Fix memory leak in ObservableObject

[mastadle]

Pydase currently has a significant memory leak if a DataService contains a dict or a list.

Changed

• Handle adding and removing to the mappings in the derived ObservableObject classes
• Use weak references in dict/list mappping to avoid memory leak

Minimal Example

import json
import time

import pydase

dict_json = """
{
    "foo": "bar"
}
"""

class Service(pydase.DataService):
    def __init__(self) -> None:
        super().__init__()
        self.dict_attr = {}

service = Service()

while True:
    service.dict_attr = json.loads(dict_json)
    time.sleep(0.01)

[mosmuell]

Thanks for debugging this!