CODEWITHSUNDEEP
pythonjsonclassobjectconfig
Blub21
Blub21

Reputation: 17

Python use .json as config file with class instances

I'm creating an application that stores it's config in a dictionary. I know I can write this to a JSON file and read this every time the app starts. But the problem is that this dictionary also contains objects. Like so(LED is an imported module with the classes APALedstrip and Arduino)

rooms['livingroom'] = {
    "room":data.room(name = 'livingroom',dataKeys = dataKeys),
    "lights":{
        "LedStrip":LED.APALedstrip(name = 'livingroom',
            room = 'livingroom')
        }
    }

rooms['bed'] = {
        "room":data.room(name = 'bed', dataKeys = dataKeys),
        "lights":{
            "LedStrip":LED.Arduino(name ='bed',
                serialPort = 'ttyUSB0',
                room = 'livingroom',
                master = {'room':'livingroom', 'light':'LedStrip'},
                roomSensors = 'livingroom')
            }
        }

I'm curious is it also possible to store this in an JSON file like so? And when it's imported into a dictionary that the objects are still created?

Upvotes: 0

Views: 726

Answers (1)

Panos Stavrianos
Panos Stavrianos

Reputation: 111

You need to serialize your objects. One way is to use "pickle".

Pickle convert an object to bytes, so the next step is to convert them to string using base64.

I choose base64 because it's safe for non-ASCII characters

In order to automatically save and retrieve the rooms use

save_rooms() and retrieve_rooms()

import codecs
import json
import pickle


def save_rooms(rooms):
    for room in rooms:
        # find all LedStrip objects
        if 'lights' in rooms[room] and 'LedStrip' in rooms[room]['lights']:
            lights = rooms[room]['lights']['LedStrip']

            # encode object to bytes with pickle and then to string with base64
            rooms[room]['lights']['LedStrip'] = codecs.encode(pickle.dumps(lights),
                                                              "base64").decode()
    with open("rooms.json", "w") as f:
        json.dump(rooms, f)


def retrieve_rooms():
    with open("rooms.json") as f:
        rooms = json.load(f)
    for room in rooms:
        # find all LedStrip objects
        if 'lights' in rooms[room] and 'LedStrip' in rooms[room]['lights']:
            lights = rooms[room]['lights']['LedStrip']

            # decode from string to bytes with base64 and then from bytes to object with pickle
            rooms[room]['lights']['LedStrip'] = pickle.loads(codecs.decode(lights.encode(), "base64"))

    return rooms


rooms = {}
rooms['livingroom'] = {
    "room": data.room(name='livingroom', dataKeys=dataKeys),
    "lights": {
        "LedStrip": LED.APALedstrip(name='livingroom',
                                    room='livingroom')
    }
}

rooms['bed'] = {
    "room": data.room(name='bed', dataKeys=dataKeys),
    "lights": {
        "LedStrip": LED.Arduino(name='bed',
                                serialPort='ttyUSB0',
                                room='livingroom',
                                master={'room': 'livingroom', 'light': 'LedStrip'},
                                roomSensors='livingroom')
    }
}
save_rooms(rooms)
loaded_rooms = retrieve_rooms()

In addition I implemented the logic so you can save any variation of rooms as long as you keep the structure the same.

ex. 

rooms['kitchen'] = {
    "room": data.room(name='kitchen', dataKeys=dataKeys),
    "lights": {
        "LedStrip": LED.APALedstrip(name='kitchen',
                                    room='kitchen')
    }
}

Upvotes: 1

Related Questions