provide API function to take tunables parameter mapping and make canonicalized vector of values

[isaacsas]

The tutorial: https://docs.sciml.ai/ModelingToolkit/stable/examples/remake/ is very useful for understanding the new parameter setup, but it would be nice if it could show a way for users to specify an initial guess for the optimizer via a parameter mapping, and/or specify the lb and ub bounds via one. Right now it isn't clear how a user can figure out the tunable parameter ordering, and hence they wouldn't easily know which components of lb and ub correspond to which parameters. While this can be done by inspection of tunable_parameters(sys), it seems like a more MTK consistent workflow would be to have a varmap_to_vars equivalent that takes a mapping of tunable_parameters to values, and a parameter object (or system?), and returns a vector of the parameter values that is consistent with canonicalize and replace!.

[isaacsas]

Along those lines, it would be nice to update this tutorial to whatever the current recommended workflow is, as it is simple example where a user is likely to need to use much of the parameter API.