Is there a way to make “X() in [X(), Y(), Z()]” re

问题:

I have a custom object in this project of mine, called Page. A Page's identifying feature is its title. A Page is normally created by calling Wiki.page, Wiki.category, or Wiki.template, or by generating them from other methods like Wiki.random. (I recommend you look a little bit at what that is before going on.)

Sometimes, users of this module might want to generate some Pages and convert that generator into a normal list. After they obtain that list of Pages, they might want to check if another page they got is in that list. However, this:

>>> wp = mw_api_client.Wiki('https://en.wikipedia.org/w/api.php')
>>> wp.page('title') in [wp.page('title'),
                         wp.page('not this'),
                         wp.page('not this either')]
False

should be True, not False, because there is a page with the title "title" in it. Is there a magic method I can use to make that True? I already tried using __eq__ (for equality) and __hash__ (for hash checking) (commit), but neither seemed to do the trick. Do lists simply use identity? Or is there something else I'm missing? How do I do this properly?

回答1:

My original answer went down the wrong rabbit hole... (see the history).

It's always worth implementing a simplified version of what is breaking... see below (using these 1,2,3,4 as inspiration)

#!/usr/bin/env python3

from pprint import pprint

class Page(object):
    def __init__(self, wiki, **data):
        self.wiki = wiki
        self.title = None
        self.__dict__.update(data)

    def __eq__(self, other):
        return self.title == other.title

class Wiki(object):
    def __init__(self, api_url):
        self.api_url = api_url

    def page(self, title, **evil):
        if isinstance(title, Page):
            return title;
        return Page(self, title=title, **evil)

w = Wiki('url')
pprint(w)
pprint(w.__dict__)

p1 = w.page('testing')
pprint(p1)
pprint(p1.__dict__)

p2 = w.page('testing')
pprint(p2)
pprint(p2.__dict__)

p3 = w.page('testing something else')
pprint(p3)
pprint(p3.__dict__)

pprint(p1 == p2)
pprint(p1 == p3)
pprint(p1 in [ p2 ])
pprint(p1 in [ p2, p3 ])

Output:

<__main__.Wiki object at 0x7f2891957d30>
{'api_url': 'url'}
<__main__.Page object at 0x7f2891957dd8>
{'title': 'testing', 'wiki': <__main__.Wiki object at 0x7f2891957d30>}
<__main__.Page object at 0x7f2891957e48>
{'title': 'testing', 'wiki': <__main__.Wiki object at 0x7f2891957d30>}
<__main__.Page object at 0x7f289190cf60>
{'title': 'testing something else',
 'wiki': <__main__.Wiki object at 0x7f2891957d30>}
True
False
True
True

As you can see, this works...

I'm not sure how I feel about your use of self.__dict__.update(data)... it caught me off guard first time round... and I'm currently suspicious of it's use here and here (both lines do the same thing...)

class Page(object):
    def __init__(self, wiki, getinfo=None, **data):
        # ...
        if getinfo is None:
            getinfo = GETINFO
        if getinfo:
            self.__dict__.update(self.info())

    def info(self):
        # ...
        self.__dict__.update(page_data)
        return page_data

Can you make sure that these calls don't override the title?

回答2:

Wait, now it works!

>>> w = mw.Wiki('https://en.wikipedia.org/w/api.php')
>>> a = [w.page('hi'), w.page('ih'), w.page('ij')]
>>> w.page('hi') in a
True

Okay, I don't know what went wrong before. This is solved now. I'll keep Attie's answer in mind, though.