names": "Cheadle"}},
       {"match": {"genres": {"query": "Action", "boost": 2.0}}},
       {"match": {"genres": {"query": "Sci-Fi", "boost": 2.0}}}],
    "filter":
       [{"range": {"imdb_score": {"gte": 6.0}}}]
    }
  }
}

• search_analyzer

Recommendations

• Boosting

Demo