[Scientific & Financial Computation ] — Shawn's blog

When Dynamic Type comes to Trouble

As a dynamic programming languages , Python don't have a type on variables , as well as a arguments . This feature save us a lot workload and it enable us to focus on logical of program .

def add(x, y):
    return x+y

print add(5,6)
>> 11
print add("D","E")
>> "DE"
print add(5.0, 6)
>> 11.0

Above function is straightforward and works fine if x and y are passed with int/ float/ string .

But things may not expect as we thought, what if we pass a list parameter into this function ?

print add([5],6)
Traceback (most recent call last):
 File "C:\Users\IBM_ADMIN\workspace\euler\test_blog.py", line 4, in <module>
 print add([5],6)
 File "C:\Users\IBM_ADMIN\workspace\euler\test_blog.py", line 2, in add
 return x+y
TypeError: can only concatenate list (not "int") to list

When Decorator comes to Help

I won't take too much time explain on how decorator works . Pls refer to Bruce Eckel's post and Brian Holdefehr's post . Basically , we need a decorator to constrain the arguments passed into function, if the arguments are not consist with predefined value , we raise an exception and tell the user that parameters that won't match .

class ArgumentsCheck(object):
    def __init__(self, *argstype):
        self.argstype = argstype # store all parameters of decorater

    def __call__(self, f):
        def wrapped_f(*args):
            # test if all type of params of decoratoer equals to those of functions
            if not all([ type(a) == type(b) for a,b in zip(args, self.argstype) ]) : 
                # Not all equals, then print all type of decoratoer and those of functions
                raise Exception("Invalid Params Type for Function %s,Argument Type required (%s) but a Type Supplied (%s) " % (f.__name__,','.join([ type(_).__name__ for _ in self.argstype]),','.join([ type(_).__name__ for _ in args])) )
            else:
                # if all params type matches ,then call the funtion with params
                f(*args)
        return wrapped_f

Define Parameter types of Function by Decoration

@ArgumentsCheck(0,0)
def func1(num1, num2): #the two params must be integers
    print num1,num2

func1(1, 2) # this will be good
print "==================\n"
func1(1, 2.0) # this will fail

Output:

1 2
==================

Traceback (most recent call last):
  File "C:\Users\IBM_ADMIN\workspace\euler\test_blog.py", line 27, in <module>
    func1(1, 2.0) # this will fail
  File "C:\Users\IBM_ADMIN\workspace\euler\test_blog.py", line 11, in wrapped_f
    raise Exception("Invalid Params Type for Function %s,Argument Type required (%s) but a Type Supplied (%s) " % (f.__name__,','.join([ type(_).__name__ for _ in self.argstype]),','.join([ type(_).__name__ for _ in args])) )
Exception: Invalid Params Type for Function func1,Argument Type required (int,int) but a Type Supplied (int,float)

Conclusion:

By this decoration, we make Strong Typed  arguments call in Python, even 0 and 0.0 behaves differently in this case .