Question

I implemented a class in python which currently inherits from dict, but really, I don’t want it to. The main reason for inheriting is so that I can use the **kwargs construct for copying the contents into an argument list.

I presume python does some sort of iteration over the dictionary, but I can’t find any documentation.

Is this possible, and, if so, how?

Code sample just to make things clearer:

   class MyThing():
       def __init__(self):
            self.dictionary = {}


   thing = MyThing()
   # code that causes thing.dictionary to be populated
   somefunc(**thing)

results in this:

TypeError: somefunc() argument after ** must be a mapping, not instance

Answer 1

I went and read the python source, it seems to use the keys() and __getitem__ methods

class A:
    def __init__(self, data):
        self.data = data
    def keys(self):
        return self.data.keys()
    def __getitem__(self, key):
        return self.data[key]

def f(alpha):
    print alpha

f(**A({'alpha': 2}))

For the curious, If you use ** on a non-dict, CPython creates a new empty dict, and then calls dict.update() to copy your object’s keys into the dict.

Its probably to best to implement this as a subclass of collections.Mapping

class A(Mapping):
    def __init__(self, data):
        self.data = data

    def __iter__(self):
        return iter(self.data)

    def __len__(self):
        return len(self.data)

    def __getitem__(self, key):
        return self.data[key]

Due to the inheritance from Mapping, pretty much all dict methods are supported. Given the error message, I think you can argue that this should qualify in all implementations of python.

I’ve tried PyPy and it accepts the first version.