Why repository does not found entity_id ?

Hello,

I’m trying to work with eventsourcing library and for the simple example :

class Ingredient(AggregateRoot):
    class Event(AggregateRoot.Event):
        pass

    @classmethod
    def create(cls, name: str):
        return cls.__create__(name=name)

    class Created(Event, AggregateRoot.Created):
        pass

    def __init__(self, name: str, **kwargs):
        super().__init__(**kwargs)

        self.name = name

    def __repr__(self):
        return f"<Ingredient (id={self.id} name={self.name}) object at {hex(id(self))}>"


class Ingredients(ProcessApplication):
    @applicationpolicy
    def policy(self, repository, event):
        pass

    @staticmethod
    def create_ingredient(name: str):
        return Ingredient.create(name)

When i try to execute the following code from a runner:

with SingleThreadedRunner(system, infrastructure_class=PopoApplication) as runner:
    ingredient = runner.system.ingredients.create_ingredient(name="Test")
    print(ingredient.id in runner.system.ingredients.repository)

the print returns “False” (the ingredient is not persisted ?) where i’m doing something wrong ?

[ghost]

It works now, thank you. You did a great work !

I will open other issues if I have others questions.

Note: https://eventsourcing.readthedocs.io/en/stable/topics/process.html#processes is not up to date 😃

[johnbywater]

@JeanDeuns I think it’s up to date? There’s a small difference between the code that you linked to and the code you wrote. The “general rule” is that the __save__() method should be called on new or changed aggregates, except when they are created or updated within a process application policy (so that the domain events can be persisted atomically with the tracking and notification information). In particular, in that bit of the documentation, there is a command method on the Commands process application which does call __save__(). The other applications create and update aggregates with their process application policy, and so correctly do not call __save__(). It’s a subtle aspect perhaps, but it reflects the atomicity of the recording. __save__() is atomic only for the new domain events of a single aggregate. Hope that helps 😃