Analyser: add a strict mode

[agarny]

Right now, if a variable has an initial value (i.e. its initial_value attribute is set) then it can end up being categorised as either a constant, a state variable, or an NLA variable.

When such a variable is computed using an ODE then the analyser will consider this variable as being a state variable. If the variable is used in a different context then the analyser may consider the variable as being a constant or an NLA variable. And, sometimes, the classification may not be what the user would have expected (although still valid).

The analysis can be better if people use an equation to initialise a constant. Still, if a variable has an initial value then the analyser may still end up categorising it as a constant or an NLA variable (if it is not computed using an ODE).

So, it might be interesting if we could use the analyser in a strict mode where a variable that has an initial value could only be categorised as either a state variable or an NLA variable, and never a constant. This would constrain the analysis and might help with some models.

@RLee64, this is something that you might want to look into? In the meantime, could you please upload the CellML file we discussed this morning?

[RLee64]

Sounds like something that'd be interesting to have, I'll keep it in mind while working on #1354

Can't seem to upload cellml files so I've attached it as an xml instead
simple_capillary_causal1.xml

[agarny]

Sounds like something that'd be interesting to have, I'll keep it in mind while working on #1354

Thanks, that would be good.

Can't seem to upload cellml files so I've attached it as an xml instead simple_capillary_causal1.xml

Yes, that's "normal" and you did the right thing by renaming the extension.