Javascript-style Object

jo is implemented as a subclass of dict, mapping attribute access and manipulation into item access and manipulation.

>>> from jo import jo
>>> j = jo()
>>> j.a = 1
>>> j.a
1

Reading attributes that don't exist produces None. If you need an exception raised, use item access instead.

>>> print(j.b)
None

>>> j['b']
Traceback (most recent call last):
 ...
KeyError: 'b'

>>> j['b'] = 'b'
>>> j.b += 'e'
>>> print(j.b)
be

By definition, attributes and items are always the same, as demonstrated by the invariant:

>>> vars(j) == dict(j)
True

Because jo is a subclass of dict, serializing can be handled simply

>>> import json
>>> print(json.dumps(j, sort_keys=True))
{"a": 1, "b": "be"}

This is exactly how string conversion is implemented as well

>>> print(j)
{"a": 1, "b": "be"}

repr performs in keeping with the python data model description

>>> del j.b

>>> j
jo({'a': 1})

To deserialize from json, simply pass jo as object_hook

>>> print(json.loads('{"a":1}', object_hook=jo))
{"a": 1}

As a convenience, the classmethod from_json performs the same task (and auto-detects whether it is loading from text or from a file)

>>> print(jo.from_json('{"a":1}'))
{"a": 1}

>>> from io import StringIO
>>> sample = StringIO(u'{"a":1}')

>>> print(jo.from_json(sample))
{"a": 1}