jsonalize 0.2

A JSON data binding library for Python

Introduction

This library supports two-way data binding between JSON and Python class.

Turorial

Here is a simple example to use jsonalize:

from jsonalize import *


# Define a class
class MyData(JSONObject):
    def __init__(self):
        JSONObject.__init__(self)
        self.id = JSONString()
        self.name = JSONString()
        self.age = JSONInt()
        self.weight = JSONFloat()


# Create an object of MyData
my = MyData()
my.id = "20190101"
my.name = "Stanley"
my.age = 28
my.weight = 60

# jsonalize the object
json_str = my.to_json()
print(json_str)

# restore the object from json
my2 = JSONObject.from_json(MyData, json_str)
print(my2.to_json())

This example should output the following message:

{"age": 28, "id": "20190101", "weight": 60.0, "name": "Stanley"}
{"age": 28, "id": "20190101", "weight": 60.0, "name": "Stanley"}

Another example shows how to prevent creating real objects for attributes when constructing a new JSONObject instance.

class MyData(JSONObject):
    def __init__(self):
        JSONObject.__init__(self)
        self.id = JSONNoneInt
        self.name = JSONNoneString
        self.attr = JSONObject.NoneValue(Attr)


class Attr(JSONObject):
    def __init__(self):
        JSONObject.__init__(self)
        self.value = JSONString()
        print("Attr object created")


my = MyData()
print(my.id is None, my.name is None, my.attr is None)

my.id = 10
my.name = "Stanley"
my.attr = Attr()
my.attr.value = "attr"
print(my.to_json())

# Outputs:
# (True, True, True)
# Attr object created
# {"attr": {"value": "attr"}, "name": "Stanley", "id": 10}

Key points from this tutorial

• A serializable class should inherit the JSONObject class
• Don't forget to invoke the __init__ method in your class
• The serializable class attributes should be set as JSON** types

List of supported JSON types

• JSONInt
• JSONLong (Only in Python 2)
• JSONFloat
• JSONComplex
• JSONBool
• JSONString
• JSONList
• JSONSet
• JSONDict
• JSONObject

Most of the types can be initialized with an initial value, for example:

a_string = JSONString("hello jsonalize")

Remarks for JSONBool

You can't test the value of a JSONBool object with the is keyword, because is will compare the instance id of two objects, but an object of JSONBool is never an instance of True or False.

You can do the test as follows:

a_bool = JSONBool(True)
print(a_bool is True, a_bool == True)
# False, True
print(a_bool.true(), a_bool.true() is True, a_bool.equals(True))
# True, True, True

A more complex example

You can have an object of JSONObject in another JSONObject class:

class Monitor(JSONObject):
    def __init__(self):
        JSONObject.__init__(self)
        self.size = JSONFloat()
        self.power = JSONFloat()
        self.color = JSONString()


class Computer(JSONObject):
    def __init__(self):
        JSONObject.__init__(self)
        self.brand = JSONString()
        self.monitor = Monitor()


computer = Computer()
computer.brand = "Lenovo"
computer.monitor.size = 23.0
computer.monitor.power = 25.0

json_str = computer.to_json()
print(json_str)
#{"brand": "Lenovo", "monitor": {"color": "", "power": 25.0, "size": 23.0}}

computer2 = JSONObject.from_json(Computer, json_str)
print(computer2.to_json())
#{"brand": "Lenovo", "monitor": {"color": "", "power": 25.0, "size": 23.0}}

A list of JSONObject objects?

Look at the following example:

class Student(JSONObject):
    def __init__(self):
        JSONObject.__init__(self)
        self.id = JSONString()
        self.name = JSONString()


class School(JSONObject):
    def __init__(self):
        JSONObject.__init__(self)
        self.address = JSONString()
        self.students = JSONList()


stu1 = Student()
stu1.id = "20190202"
stu1.name = "Stanley"

stu2 = Student()
stu2.id = "20190203"
stu2.name = "Cyrus"

school = School()
school.address = "Central Street No.23"
school.students.append(stu1)
school.students.append(stu2)

json_str = school.to_json()
print(json_str)
#{"students": [{"id": "20190202", "name": "Stanley"}, {"id": "20190203", "name": "Cyrus"}], "address": "Central Street No.23"}

school2 = JSONObject.from_json(School, json_str)
print(type(school2.students[0]), school2.students[0])
#(<type 'dict'>, {u'id': u'20190202', u'name': u'Stanley'})

As you can see here, the deserializing of the School object is incorrect. Because any type of object could be appended to school.students, so jsonalize don't know what to restore when deserializing.

This problem is currently not resolved, maybe I can add a type mapping structure in the future.