Question

I need some help with figuring out Pythons *args and **kwargs. It’s simple but I haven’t entire wrapped my head around them. Here’s one of scenarios that’s bewildering to me.

I have two functions mainfunc and wrapperfunc (which is a wrapper function for the main function). It looks like this.

def mainfunc(fname=None, lname=None):
    print 'Firstname: ' + fname
    print 'Lastname: ' + lname

def wrapperfunc(uname, *args):
    print uname
    mainfunc(*args)

I can call wrapperfunc like this:

wrapperfunc('j.doe', 'john', 'doe')

In this method, all three parameters are positional. Since j.doe comes into uname, the other two params can be accessed by *args

..but is it possible to pass some of the params to wrapperfunc from a dict so that I can still access uname inside wrapperfunc directly and then pass the remaining positional parameters to the mainfunc. Something like the following snippet:

params = {'uname':'j.doe'}
wrapperfunc(**params, 'john', 'doe')

I want to access the named parameters directly inside wrapperfunc but pass all the positional parameters to mainfunc.

Answer 1

Keyword arguments must come after position arguments in Python.

params = {'uname':'j.doe'}
wrapperfunc('john', 'doe', **params)

will pass the keyword arguments after the two positional arguments,

If you want to look at an argument, but otherwise do the call normally, do:

def wrapper(*args, **kwargs):
    print kwargs["uname"]
    return mainfunc(*args, **kwargs)

You can generalize this to work on any function you want as a decorator.

def wrapper(f):
    def wrapped(*args, **kwargs):
        print kwargs["uname"]
        return mainfunc(*args, **kwargs)
    return wrapped

@wrapper
def foo(uname="Test"):
    return uname + "bar"

# the @decorator is equivalent to `foo = wrapper(foo)`