movierama

A movie recommendation system in the form of an API. It contains the implementation of a User/User collaborative filtering algorithm to recommend movies.

Instructions

The docker-compose command will automatically bring up a postgres server, load it with data and finally bring up the movierama service.

• Compile: $ mvn clean package

• Run: $ docker-compose -f docker/docker-compose.yml up

• Swagger URL: http://localhost:8080/swagger-ui.html

Endpoints

• A user can rate a movie he viewed.

  • POST: http://localhost:8080/users/1/movies/1/ratings/5

• A user can retrieve information about the movies he viewed. The response is paginated.

  • GET: http://localhost:8080/users/1/movies/pages/1

• A user can retrieve a number of movie recommendations.

  • GET http://localhost:8080/users/1/movies/recommendations/5

NOTE: The movie descriptions are consumed from the external API: https://api.themoviedb.org.

Data

The movielens dataset has been used to populate the user profiles.

• Location: $ wget http://files.grouplens.org/datasets/movielens/ml-latest-small.zip

The ratings.csv and movies.csv have been preprocessed with the scripts located at: sql/data/scripts. These are:

• preprocess-movie-data.sh

• preprocess-rating-data.sh

The SQL statements which are produced by these two scripts and that populate the DB are located at sql/schema/DATA-[MOVIES|RATINGS].sql.

The DB schema is located at sql/schema/TABLES.sql.

All schema and data are loaded automatically with the docker-compose command.