This is a largely unfinished programme, which gives a broad indication of how I would write a programme to process a file of JSON objects separated by new lines, given terms from an API request body and serving the result in an API response.
Unfortunately I've run out of personal time to complete it, so I'm showing what I have so far!
You can clone the repo, navigate to the root directory and run sbt run on the command line to start the server locally.
In another terminal window you can run the following to make an API request:
curl -X POST http://localhost:8080/amazon/best-rated -H "Content-Type: application/json" -d '{"start": "01.01.2010","end": "31.12.2020","limit": 2,"minNumberReviews": 2}'
I've made a short video to walk you through the code: https://www.loom.com/share/4eb6bc08c025431bb705ee82a7e95725.
The main point is that the commented out code in BestRated.scala is the main thing I wanted to complete and the function
signatures for this are located in Utils. If I had more time, I would write unit tests for these functions and finish
implementing them. The main thing I am unsure about is how to implement some kind of sorting so that I can provide the
highest rated products.
- the POST request body follows JSON protocols so that it works with the http4s circe decoders
- assumed the file is not zipped
- assumed average rating meant the mean average. I also wondered how many decimal points to return.
- how can I think creatively around sorting a stream of products with the highest average rating?
- will there be any constraints on the limit value? What if the limit is very high? How will the programme manage this?
- what are the constraints on the dates given in the request? It would be good to write unit tests for all the possible date scenarios. For example, what if the end date is in the future? How will the programme respond?
- caching - what if the same request is repeated many times by the user? It's hard to know without understanding the use case for this API. Would caching by useful?