How to use *args in python

Question

I have a functions module that has some functions all with some common inputs, and others that are particular to them. e.g.

def func_a(time_series, window ='1D'):

def func_b(time_series, window ='1D', availability_history ): 

def func_c(time_series, window ='1D', max_lag=25, use_probability ='T'):

I am trying to run these functions in a loop as follows:

func_list = [func_a, func_b, func_c]
windows = ['1D', '5D']
params = ['', hist, (25, 'T')]

for i_func, func in enumerate(func_list):
    class_obj = class_X(A,B,func)

    for window in windows:
        args = (window, params[i_func]) # params is a list or tuple of other params for funcs e.g. 
        class_obj.run_func(args)

And in another module

class class_X(object):
    def __init__(self, a, b, func_to_run):
        self.a = a
        self.ts = b
        self.method = func_to_run

    def generate_output(self, *args):
        return self.method(self.ts, args) # time series is common and fixed for all, other params differ or change

The above code wouldn't work because I think the functions that I am calling need to be changed to make use of *argsrather than having fixed defined params.

I think *args is meant for functions where number of input params are not known, but I am trying to use it in a case where the number of input params is known, but varies across different functions in a loop.

Is there any fix for this where I don't have to modify the functions module and can still pass all the required params as a single object (e.g. list or tuple)?

EDIT-

macromoonshine's answer states I can use kwargs like this:

def generate_output(self, **kwargs):
    return self.method(self.ts, kwargs)

With this modification you can call generate_outputs() as follows:

x.generate_outputs( window ='1D', max_lag=25, use_probability ='T')

where xis an instance of your class X

Can this be enhanced so I can pass args other than time_series and window as a lookup value in a loop e.g.

x.generate_outputs( window ='1D', params[iloop])

where

params[iloop] = max_lag=25, use_probability ='T'

I tried doing this:

params = (30, "F")
x.generate_outputs( window, *params)

but get an error

TypeError: generate_output() takes 1 positional argument but 4 were given

look up keyword arguments. **kwargs

flakes
– flakes

2017-11-11 09:19:51 +00:00
Commented Nov 11, 2017 at 9:19 — flakes
– flakes, Commented Nov 11, 2017 at 9:19

clemens · Accepted Answer · 2017-11-12 09:00:15Z

1

You can use the **kwargs instead which allows arbitrary keyword parameters. This should be easier than chinging each function. You have just to modify your generate_outputs() method in your code:

def generate_output(self, **kwargs):
    return self.method(self.ts, kwargs)

With this modification you can call generate_outputs() as follows:

x.generate_outputs(time_series, window ='1D', max_lag=25, use_probability ='T')

where xis an instance of your class X.

If you want to pass the kwargs from a dict instead named parameter, you have to prefix the dictionary variable with **. The adapted code should look like this:

params = [{max_lag: 35, use_probability: 'F'}, ... ]
TS= [1,2,3,4]

for i_func, func in enumerate(func_list):
    class_obj = class_X(TS, func)
    for window in windows:
        req_args = dict(params[i_func])
        req_args['window'] = 0
        class_obj.generate_output(**req_args)

edited Nov 12, 2017 at 9:00

answered Nov 11, 2017 at 9:25

clemens

17.9k12 gold badges52 silver badges71 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

dayum Over a year ago

Thanks, how can i pass the arguments other than time_series and window as something I lookup via a list/tuple etc. e.g. x.generate_outputs(time_series, window ='1D', params [i]) where params[i] = max_lag=25, use_probability ='T'

clemens Over a year ago

Sorry, you're right. I've adapted my post for this.

Collectives™ on Stack Overflow

How to use *args in python

1 Answer 1

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related