Minor apparent bug in "error catch" for gmse_apply

[jejoenje]

Very minor apparent bug which does not seem to be affecting anything much, but points erroneous error specification in manage_target to scaring, while the latter is specified correctly.
Haven't had time to investigate specifically but the reference to input_list[41] which should produce the error message, by my count relates to manage_target in input_list.

Example code to reproduce the "issue" below:

library(GMSE)

### Baseline check:
test1 = gmse_apply(get_res = "Full", scaring = TRUE, manage_target = 1200)

### Reading manager target as vector
MANAGE_TARGET1 = c(1200, 1200)

### Specifying manager target as element of vector, works as expected:
test2 = gmse_apply(get_res = "Full", scaring = TRUE, manage_target = MANAGE_TARGET1[1])

### "Mistake" not specifying manage_target correctly. Note error message relates to scaring, not manage target!
test3 = gmse_apply(get_res = "Full", scaring = TRUE, manage_target = MANAGE_TARGET1)

### Oddly, misspecifying manage_target when it is a list does not produce the same error relating to scaring:
MANAGE_TARGET2 = list(1200, 1200)
test4 = gmse_apply(get_res = "Full", scaring = TRUE, manage_target = MANAGE_TARGET2)

[bradduthie]

Thanks @jejoenje -- I'll take a look. I count the index input_list[41] to be scaring though?

[jejoenje]

Huh, I must have miscounted! Not sure what happened there... Sorry!

Still not sure whats going on re. the error relating to manage_target above, though. It appears I can give it a single TRUE or FALSE without issue, it's just that list it's choking on and blaming scaring.

As I said though, not a big issue at all, just thought I'd put it in here for reference.

[bradduthie]

Odd. I'll take a look @jejoenje -- I wonder if the vector form of manage_target is somehow screwing up the input_list by not taking the first element? And it does with list? I'm just guessing here, but would be good to find out.

[jejoenje]

Ah, yeah, it probably is that actually. I note input_list is a "straight" concatenation of the input variables, so I guess list or vector inputs where a single value is expected will mess up the order in that vector, and hence the numeric references.
This is an example of a similar issue, with another error message relating to something else:

gmse_apply(get_res = "Full", scaring = c(TRUE, FALSE), manage_target = 1000)

I guess this is OK as long as such a parameter input mistake is actually caught at the error checking stage... Not sure what would happen if not - more than likely just a crash, but I can imagine there may be a (perhaps exceptional) case where everyhing is passed and gmse_apply appears to run but just not with the expected inputs?

[bradduthie]

It might be good to write a function that just checks all relevant inputs to make sure that they are scalars? At least, we might want to check the obvious ones that should be TRUE/FALSE for the user and raise a flag if they're trying to insert something with the wrong structure. I'll add it to the TODO list for v0.7.

[jejoenje]

Sounds good @bradduthie. I will definitely help with the TODO list (sorry for being rubbish so far!).
Was also wondering if it would be possible to switch to named vectors/lists for parameter inputs, just to avoid confusion, but this may be rather a lot of rewriting and perhaps overkill.

[bradduthie]

Cheers! I guess a named vector/list switch would depend on how much would need to be refactored and changed, and how it might affect the backwards compatibility.

[bradduthie]

I think I have this one fixed (or, at least, giving an appropriate error message). I'll tag this and then close it when I push the commit.