How to convert unicode objects to normal objects in Python.

Question

I currently have a deep object, and it is all unicode (sadly).

I am to a point where a variable is either going to be a dict, or a bool. In this case, I do

if type( my_variable ) is BooleanType:

But this is not triggered because the type is actually Unicode for all values.

How do I convert this unicode object to a normal object so I can correctly read the type, without destroying the data?

Thanks!

Here is the result of print(repr(variable)). It shows the Bools as not being unicode (unlike what I first though) but still giving me troubles.

{u'forms': {u'financing': {u'view': True, u'delete': True}, u'employment': {u'view': True, u'delete': True}, u'service': {u'view': True, u'delete': True}}, u'content': {u'articles': {u'edit': True, u'add': True, u'view': True, u'delete': True}, u'slideshow': {u'edit': True, u'view': True}, u'pages': {u'edit': True, u'add': True, u'view': True, u'delete': True}}, u'people': {u'edit': True, u'sort-staff': True, u'sort-riders': True, u'add': True, u'delete': True, u'view': True}, u'events': {u'edit': True, u'add': True, u'view': True, u'delete': True}, u'settings': {u'edit': True, u'view': True}}

Answer 1

Do not use type unless you are really really sure that you want to.

In this case, you don't -- especially checking for bool, given Python's flexibility for what can be considered as boolean! For instance, what if you are given None? How about an empty string? How about []?

The solution to this problem is the use of Abstract Base Classes (ABCs), which allow you to specify exactly what an object should be able to do, instead of what type it is. The collections module comes with a bunch of these:

import collections
if isinstance( ..., collections.MutableMapping ):
    ...

This permits anything 'dictionary-like', so that you retain polymorphism. If you need more careful specification ("I want __getitem__ and __delitem__ but not necessarily __setitem__!"), you can write your own -- see the definition of the ABCs in the source of the collections module for starters.

Are you sure that you want this functionality? If you do it correctly (with ABCs), it's not an inherently bad idea, but that's not to say you should abuse it!

---

Edit: I'm not sure that you understand what Unicode is or how Python handles it. This is one of the major differences between Python 2.x and Python 3.x, which are you using?

Re-edit: Ah, ok, you are using Python 2.x and you have a dictionary with Unicode string keys. I'm not sure what you were doing that caused a problem, since Unicode strings work basically just like ordinary strings. The MutableMapping check above will work fine.

Answer 2

Maybe you should learn about the pretty print module as it makes you check things (even I agree that you put yourself in mess. This usually happens as result recursion with wrong kind of result, for example doing append, when you should do extend for list)

Here your variable contents pretty printed:

{u'content': {u'articles': {u'add': True,
                            u'delete': True,
                            u'edit': True,
                            u'view': True},
              u'pages': {u'add': True,
                         u'delete': True,
                         u'edit': True,
                         u'view': True},
              u'slideshow': {u'edit': True, u'view': True}},
 u'events': {u'add': True, u'delete': True, u'edit': True, u'view': True},
 u'forms': {u'employment': {u'delete': True, u'view': True},
            u'financing': {u'delete': True, u'view': True},
            u'service': {u'delete': True, u'view': True}},
 u'people': {u'add': True,
             u'delete': True,
             u'edit': True,
             u'sort-riders': True,
             u'sort-staff': True,
             u'view': True},
 u'settings': {u'edit': True, u'view': True}}

From this it is evident that you have True value only, not False. What is use case of these values? Why for example you are not using set: {u'add',u'delete', u'edit', u'sort-riders',...} ?

from pprint import pprint

def alternative(yourdict):
    for key in yourdict:
        if yourdict[key] is True:
            yield set(yourdict.keys())
            break
        else:
            yield tuple((key,tup) for tup in alternative(yourdict[key]))

my_variable = {u'forms': {u'financing': {u'view': True, u'delete': True}, u'employment': {u'view': True, u'delete': True}, u'service': {u'view': True, u'delete': True}}, u'content': {u'articles': {u'edit': True, u'add': True, u'view': True, u'delete': True}, u'slideshow': {u'edit': True, u'view': True}, u'pages': {u'edit': True, u'add': True, u'view': True, u'delete': True}}, u'people': {u'edit': True, u'sort-staff': True, u'sort-riders': True, u'add': True, u'delete': True, u'view': True}, u'events': {u'edit': True, u'add': True, u'view': True, u'delete': True}, u'settings': {u'edit': True, u'view': True}}
pprint(my_variable)

print 50 * '-'
print 'Alternative datastructure'
pprint(tuple(alternative(my_variable)))