Alright, so I have been working on an algorithm with lots of variables and lots of repeated function calls, so I put all of my functions into a module and everything works just fine. As I’m nearing the publishing of a paper on this algorithm, I’m trying to do some clean up to make things look nicer and perhaps be less confusing for people who may end up using the algorithm.
The question I have, is this, is there a way to keep my functions in a module, but not have to pass every single variable into the function every time I use it? Let me provide an example.
In the REPL I can make a function that only takes 1 variable as the input, but uses another variable without a problem.
julia> f = X -> X+Y #9 (generic function with 1 method) julia> Y=7; julia> f(3) 10
I have numerous functions that are beyond my ability to craft in a single line but I can also use:
julia> function ff(X) X+Y end ff (generic function with 1 method) julia> Y=7; julia> ff(3) 10
Now, I’ve got a bunch of different functions, and I don’t want them cluttering the file where the main algorithm is constructed, so I put the function into a module.
module testlib export ff function ff(X) X+Y end end
So I load up my module to the REPL, but it no longer works. I think this what you call is a scope problem.
julia> using testlib julia> Y=7; julia> ff(3) ERROR: UndefVarError: Y not defined Stacktrace:  ff(::Int64) at C:\Users\#######\Documents\GitHub\####\testlib.jl:5  top-level scope at none:0
I have gotten around this by writing the functions in the module so that they take all of the variables as arguments. i.e.
module testlib export ff function ff(X,Y) X+Y end end
This works just fine, but is not ascetically pleasing when you have massive amounts of arguments getting passed into functions. I have gone so far as to write data structures to pass these variables around, but its still unwieldy.
Ultimately, I’d like to know, is there a more proper way for me to do this?