movie-search's Introduction

Movie Search

API for searching movies by its synopsis. The movie data was gathered from Metacritic, using a web crawler. A reverse index was created based on the terms of the movies synopsis.

The Jaccard Index was used to compute a score value assigned to each document according to specific queries. For each query, a ranking of the 10 most relevant movies is returned, taking into account the terms in the synopsis of each one.

Run with docker

docker pull ghcr.io/vncsmyrnk/movie-search:latest
docker run --rm \
    -p 5000:5000 \
    movie-search:latest

Development

docker run --rm -it \
    -v "$(pwd)"/src:/var/app \
    -p 5000:5000 \
    --workdir /var/app \
    --cpus 2 \
    --name movie-search \
    python:3.9-slim bash

# Inside container
pip install -r requirements.txt
flask --app server run --host 0.0.0.0

Examples

curl -X GET http://localhost:5000/api/query?q=music%20play -s | jq .

Returns:

[
  {
    "movie": {
      "avg_score": 70,
      "description": "Six urbanites play musical beds.",
      "description_cleaned": "six urbanit play music bed",
      "movie_uri": "/movie/your-friends-neighbors/",
      "platform": "metacritic",
      "scores": [
        {
          "reviewer_name": "The A.V. Club",
          "score": "100"
        },
        {
          "reviewer_name": "Newsweek",
          "score": "90"
        },
        {
          "reviewer_name": "TV Guide Magazine",
          "score": "80"
        },
        {
          "reviewer_name": "San Francisco Examiner",
          "score": "75"
        },
        {
          "reviewer_name": "The New Republic",
          "score": "70"
        },
        {
          "reviewer_name": "Los Angeles Times",
          "score": "50"
        },
        {
          "reviewer_name": "San Francisco Chronicle",
          "score": "25"
        }
      ],
      "title": "Your Friends & Neighbors",
      "year": "1998"
    },
    "movie_id": 3304,
    "score_jaccard": 0.4
  },
  {...}
]

Recommend Projects