HAVE YOU READ THIS BOOK?(H.Y.R.T.B)

Book recommendation platform automatically update results based on user interaction. This project suggest a way to connect big data processing environment and commercial webapp platform

Prerequisites

• You need to prepare docker swarm environment to laucn this platform.
• The sample data to be used in this projects are prepared in mongo directory. You can get original data from here.

Getting Started

• You can start this project in develop mode with docker-compose up in development mode
• To start this porject in production mode you need to prepare few more docker settings.
• First you need to create docker swarm overlay network with docker network create --attachable -d overlay mongo and docker network create --attachable -d overlay sparknetwork
• Second you need to add label to each node.

Mongo data nodes

docker node update --label-add data1=true prod-mongodata-1
docker node update --label-add data2=true prod-mongodata-2
docker node update --label-add data3=true prod-mongodata-3

Mongo config nodes

docker node update --label-add cfg1=true prod-mongocfg-1
docker node update --label-add cfg2=true prod-mongocfg-2
docker node update --label-add cfg3=true prod-mongocfg-3

Mongos nodes

docker node update --label-add mongos1=true prod-mongos-1
docker node update --label-add mongos2=true prod-mongos-2

• Finally you can deploy your fully scalable service with docker stack deploy -c docker-compose-prod.yml up commands.

• If you want to login as a user in README file you can login with

ID:[email protected]
PASSWORD:test1234

Platform composition

• Webapplication built by Flask frameworks.

• Standalone Spark Cluster have been used to handle and extract useful data from databases and utilize in webapp.

• Handle async and scheduler job with Celery Beat and Celery Worker.

• Compose MongoDB Cluster to form a high-available and scalable db to fully utilize big data.

Platform main functions

• This platform offer 3 main ways to recommend books to you.

  • Book recieved best reviews from users.
  
  • Most close book that recieved most similar review point to last review based on cosine similarity
  
  • Recommended books with collaborative filtering
  

• All of recommendations are updated with user interaction with webapp

Running the tests

• This platform is built on automated test environment with python unittest. In development mode you can test your application with docker exec -it webapp python3 tests.py

License

React is MIT licensed.