Question

I have been exploring the json library, and am attempting to convert an object into JSON data and back again. I have run into trouble running this example code:

import json

class Obj:
    '''
    classdocs
    '''

    def __init__(self,s,hello="Hello world!"):
        '''
        Constructor
        '''
        self.s = s
        self.hello = hello
    def __repr__(self):
        return '<MyObj(%s,%s)>' % (self.s, self.hello)

def objToJSON(obj):
    return obj.__dict__

def jSONToObj(json):
    print(json)
    return Obj(**json)

if __name__ == '__main__':
    str = json.dumps(Obj("Hello","World"), default=objToJSON, sort_keys=True)
    print(str)
    print(json.loads(str,object_hook=jSONToObj))
    str = json.dumps(Obj("Text",{"a":"aaaa","b":"BBBBB","C":"ccccc"}), default=objToJSON, sort_keys=True)
    print(str)
    print(json.loads(str,object_hook=jSONToObj))

The output of which is:

{"hello": "World", "s": "Hello"}
{'s': 'Hello', 'hello': 'World'}
<MyObj(Hello,World)>
{"hello": {"C": "ccccc", "a": "aaaa", "b": "BBBBB"}, "s": "Text"}
{'a': 'aaaa', 'C': 'ccccc', 'b': 'BBBBB'}
Traceback (most recent call last):
  File "C:\Users\dimo414\src\test.py", line 27, in <module>
    print(json.loads(str,object_hook=jSONToObj))
  File "C:\Python31\lib\json\__init__.py", line 318, in loads
    return cls(**kw).decode(s)
  File "C:\Python31\lib\json\decoder.py", line 339, in decode
    obj, end = self.raw_decode(s, idx=_w(s, 0).end())
  File "C:\Python31\lib\json\decoder.py", line 355, in raw_decode
    obj, end = self.scan_once(s, idx)
  File "C:\Users\dimo414\src\test.py", line 22, in jSONToObj
    return Obj(**json)
TypeError: __init__() got an unexpected keyword argument 'a'

It seems that when a dictionary is a value in the object’s dictionary, the data passed to jSONToObj is the internal dictionary, not the full dictionary. Why is that?

Answer 1

Since you specified that objects should be reconstructed using the function jSONToObj, the deserializer assumes that all dicts are to be objects and tries to call your deserializer on them.

From the docs:

object_hook is an optional function
that will be called with the result of
any object literal decode (a dict).
The return value of object_hook will
be used instead of the dict. This
feature can be used to implement
custom decoders (e.g. JSON-RPC class
hinting).

The inner dict is the first to be deserialized, probably because loads works recursively and starts from the bottom layer. It makes sense that before deserializing an object, you have to deserialize its arguments.