Support serialization of state

[markaren]

Not able to test this using PyFMI as it raises a NotImplementedError.
But works with FMI4j.

Closes #54

[fcollonval]

I would separate the (de)serialization action in Python from the state getter/setter.

Should we also raise an error if the is false and the FMU user tries to call those functions?

Side question: regarding the integration test, would it be easy to set a CI job testing the capabilities through FMI4J as it covers area not available through pyfmi?

[markaren]

Side question: regarding the integration test, would it be easy to set a CI job testing the capabilities through FMI4J as it covers area not available through pyfmi?

It's possible, but keeping it up to date is not so straightforward... So I dunno. If we get FMPy up and running we could use that.

[fcollonval]

Side question: regarding the integration test, would it be easy to set a CI job testing the capabilities through FMI4J as it covers area not available through pyfmi?

It's possible, but keeping it up to date is not so straightforward... So I dunno. If we get FMPy up and running we could use that.

Thanks for the feedback. From my trial on subinterpreter, using FMPy seems really tricky. It may have something to do with how to interact with c code.

[markaren]

How about this?

[fcollonval]

Thank you for the modifications. This fits perfectly what I meant.

There are just two typing changes to revert. But otherwise this is good to go.