A naukri.com initiative
PlanetPython
1w
394
Glyph Lefkowitz: Stop Writing `__init__` Methods
- Before the introduction of dataclasses in Python 3.7, the __init__ method was commonly used for object initialization.
- Different alternatives to using __init__ method had their own drawbacks, such as making objects mutable or complex inheritance structures.
- Having an empty __init__ method for simple object initialization was considered acceptable.
- Problems arise when __init__ is used for complex object setups, like handling I/O operations.
- The article suggests using dataclasses to define attributes and classmethods for object creation instead of __init__ for complex objects.
- Using precise types and NewType to enforce object validity is recommended for better design.
- Dataclasses with classmethods and NewType can provide more readable and maintainable code compared to manual __init__ methods.
- The recommended best practices include using dataclasses, classmethods, and NewType for defining Python classes.
- By adhering to these best practices, the code becomes more future-proof, testable, flexible, and easier to understand.
- The article emphasizes that defining a custom __init__ method is considered an anti-pattern in modern Python development.
Read Full Article
23 Likes
For uninterrupted reading, download the app