Basic inheritance not works while using dependency_injector library DeclarativeContainer in Python. Why?

Question

It is known the Python have ability to create child classes and add methods into them. However when I inherit class from DeclarativeContainer it fails and I cannot understand why. This code works fine if I inherit from different classes. Please help to understand why inheritance not works as I expect in this case.

from dependency_injector import containers

class Container(containers.DeclarativeContainer):
    def printContainer(self):
        print(f'Hello from container!')

if __name__ == "__main__":
    container = Container()
    container.printContainer() # error here

The error, surprisingly, says my code dealing with DynamicContainer class, however i not use it but DeclarativeContainer. Actually I expected method to be called on a Container class:

AttributeError: 'DynamicContainer' object has no attribute 'printContainer'

Here some context. I am trying to make use of dependency_injector library instead of manually injecting dependencies. I got some errors and confusions and made as as simple as possible code example to understand what is going on.

Answer 1

Please note that the dependency_injector documentation states:

The declarative container can not have any methods or any other attributes then providers.

---

Now let's find out the reason for the confusing error. According to the source code of dependency_injector, DeclarativeContainer.__new__ returns container which is an instance of the DynamicContainer class.

@six.add_metaclass(DeclarativeContainerMetaClass)
class DeclarativeContainer(Container):

    # omitted code

    instance_type = DynamicContainer
    """Type of container that is returned on instantiating declarative
    container.
    :type: type
    """

    # omitted code

    def __new__(cls, **overriding_providers):
        """Constructor.
        :return: Dynamic container with copy of all providers.
        :rtype: :py:class:`DynamicContainer`
        """
        container = cls.instance_type()

        # omitted code

        return container

Thus, when you create your Container object, you receive a DynamicContainer instance. You can verify it by checking the type and attributes of the object.

from dependency_injector import containers

class Container(containers.DeclarativeContainer):
    def printContainer(self):
        print(f'Hello from container!')

if __name__ == "__main__":
    container = Container()
    print(type(container))
    print(dir(container))

Outputs:

<class 'dependency_injector.containers.DynamicContainer'>
['__IS_CONTAINER__', '__class__', '__deepcopy__', '__delattr__', '__dict__', '__dir__', '__doc__', '__eq__', '__format__', '__ge__', '__getattribute__', '__gt__', '__hash__', '__init__', '__init_subclass__', '__le__', '__lt__', '__module__', '__ne__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__self__', '__setattr__', '__sizeof__', '__str__', '__subclasshook__', '__weakref__', 'apply_container_providers_overridings', 'assign_parent', 'check_dependencies', 'declarative_parent', 'dependencies', 'from_json_schema', 'from_schema', 'from_yaml_schema', 'init_resources', 'is_auto_wiring_enabled', 'load_config', 'overridden', 'override', 'override_providers', 'parent', 'parent_name', 'provider_type', 'providers', 'reset_last_overriding', 'reset_override', 'reset_singletons', 'resolve_provider_name', 'set_provider', 'set_providers', 'shutdown_resources', 'traverse', 'unwire', 'wire', 'wired_to_modules', 'wired_to_packages', 'wiring_config']

It explains why you encounter AttributeError when trying to execute container.printContainer().