* Page: added about 400 lines of rambling nonsense.
* Exceptions: added RedirectError.

* Site: api_query() -> _api_query(); api_query() acts as a wrapper for _api_query(), accepting API params as **kwargs.
* Various cleanup throughout and minor fixes.

* Exceptions: New PermissionsError; reworded docstring of SiteAPIError.
* Site: __init__() accepts an optional cookiejar parameter, otherwise we
  use CookieJar(). Added five new cookie/username-related methods. Only
  login from __init__() if we are missing valid login cookies and a user/
  pass was provided. _login() and _logout() both try to save cookies via
  _save_cookiejar(). _load_attributes() automatically refreshes all
  attributes other than namespaces if at least one is missing, instead of
  only the missing ones. api_query() raises SiteAPIError if either
  self._base_url or self._script_path is missing. Removed some pointless
  methods and renamed one; added domain().
* Functions: _get_site_object_from_dict() is cleaner, adds our cookiejar
  to Site instances using _get_cookiejar() to load a LWPCookieJar() object
  from the ".cookies" file in our project root. The same cookiejar is
  returned for every site, enabling crosswiki login, via a global variable.
* User: Renamed some methods.
* .gitignore: Added .cookies file.

* Site: New _get_logged_in_user() method, name self-explanatory. This acts
  as a replacement for the former crud in get_user(), which now calls this
  when the username arg is None. This method will first try to determine
  our username based on a special cookie in self._cookiejar (cookie.name is
  self._name + "UserName", e.g. "enwikiUserName"), and will only do an API
  query if no cookie was found. This removes an API query that is usually
  only necessary if we are not logged in.
* Site: silly bugfix in __init__().
* User: Reverted earlier change to _get_attribute() (addition of
  raise_exception arg); name() and exists() now use their own code, which
  is simpler.
* User: Calling name() does not do an API query unless force=True, unlike
  the other "get" methods.
* User: .join() instead of .format() because I feel it looks cleaner and is
  probably more efficient.

* Site's api_query() is much smarter. It uses a custom urllib2 URL opener with cookie support and catches URLErrors, raising its own brand new exception (SiteAPIError) when something is wrong.
* The opener now uses a custom User-Agent, which is a constant in wiki.tools.constants.
* Site instances automatically login via _login(), which accepts a username and password (provided via config by get_site()) and uses two api_query()s and stores the login data as cookies in self._cookiejar. Login data is not preserved between bot restarts yet. Login errors, e.g. a bad password or username, raise the new LoginError.
* Site's get_user()'s username argument is now optional. If left blank, will return the current logged-in user, provided by an API query.
* Misc cleanup throughout.