Queries become unwieldy for larger graphql schemas

[AdrielVelazquez]

I'm utilizing epoxy for sqlalchemy models, which is working out very well; however, when attempting to separate out the models an issue arises with inheritance.

class QueryMixin(R.ObjectType):
    field1 = R.String
    field2 = R.Integer


Class Query(QueryMixin):
    pass #or possibly have another field

When you create a situation like this, Query doesn't have any of the parameters of QueryMixin and thus the execution fails when creating a schema.

I propose a better handling of inheriting mixins. This may be a circumstance of graphql-core not handling inheritance first.

[syrusakbary]

@AdrielVelazquez have you taken a look into graphene? The inheritance/types is handled a little bit different, and the bug you're exposing is already solved

PS: GraphQL-epoxy is a great package too! Use the one you feel more comfortable with
PS2: SQLAlchemy automatic mapping will come to graphene soon :)

[AdrielVelazquez]

I'm trying to look in graphene the equivalent of the following

@R.Field.CanBe
class RandomClass(object):
    #Random attributes

But I can't seem to find it. Does graphene have support for normal classes?

[syrusakbary]

I'm certainly confident that you would not need the canBe decorator in graphene.
Just return the instance whenever you're using it.
If that doesn't work... open an issue in the graphene project and I will solve it ASAP :)

[jhgg]

Hmm. Mixin support is something I can do @AdrielVelazquez.

[jhgg]

I've implemented it as such:

class NameMixin():
    name = R.String

class Dog(R.ObjectType, DogMixin):
    breed = R.String