Evshved / xcom

Web Server Log Parser

This is a project to parse a web server log file and extract number of visits per page in total and unique methods and to build map or transitions between pages.

Run

First clone the repository and cd into the root of the project. Ruby 3.0.2 is used to develop.

To install dependencies execute:

$ bundle install

And to run the parser:

$ ./parser.rb data/webserver.log

Test and Lint

Rspec is utilized to write test and Rubocop to lint the code. To run tests run:

$ bundle exec rspec

and to run rubocop:

$ bundle exec rubocop

Test coverage

The code base is 100% under test coverage.

Architecture

Entry point of the project is parser.rb script. It checks if there is enough passed argument in terminal. Then, it uses FileReader class to create a file handler refer to passed argument. Finally, an instance of LogAnalyzer is created with reader and Parser and analyze method is called on it.

In order to fulfill its task, LogAnalyzer does the following steps:

• Read log file with received reader and pass lines to LinesFormatter
• Send line_formatter with lines inside to different analyzers
• Each analyzer returns array of objects with analyzed data prepared to be printed
• Each array is passed to Printer, where it's finally printed

This architecture is easy to extend because if once it's needed to add another type of analyzer, it will be needed to just add a new class with new formulas inside. Introduced models for analysis results (PageView, Transition) which know how to be printed, help eliminate dependency between printer and data it prints. Injection of dependencies to Printer and FileReader at the upper architecture level helps to achieve easy extention for other file and output formats.

To Do

The project has room for improvement:

• Add usage document to ./parser.rb command.
• Add --help command.
• Add support of different outputs
• In case it will be needed to process bigger sets of data, parallel processing might be implementd.
• Provide an option for the terminal command to select needed analyzer. In this case we can move Analyzers injections to upper architecture level.
• Encapsulate solution in a Gem.
• Store parsed data in database