[GeoMechanics] GeoMechanics with Reduced Order Models using Kratos RomApplication

[Rbravo555]

This PR implements the required changes in two python scripts of the GeoMechanics application to conform with the format expected by the RomApplication.

📝 Description
The RomApplication works with other applications in Kratos by keeping the behaviour of the specific app, while changing only the builder and solver. To do this, the format inside the scripts should match the expectations by the RomApp. For the GeoMechanics application, changes to two files were required:

• In geo_mechanics_solver.py the required functions should be named _CreateBuilderAndSolver rather than _ConstructBuilderAndSolver, and there should be a function returning the linear solver.

• In geomechanics_solvers_wrapper.py, the mechanism inside the RomApp for setting the RomBuilderAndSolver requires the existance of the CreateSolver and CreateSolverByParameters functions which did not exist as separate entitied but combined into a single function CreateSolver(model, custom_settings).

🆕 Changelog
In geo_mechanics_solver.py

• changed _ConstructBuilderAndSolver(self, param1) ==> _CreateBuilderAndSolver(self)
• added method _GetLinearSolver(self)

In geomechanics_solvers_wrapper.py

• splitted operations CreateSolver(model, custom_settings) ==> CreateSolver(model, custom_settings) + CreateSolverByParameters(model, custom_settings, parallelism)

[rfaasse]

@Rbravo555 Since I replaced @huhaas in the Kratos GeoMechanics team, I picked up this review while discussing the PR with @avdg81 and @markelov208

Thanks for making the geo solver wrapper consistent with other applications and doing a bit more clean-up! The PR looks clean, we just have a few suggestions to make the changes a bit simpler.

[rfaasse]

If possible within the limits of the framework, I would propose to rename this function to CreateSolverBySolverSettings to make it even more clear that the "solver_settings" should be passed to this function

[Rbravo555]

I agree that your proposed name is more descriptive, as it reflects reliance on solver_settings. However, this function is consistently named CreateSolverByParameters across multiple applications in Kratos, this consistency is crucial for the RomApplication.

Given the implications of renaming a function that is widely used, this change would require consideration and approval by the Technical Committee. I will certainly forward your proposal to them for review.

For the time being, we will maintain the current naming convention to ensure compatibility and consistency across applications. We appreciate your understanding and value your input in this matter.

[rfaasse]

Thank you for taking it into consideration and explaining the situation, then indeed it is out of scope to change the name in this PR. Also, thank you for implementing the other suggestions, the PR looks ready for merging now from my point of view!

[rfaasse]

As mentioned, thank you for processing the suggestions, this PR is ready for merge from my point of view