Developers often find themselves working on scripts or applications that require a high degree of customization, with various behaviors controlled by script options or application parameters. As the number of behaviors increases in proportion to the combinations of options, the complexity of the code can become overwhelming, leading to issues such as duplicated code, poor code structure, and difficulty in maintaining the codebase. Design patterns in Python can be used to enhance code quality in such development scenarios, and examples of these design patterns can be easily found online. Package Dyndesign (https://github.com/amarula/dyndesign) offers a solution with a set of tools to load and merge optional classes with base classes dynamically, also enabling safe invocation of methods or use as decorators without causing fatal errors if the optional classes are not loaded. It also provides the ability to choose between method overloading or calling all methods of the merged classes with the same name.

Dyndesign’s main objective is to adhere to SOLID principles, particularly the Single Responsibility principle and the Open-closed principle, and eliminate code duplication by

1. confining code related to each option to specific classes,
2. building option-customized class by loading classes to merge based on selected options,
3. minimizing the impact on the code when adding new options, by only requiring the addition of new modules containing the optional classes to be merged.

Dyndesign stands out from common design patterns, such as adapter, class builder, facade, etc. by simplifying ease of use, readability, and maintainability of code. Developers don’t need to manage dependency injections, separation of abstract/concrete classes, or MRO for multiple inheritance to create customizable apps. The only concern for developers using Dyndesign is building basic classes and option-specific extension classes, and merging them dynamically using the mergeclasses function based on selected options.