using rob rix's madness to parse lowercase/numeric tuples cuz that's what we do when we learn

tupleparse.swift

func flatten(x: [String]) -> String {
    return reduce(x, "", +)
}

let x = "(1,2,(3,4,(5, potato,7)),8,(9))"

enum Type: Printable {
    case Element(String)
    case Tuple([Type])
    
    var description: String {
        switch self {
        case .Element(let x): return toString(x)
        case .Tuple(let types): return toString(types)
        }
    }
}

let element =           ignore((%" ")*)                                     // any leading whitespace
                        ++ ((%("0"..."9") | %("a"..."z"))+ --> flatten)     // 1 or more elements
                        ++ ignore((%",")* ++ (%" ")*)                       // optional comma/trailing whitespace
                        --> { Type.Element($0) }

let fixed = fix { tuple -> Parser<Type>.Function in
    let tupleElement = tuple ++ ignore(%(","))*                                     // recursive pattern + comma
    let tupleOrNumber = (tupleElement | element)* --> { Type.Tuple($0) }            // array of tuple elements or singleton array of other element
    
    let p = ignore(%"(") ++ tupleOrNumber ++ ignore(%")")
    
    return p
}

println(parse(fixed, x)) // Optional([1, 2, [3, 4, [5, potato, 7]], 8, [9]])

Another neat library for exploring this type of parsing is https://github.com/brotchie/Parcoa if anyone is interested.

For the sake of making the URL available, https://github.com/robrix/Madness is the one referenced at top.