star unpacking for own classes

问题:

I was wondering if it's possible to use star unpacking with own classes rather than just builtins like list and tuple.

class Agent(object):
    def __init__(self, cards):
        self.cards = cards
    def __len__(self):
        return len(self.cards)
    def __iter__(self):
        return self.cards

And be able to write

agent = Agent([1,2,3,4])
myfunc(*agent)

But I get:

TypeError: visualize() argument after * must be a sequence, not Agent

Which methods do I have to implement in order to make unpacking possible?

回答1:

The exception message:

argument after * must be a sequence

should really say, argument after * must be an iterable.

Often star-unpacking is called "iterable unpacking" for this reason. See PEP 448 (Additional Unpacking Generalizations) and PEP 3132 (Extended Iterable Unpacking).

Edit: Looks like this has been fixed for python 3.5.2 and 3.6. In future it will say:

argument after * must be an iterable

---

In order to have star unpack, your class must be an iterable i.e. it must define an __iter__ that returns an iterator:

class Agent(object):
    def __init__(self, cards):
        self.cards = cards
    def __len__(self):
        return len(self.cards)
    def __iter__(self):
        return (card for card in self.cards)

then:

In [11]: a = Agent([1, 2, 3, 4])

In [12]: print(*a)  # Note: in python 2 this will print the tuple
1 2 3 4