Wraps the best available JSON implementation available in a common interface
Overview
Anyjson loads whichever is the fastest JSON module installed and provides a uniform API regardless of which JSON implementation is used.Originally part of carrot (http://github.com/ask/carrot/)
Examples
To serialize a python object to a JSON string, call the serialize function:>>> import anyjson
>>> anyjson.serialize(["test", 1, {"foo": 3.141592}, "bar"])
'["test", 1, {"foo": 3.141592}, "bar"]'
Conversion the other way is done with the deserialize call.>>> anyjson.deserialize("""["test", 1, {"foo": 3.141592}, "bar"]""")
['test', 1, {'foo': 3.1415920000000002}, 'bar']
Regardless of the JSON implementation used, the exceptions will be the same.
This means that trying to serialize something not compatible with JSON
raises a TypeError:>>> anyjson.serialize([object()]) Traceback (most recent call last): <snipped traceback> TypeError: object is not JSON encodableAnd deserializing a JSON string with invalid JSON raises a ValueError:
>>> anyjson.deserialize("""['missing square brace!""")
Traceback (most recent call last):
<snipped traceback>
ValueError: cannot parse JSON description
No comments:
Post a Comment