/sp_log_parser

Primary LanguageRuby

Smart Pension - tech exercise

Setup and tests

The script requires:

  • Ruby 2.5.3
  • bundler

Before run:

$ bundle

To run the specs run

$ bundle exec rspec

To run Rubocop run:

$ bundle exec rubocop

The script

To run the script:

$ ruby ./parser webserver.log

The output will be:

------------------------
/about/2 90 visits
/contact 89 visits
/index 82 visits
/about 81 visits
/help_page/1 80 visits
/home 78 visits
------------------------
------------------------
/index 23 unique visits
/home 23 unique visits
/contact 23 unique visits
/help_page/1 23 unique visits
/about/2 22 unique visits
/about 21 unique visits
------------------------

Assumptions

  • Give the relatively small time window, some assumptions have been made:
    • The file is correctly formatted
    • The paths visited are only one level deep, as the ones present in the file (i.e: no validation for /pages/1/tags/3)
    • The input is manageable in-memory

Trade offs

  • I kept the 2 methods in the presenter separated in the public interface:
    • for more granular testing
    • to allow to be used separately

Next steps

  • Give more time, the code could be improved to provide:
    • file validation (i.e: checking the file size before processing and splitting the file if needed)
    • Introducing a Visit object, to manage large collections
    • output as a csv file
    • output as json (as part of a Gem) to be served in an API response