将None而不是Series传递给函数时出错

Question

I have a module fi with the following classes defined:

class CashFlows(Asset):

   def __init__(self, amounts : pandas.Series, probabilities : pandas.Series = None):
       amounts = Asset.to_cash_flows()

       if probabilities is None:
           probabilities = pandas.Series([1]*len(amounts), index=amounts.index)

I then have another class Bond(fi.Asset) with this method within it, which is where CashFlows is being called:

def to_cash_flows(self, notional : float = 100.0) -> fi.asset.CashFlows:
    #Bunch of stuff here which is not related
    series = pandas.Series(list_of_data, indices_of_data)
    return fi.CashFlows(series)

I get the error obiwan.ObiwanError:__init__(probabilities) is None but should be <class 'pandas.core.series.Series'> , but I thought I was addressing the possibility of such an Error with the if statement. obiwan is just a Python type-checking package that I am using during Unit-Testing.

Thank You

Answer 1

According to the obiwan PyPI page :

You can specify alternative constraint types using sets:

def example5(x: {int,float}):
    ...

Taking that at face value, it would seem you should replace:

probabilities : pandas.Series = None

with:

probabilities : {pandas.Series, type(None)} = None

This will tell obiwan that None is an acceptable value for the parameter.

obiwan is not actually looking into the function itself and doesn't care about the if statement that compares the parameter to None. It is only looking at the function declaration.