Say I have two models set up like so:
class ParentFoo(SqlBase):
__tablename__ = "parents"
id = Column(String, primary_key=True)
children = relationship("ChildFoo", backref="parents")
class ChildFoo(SqlBase):
__tablename__ = "children"
id = Column(String, primary_key=True)
Now in a request handler, say I want to get all children entities from the database given a parent:
class RequestHandler():
def get(self, parent_id):
parent = database.query(parent_id)
children = parent.children
send_response(children)
The SQLAlchemy ORM API allows this type of query. The problem for me is when I'm writing my tests. I have a generic RequestHandler
class that takes a ParentFoo
and a ChildFoo
class as arguments in the constructor. So my request handler class actually looks like this:
class GenericRequestHandler():
def __init__(self, parent_class, child_class):
self.parent_class = parent_class
self.child_class = child_class
def get(self, parent_id):
parent = database.query(self.parent_class, parent_id)
children = parent.getattr(self.child_class.__tablename__, None)
As long as the column name on the parent matches the __tablename__
for the child entities, this will work. I enforce this convention. The problem is testing this.
Python's mock
library doesn't seem to provide for return values on attribute lookups and I don't want to mock out the getattr
function, because it's used elsewhere in my get method.
Any thoughts?
Aucun commentaire:
Enregistrer un commentaire