Recommendation Microservices

Note: * You will only need docker installed on your computer to run this app

Git Steps

1. Fork Branch
2. Open terminal and clone forked branch: git clone https://github.com/<YOUR USERNAME>/recommendation.git
3. Go inside point-system directory: cd recommendation
4. Add upstream repo: git remote add upstream https://github.com/fcgl/recommendation.git
5. Confirm that you have an origin and upstream repos: git remote -v

Build & Run App

This build should work for both macOS and Linux

1. Download docker for your operating system
2. From project root run the following commands:
   • Build And Run: docker-compose up --build

Health Endpoint

Confirm everything was ran correctly by going to the following endpoint: * http://localhost:5000/health/v1/marco

Toggling the User Recommendation process... Should be ran every 24 houra

TODO: Automate this so the process is automatically ran every 24 hours

Logic:

1. Endpoint user_recommendation/v1/toggle is ran every 24 hours
2. The endpoint runs a process that inserts UserRecommendation objects into the database. The UserRecommendation objects hold a list of Product Ids. The UserRecommendation Id's are referenced in the User object
   • Initially the user is given recommendations based on what's popular
   • As we get more data we will turn on our recommendation process which gives more user specific recommendations
3. When the user opens up the application an API request will be made to endpoint user_recommendation/v1 to get the products recommended to the user

Development Testing

1. Make an api request to http://localhost:5000/dev/v1/populate_data
   • This will populate your database with 2 Users and 3 Products. This is enough to test the generic version of the recommendation system (based on popularity)
2. To toggle the recommendation process make an API request to http://localhost:5000/user_recommendation/v1/toggle
3. To see a user's recommendation make an api request to http://localhost:5000/user_recommendation/v1?userId=1 (Where userId can be the id of any user in the database)