Make PySD Created Variables Private

[JamesPHoughton]

As @kaetze66 pointed out in kaetze66@c84030b, sometimes in the processing of a variable we create a new variable that didn't exist before. This may create the possibility for namespace conflicts:

For example, a stock like:

accumulation= INTEG (rate, 0)
	~	
	~		|

Has a new object created which is the stateful object, which is called each time you access a stock variable.

def accumulation():
    return integ_accumulation()

integ_accumulation = functions.Integ(lambda: rate(), lambda: 0)

Or, the INITIAL function:

StockA Initial Value=
	INITIAL(InflowA)
	~	
	~		|

Which becomes:

initial_inflowa = functions.Initial(lambda: inflowa())

def stocka_initial_value():
    return initial_inflowa()

if the model already defines a variable named initial_inflowa (which is a completely reasonable thing to do), I'm not sure if the function that creates pysafe names will properly account for this in the namespace - its been a while since I looked at that part of the code.

We need to put together a test model to try this explicitly (probably based on @kaetze66's ).

One way to fix it may be to make the created variables 'private' - by listing underscores before their names. we know no model variable will have that syntax.

[kaetze66]

Just a quick update. I have been testing for a while now the leading underscore (for delay, smooth, trend, initial) for the model database and I haven't run into any of these problems anymore. The model database is limited to .mdl files currently so xmile might be different. But based on the answer from @alexprey it seems like xmile identifiers cannot begin with an underscore which in my opinion would solve the problem or at least make it highly unlikely to occur.

[alexprey]

Hi, @kaetze66, solution with adding underscores is fine, I'm already done this thing in my experimental branch and it works fine.