Fix tuple parsing bug

[jasmith-hs]

In Jinjava, we parse something surrounded with parentheses either as an AstTuple or an AstNested. When there is a single element inside the parentheses, we'll parse it as AstNested, otherwise it will become an AstTuple.

This check was done by seeing if there are any commas before the first ). However, this is inadequate for two reasons:

1. We can have an expression like: #{(range(0,1))}. Clearly this shouldn't be parsed as a tuple, but this bug caused it to be.
2. We can have an expression like: #{((1),2)}. Clearly this should be parsed as a tuple, but the bug causes it to be parsed as an AstNested

The fix here simply maintains a counter of depth into parentheses or brackets, and only checks for the existence of a comma symbol when at 0 depth. We don't need to worry about matching ( to ) and [ to ] here because if they are mismatched, the correct parsing exception will get thrown down the line anyway.