Make it possible to interpret a model (initially algebraic and ODE models)

[agarny]

In some cases (e.g., from WebAssembly), to generate some code with the view of compiling it on the fly is next to impossible at this stage (at least, using LLVM+Clang; see opencor/libopencor#223). For this reason, it would be nice to have a way to interpret a model. This could (greatly) take advantage of what has been done for the Generator, especially when it comes to which equation should be evaluated and when.