Module 5: Object-oriented design

In this module you will benefit from the automatic testing that you have added in the last module. Let’s explore some other object-oriented designs for our code base…

• Create new branch “oop-dataclasses” (Make sure changes from last module have been merged, and that you start from the main branch)
• 5.1 Type Hints
  • Add type hints to all functions and methods. Commit
• 5.2 Data class
  • Make all the cleaner classes dataclasses.

  from dataclasses import dataclass
  
  @dataclass
  class SpikeCleaner:
      ...

  • remove the __init__ method (not needed anymore)
  • Check that the notebook still runs and that the classes indeed work as data classes (e.g. have a string representation and support equality testing etc)
  • Commit
• 5.3 Module level function
  • Make a private module function _print_stats() that prints the number of cleaned values
  • call the function from each of the clean methods (note: inheritance is not required to obtain common functionality)
• 5.4 Composition or inheritance
  • Create a new cleaner class called CleanerWorkflow that takes a list of cleaners when constructed and has a clean method that run all the cleaners’ clean methods.

  class CleanerWorkflow:
      def __init__(self, cleaners) -> None:
          self.cleaners = cleaners
  
      def clean(self, data: pd.Series) -> pd.Series:
          data_cleaned = data.copy()
          for cleaner in self.cleaners:
          ...

  • Modify the notebook to use the CleanerWorkflow instead of looping over the cleaners
  • Consider what type of validation you would want CleanerWorkflow to have? Is it better check validity up front or to just go ahead and handle problems afterwards?
  • Consider whether it would be better to create a base class BaseCleaner - write down your considerations as a comment in the pull request, refer to specific lines of code
  • e.g. how would you handle e.g. common plotting functionality in the cleaner classes?
  • Create pull request in GitHub and “request review” from your reviewers
  • Get feedback, Adjust code until approval, then merge (and delete branch)

Back to course project overview