Extend the base SQLAlchemy utility methods in models/utils.py to allow loading relationships in asyncpg

[devraj]

Due to how the 2.0 syntax works and using asyncpg writing generic helper methods when using relationships becomes a lot trickers. Essentially you have to write a select as follows:

        query = select(cls).options(selectinload(cls.memberships),\
            selectinload(cls.card_holder))

for the relationships to work. This makes writing generic CRUD wrapper difficult.

The proposal here is to write a generic method called _get_select which provides the joined configuration of the select statement which other get methods can use to apply conditions to e.g:

        query = select(cls).options(selectinload(cls.memberships),\
            selectinload(cls.card_holder))\
            .where(or_(cls.email == email,\
                cls.mobile_number == mobile_number))

a refactored version would then look like

        query = cls._get_select()
            .where(or_(cls.email == email,\
                cls.mobile_number == mobile_number))

each class inheriting from the ModelCRUDMixin can override that to provide the appropriate configuration, allowing the global getting to work across models and making it easier to write other getters.

Another benefit of this approach will be that as the relationships are modified they can be changed in one spot and all other getters remain unchanged.

See also #38