This package makes it easier to work with dictionaries in sessions. It provides a class which makes it easy to retrieve/create a path into request.session by specified the path as a list.

For instance ['foos', foo_id, 'bars', bar_id] represents request.session['foos'][foo_id]['bars'][bar_id]

Any int elements in path will be converted to strings

Once created you can work with the object, without needing to specify the path anymore.

pip install django-namespaced-session

Make sure 'django.contrib.sessions' is in INSTALLED_APPS. This package isn't a proper Django 'app' - simply a class you can import - so it doesn't need to be added to INSTALLED_APPS itself.

from namespaced_session.session import NamespacedSession

def some_view(request, foo_id, bar_id):

    # returns an object which wraps session['foos'][foo_id]['bars'][bar_id]
    # initializes the path via defaultdict if the path doesn't exist
    nss = NamespacedSession(request,
                            ['foos', foo_id, 'bars', bar_id])

    # returns the dict stored at session['foos'][foo_id]['bars'][bar_id]
    sess_info = nss.get_sess_info()

    # sets session['foos'][foo_id]['bars'][bar_id]['some_field'] = some_value
    nss.update({'some_field': some_value})

    # sets session['foos'][foo_id]['bars'][bar_id]["baz"] = 2
    nss.set(["baz"], 2)

    # sets session['foos'][foo_id]['bars'][bar_id]['boo']['qux'] = {"quux": 1}
    nss.set(["boo", "qux"], {"quux": 1})

    # returns {"qux": {"quux": 1}}
    nss.get(["boo"])

    # deletes "qux" key from "boo". sessions['foos'][foo_id]['bars'][bar_id]['boo']
    # is now an empty dict
    nss.delete(["boo", "qux"])

    # sets session['foos'][foo_id]['bars'][bar_id] to an empty dict
    nss.clear()

    # sets request.session.modified = True and also calls request.session.save()
    # this is called automatically by all methods which modify the session
    # probably should remove the call to request.session.save() for efficiency reasons
    nss.save()

    # clears contents stored at session['foos'][foo_id]['bars'][bar_id] and then, working
    # backwards from 'bars' all the way down to 'foos', if the
    # key's value is an empty dict, will delete that key. deletion will
    # stop at the first key whose value is not an empty dict
    nss.destroy()