Maintenance of Indirect/Inelastic interfaces

[robertapplin]

Description

The Indirect/Inelastic interfaces are among the oldest interfaces in the Mantid codebase. They've accumulated a lot of maintenance needs over time and require significant restructuring to make them easier to maintain and expand. There are several problems with the way these interfaces are set up currently. For example:

• The Qt Framework is pervasive throughout the code, which makes it hard to debug and test code.
• The logic isn't organized into smaller, testable modules/functions.
• The use of multiple layers of base classes makes the current structure unclear what each class is meant to do. This has also lead to a large amount of unused code in the base classes.
• Commonly used functions are attached to these base classes, when they would be better off being in their own independent module (not within a class at all).

As a result we are lacking in automated tests for many of the interfaces. This issue has been created to track our progress in improving the code behind the Indirect/Inelastic interfaces. Our goal is to make them more maintainable and easier to extend, which should also improve their longevity, and make them more testable.

Plan

We want to actively refactor the following Indirect/Inelastic interfaces:

• Inelastic QENS Fitting interface refactor #37230
• Inelastic Data Manipulation interface refactor #36609
• Inelastic Bayes Fitting interface refactor #37295
• Inelastic Corrections interface refactor #37296
• Indirect Data Reduction interface refactor #35912
• Indirect Diffraction interface refactor #37231

It is currently believed that refactoring the following Indirect/Inelastic interfaces would not provide enough benefit to justify the time spent refactoring the code:

• Indirect Simulations
• Indirect Tools