Compound models to simplify complex connections between parameters

[phajy]

Looking at #117 this might already be "done" but some additional documentation might really help the end user.

Specifically, it would be great to create a new model that wraps multiple old models, e.g., this example adds together two models that share parameters or have one of the parameters computed

my_new_model(m_dot, r_a, r_b, ...)
     <some definition of my_new_model parameters>
     <some code to calculate r_tr from r_a, r_b, and other parameters>
     <set some other parameters that don't need to be specified by my_new_model>
     return model_a(m_dot, r_tr, ...) + model_b(m_dot, r_tr, ...)

Could a convolution model also be embedded in this? E.g.,

return conv_model_c(conv_params)(model_a(...) + model_b(...))

I am happy to attempt to write the documentation with examples (but will likely need help in the syntax)!

[fjebaker]

I think this can be closed as of the ParameterPatch model wrapper in #172 Does that solve your use case?