Migrate models along inclusions of theories

[epatters]

This issue is the first in a series about interoperating models in different logics.

As the simplest case of model migration, it should be possible to push forward a model along an inclusion of one theory into another, such as the inclusion of the base theory for causal loop diagrams (CLDs) into a theory for CLDs with extra gadgets. At the data structure level in catlog, this should be as easy changing the model's pointer to its theory, but it will take more thinking and work to:

• decide whether all inclusions are allowed or only an explicit list of them
• if the former, decide when and where to enumerate the inclusions
• integrate this feature into the frontend, most likely by allow the theory selection widget to operate with a list of allowed target theories

[epatters]

My current thinking is that the inclusions should be listed explicitly in the frontend metadata because:

• we can continue to lazy load the theories as they're selected, rather than having to load them all into memory in order to check for inclusions
• a monomorphism between two theories need not be unique, so we'll have to explicitly give the blessed non-inclusion monomorphisms anyway (cf. Migrate models along monomorphisms of theories #375)