Movie Recommendation System

Movie recommendation system is a demo application made for purpose of demonstrating the use cases of Structured Query Engine. User can search for a movie using various filters provided and select a movie from search results to get recommended similar results.

Installation

• Recommended versions: Python >= 3.5.2 and Node >= 6.0.0
• Copy sample.config.json to config.json
• Use pip install -r requirements.txt to install python dependencies
• cd app/frontend && npm install

Running

• Edit config.json's field according to conveinence. query_engine_url is to be set to url of structured query engine
• In the root folder, to get the movie_metadata.csv do wget -O movie_metadata.csv https://goo.gl/YRj8dV
• Run the structured query engine
• Feed the structured query engine using python -m scripts.test_query_engine

Running: Production build

• From root folder cd app/frontend and do npm run build. This will generate a production build ready to be used.
• Run the backend using python start.py
• Go to config_server_url:config_server_port to see the app in action.
• You will always need to run npm run build each time to change something in js files or config.json

Running: Development build

• Run the backend using python start.py
• From root folder cd app/frontend and do npm start. This will open a page on localhost:3000 which will hot reloaded whenever a change is made to frontend files.

Note: You can use NVM to install versions of node.

Features

• Built using ReactJS, React Router, React Bootstrap and Tornado Web Framework
• Single page application
• Use axios promise based AJAX requests for backend communication.

Architecture

Search controller handles multi filter search requests from frontend and Recommendation controller handles gathering recommendations for a particular movie from Structured Query Engine.

Screenshot

Authors

• Amanpreet Singh @apsdehal
• Karthik Venkatesan @gamemaker007
• Simranjyot Singh Gill @simranjyotgill

Credits

We will like to thank our Search Engine Architecture course at NYU's professor Matt Doherty.

License

Apache License V2