Created Stratification DSL. Required adding getter methods for lss, …

[neonWhiteout]

…lssv, etc to StockFlow.jl, several methods to Syntax.jl (incl infer_links, which is now exported), and a Stratification.jl file. Also created a Stratification tests file and added more tests to Syntax.

[neonWhiteout]

Beyond maybe some comments and formatting, this should be fine.

[neonWhiteout]

Composition DSL has been separated out into #85

[neonWhiteout]

We may want to discuss the use of global flags, which I also use in Composition. I'm unsure what the best practice is. We might just want to delete them altogether.

[neonWhiteout]

If we want to hold off longer I can try generalizing to allow for any amount of arguments. Would need to revamp the stratification syntax, but beyond that shouldn't be too difficult.

Struggling with macros which take a variable number of arguments...

[neonWhiteout]

Have any good ideas for a syntax to generalize beyond a pullback with only 2?

figure it'd be something like [(a1, ..., an), (b1, ..., bm), ... (x1, ..., xk)] => t1, but I dunno if there's something better.

Also, converting all those cumbersome vector and dicts to a struct with all the information for each stock flow.

I've sunk a lot of time into this...

[neonWhiteout]

https://github.com/neonWhiteout/StockFlow.jl/tree/Multi_Stratification Generalized stratification. Now works for an arbitrary number of stockflows.

Significantly different from present stratification, however. Uses structs and a lot of broadcasting. Notation is as I described in the previous comment.

I'm thinking I port over the structs and some of the structure into this branch's stratify, then make a separate stratify macro for the generalized form. Currently using the name 'nstratify', because it can have 'n' inputs, but idk.

[neonWhiteout]

Little funny. When I call a macro with something like A => B, it interprets that as an expression. In a function call, it's bound to a variable. When I try getting the value of that variable inside another expression, it tries finding the variables for A and B, but it isn't defined. In effect, it's unwrapping the original expression.

As such, need to wrap it in another expression. Expr(:quote, my_var)

[neonWhiteout]

Generalized stratification. Now works for an arbitrary number of stockflows.

Please see here for both generalized stratification and a rework of Stratification. We object oriented now.

Need to write tests for generalized stratification, but all the old tests still pass (after slight modification to use the right types)

[neonWhiteout]

May want to close this in favor of #87. Started a new branch because it is quite different from this one, though it's based off of it.