py-json-serialize

Serialize in JSON format

Features:

• Simple api: @json_serialize, to_json(), from_json();
• Version support: version-based data migration;

Install

pip install py-json-serialize

Test

in the project's root folder, run pytest:

pytest

Dependencies

None

API

• class decortator @json_serialize

  • format 1: no parameter

   from py_json_serialize import json_serialize
  
   @json_serialize
   class A(object):pass
  

  • format 2: with parameter

  @json_serialize(clsid = "", [version=0*|n])

  1. clsid: the unique string to identify the class. the class-id will be the name of decorated if not specified.
  2. version: optional parameter to specify the version of serialized data format. the default value is 0 if not specified.

   from py_json_serialize import json_serialize
  
   # old version
   @json_serialize("app-config", version=1)
   class AppConfigV1(object):
       servers = []
  
   # new version
   @json_serialize("app-config", version=2)
   class AppConfigV2(AppConfigV1):
       timeout = 600
  

• to_json()/from_json()

  The decorated class will have two new functions:

  1. to_json(): convert class instance to json string

     @json_serialize
     class Hello(object):
         def __init__(self, who = "World"):
             self.who = who
     
     a = Hello()
     print(a.to_json())
     

     outputs:

     {
         "_clsid_": "Hello",
         "who": "World"
     }
     

  2. from_json(str|dict): reads json string or dictionary object to return an class object.

     It is actually a staticmethod that can be called to return an object of any type deduced from the class-id in the data string. so don't be surprised that you might get an object of different type if the input json data string is serialized from another class.

     So the from_json() is simply a handy helper method to make your code more readable if your app only handles one type of data.

• function json_decode(jstr: str|dict)-> object: convert json string or dictionary to python object

  This is a function to decode the serialized json string, its typical usage is as following:

  class Task(object):pass
      @staticmethod
      def from_json(jstr):
          return json_decode(jstr)
  
  @json_serialize
  class CopyFile(Task):pass
  
  @json_serialize
  class UploadFile(Task):pass
  
  task1 = Task.from_json({ CLASS_ID:'CopyFile' })
  assert isinstance(task1, CopyFile)
  
  task2 = Task.from_json({ CLASS_ID:'UploadFile' })
  assert isinstance(task2, UploadFile)
  

• function json_encode : convert python object to json string

  def json_encode(obj, pretty = True, encode_all_fields = False)
  

  When pretty is True, the fields are sorted by their name and the json string is intented properly for human reading, otherwise the json string can save some storage space and more efficient for machine processing.

  If encode_all_fields is true, then all class fields are serialized, otherwise the internal fields (field name starts with '_') are ignored.

• CLASS_ID: the unique key indicating which class is serialized.

  It is published in case a json string or dictionary object needs to be prepared manually as input to json_decode()/from_json().

  The internal constant value of CLASS_ID is "_CLSID_"

  json_decode({ CLASS_ID: "Hello", who: "World" })
  json_decode('{ "%s": "Hello", "who": "World" }' % CLASS_ID )