The best friend of streaming platforms.
We will need to install Ruby (3.2.0 is recommended). With Ruby installed run the command line application using the command below:
$ ./playlist_pal.rb sample-data/spotify.json sample-data/changes.json new_spotify.jsonTo validate its output run the unit tests using RSpec:
$ bundle install && rspec specs/It's also possible to Rubocop for static analysis:
$ rubocopThis app was built in just 1 hour and 59 minutes, so there are plenty of room for improvements, particularly for handling very large files.
Some ideas that could to handle this scenario:
- avoid very large files, for example split the input file according to the data type and changes file according to the operation;
- don't put the entire files on memory like I did. To solve this we could try oj or using Streams (I did this using Elixir, I don't if
ojuses streams of if there other gems that use); - adopt a map and reduce approach, splitting into other servers/processes each operation;
- use Sidekiq workers and control the number of file processing per worker, using for example sidekiq-limit. We doesn't solve the initial problem, but it can solve another problem of having multiple files to be processed;
The main idea during the development was how to create a simple, efficient and maintainable code in only 2 hours. To do that some concepts and approaches were taken:
- create unit test to ensure we are in the right path and avoid time debugging;
- create small function with meaningful names;
- failfast and with good feedback for the user;
Due the time constraint, there were certain tasks I had intended to develop, but I couldn't complete:
- create unit tests for
PlaylistUpdater. In the end I just created two integration tests to help me to validate the code; - validate the input file format deeply (e.g.: if all songs have an ID);
- validate the changes file format deeply (e.g.: all songs ids exist);
- parse/sanitize data to avoid convertions during the output generation;
- use the class
Logger, instead of,puts;