injection of a dynamic class

Given a class known at runtime (dynamic) with dependencies in the c’tor, can we use the container to initialize such class?

E.g:

from dependencies import Injector

class Person:
    def __init__(self, name: str):
        pass

def get_cls():
    return Person

class Container(Injector):  # (1)
    ...
    # (2)

# (3)
person = ...  # (4)

1. I tried to inherit from more classes, but got an exception: _dependencies.exceptions.DependencyError: Multiple inheritance is allowed for Injector subclasses only
2. I tried to use __new__ but with no luck
3. I tried to register a class variable after the class definition, but got an exception: _dependencies.exceptions.DependencyError: 'Injector' modification is not allowed

Is there away to create such dynamic behavior? Or is it inevitable becuase of the nature of the simplified and non-boilerplate API?

@proofit404 Looking forward to hear your opinion 😃

[proofit404]

Feel free to reopen this issue if you have further questions.

Have a good day 🌴 🍸

Best regards, Artem.

[proofit404]

>>> from dependencies import Injector
>>> from dataclasses import dataclass
>>> @dataclass
... class Person:
...     name: str
...
>>> @dataclass
... class Animal:
...     age: int
...
>>> def get_plugin_classes():
...     return [Person, Animal]
...
>>> @dataclass
... class Plugins:
...     plugins_lst: ...
...
>>> class AllDeps(Injector):
...     name = 'hello'
...     age = 5
...
>>> class Container(Injector):
...     plugins = Plugins
...
>>> Container(plugins_lst=[AllDeps(plugin=plugin).plugin for plugin in get_plugin_classes()]).plugins
Plugins(plugins_lst=[Person(name='hello'), Animal(age=5)])

We could try something like this.

It still requires you to know names of possible dependencies for all classes which could be returned from get_plugin_classes.

What do you think?