Feature request: Support for MFront parameters

[chleh]

First of all thanks for this library, based on the unit tests we could implement MFront support into our FEM code in seven days, including tests. Especially your unit tests were really helpful.

I recognized that Behaviour::parameters is always empty. There is some code to parse the parameters from MFront behaviour libraries https://github.com/thelfer/MFrontGenericInterfaceSupport/blob/9b0884a042c288d4d6ac55db458ec851dc1b1b89/src/Behaviour.cxx#L286 However, it's commented and seems somewhat outdated. Are you planning to implement support for parameters in the near future?

[thelfer]

Hi @chleh, Thanks for using MGIS ! This library is still young, so it does have a few shortcomings, like parameters' support. I'll implement this rapidely. Beware that, do to the support of the stored and disspated energies, you must use the development branches of TFEL to use the development branches of MGIS. Do no hesitate to make other feed-backs or feature requests. Best, Thomas

[thelfer]

Initial support for parameters has been merged in the master branch.

The Behaviour class now has the following members:

• params: list of real parameters
• iparams: list of integer parameters
• usparams: list of unsigned short parameters

The parameters' values can be changed through the setParameters functions in C++.

Appropriate bindings have been made for python.

What is missing:

• ability to get the default value of a parameter
• ability to get the current value of a parameter
• ability to get the bounds of a parameter
• C bindings.
• comprehensive tests
• backport all of this in the rliv-1.0 branch

[chleh]

That was really fast, thanks. We'll test that. For me the issue is solved now.

[thelfer]

Just add tests on parameters.

[thelfer]

Bindings and tests are now ok. I'l merge master and rliv-1.0 branches and close the issue.