predictory / predictory-app

Simple real world platform for watching movies with recommender

Predictory - app

Simple real world platform for watching movies with built-in recommender

Installation

To install dependencies, run command

$ yarn (or npm install)

Configuration

This project uses package config for project configuration. Default configuration is showed in file config/default.json. To change configuration, create file config/local-development.json and edit variables to your values. More information about config package is here.

Database setup

This project uses MySQL database and MongoDB (on Recommender API). If you want to use PostgresSQL, change mysql2 dependency in package.json to pg. Also change dialect and port in database config.

Example of database config:

"dbConfig": {
    "dialect": "mysql",
    "host": "localhost",
    "port": 3306,
    "name": "recommender",
    "user": "root",
    "password": "",
    "synchronize": false
}

Movies APIs

For downloading data about movies, the project uses OMDb API and TMDb API. Register for these two services and paste your API keys to you config file.

{
    "omdbApiKey": "###key###",
    "tmdbApiKey": "###key###"
}

Data

This project uses movielens development dataset (9/2018). If you want to download latest version of this dataset, run command

yarn data:prepare

Database preparation

yarn db:prepare

Database population

yarn db:populate

This command will take some time. Also, you need to be registered at least on OMDb API which is used to download data about movies.

Training recommender API

If you want to use recommender abilities, you need to download and install API first. With recommender API downloaded and running, run command

yarn recommender:train

to call and train API (it will take some time).

Running recommender

Server

Development

yarn server:dev

Production

yarn server:start

Client

Development

yarn client:dev

Production

yarn client:build
yarn client:start