访问Python类中的属性？

Question

I have the following class:

class convert_to_obj(object):
    def __init__(self, d):
        for llist in d:
            for a, b in llist.items():
                if isinstance(b, (list, tuple)):
                    setattr(self, a, [obj(x) if isinstance(x, dict) else x for x in b])
                else:
                    setattr(self, a, obj(b) if isinstance(b, dict) else b)

    def is_authenticated(self):
        username = self.username
        password = self.password
        if username and password:
            return True

I am converting a dict to obj and then trying to access the is_authenticated method, when I do the following below:

new_array = [{'username': u'rr', 'password': u'something', }]
user = convert_to_obj(new_array)
user.is_authenticated()

it returns an error saying:

'convert_to_obj' object has no attribute 'is_authenticated'

I don't know why it's doing this. Hopefully, some other eyes might be able to point out what I am doing wrong. Thanks

Answer 1

@user2357112 is right (and good catch, because I wouldn't have seen it):

DO NOT USE TABS IN PYTHON — EVER!

That said, I have a few comments about your code.

First:

class convert_to_obj(object):

is a very bad name for a class. It'd be a good name for a function, though. You should better call it, for example:

class DictObject(object):

That being said, I'd advice you to use existing tools for doing such a thing. There's a powerful one called namedtuple , in the collections module. To do your thing, you could do:

from collections import namedtuple

# Create a class that declares the interface for a given behaviour
# which will expect a set of members to be accessible
class AuthenticationMixin():
    def is_authenticated(self):
        username = self.username
        password = self.password
        # useless use of if, here you can simply do:
        # return username and password
        if username and password:
            return True
        # but if you don't, don't forget to return False below
        # to keep a proper boolean interface for the method
        return False

def convert_to_object(d): # here that'd be a good name:
    # there you create an object with all the needed members
    DictObjectBase = namedtuple('DictObjectBase', d.keys())
    # then you create a class where you mix the interface with the
    # freshly created class that will contain the members
    class DictObject(DictObjectBase, AuthenticationMixin):
        pass
    # finally you build an instance with the dict, and return it
    return DictObject(**d)

which would give:

>>> new_array = [{'username': u'rr', 'password': u'something', }]
>>> # yes here I access the first element of the array, because you want
>>> # to keep the convert_to_object() function simple. 
>>> o = convert_to_object(new_array[0])
>>> o
DictObject(password='something', username='rr')
>>> o.is_authenticated()
True

all that being more readable and easy to use.

---

NB: for a list of dicts to convert, just make:

>>> objdict_list = [convert_to_object(d) for d in new_array]
>>> objdict_list
[DictObject(password='something', username='rr')]

And if you're working with a list of pairs instead of a dict:

>>> tup_array = [('username', u'rr'), ('password', u'something')]
>>> {t[0]:t[1] for t in tup_array}
{'password': 'something', 'username': 'rr'}

So you don't need the extra leg work in the __init__() .

HTH

Answer 2

You've mixed tabs and spaces, so the is_authenticated definition is erroneously nested inside the definition of __init__ . Turn on "show whitespace" in your editor to see the problem, and run Python with the -tt flag to make it tell you when you do something like this. To fix the problem, convert the tabs to spaces; your editor most likely has a function to do this automatically.