graphene-federation 0.0.4

Federation implementation for graphene

graphene-federation

Federation support for graphene

Build:

Draft version, of federation specs implementation on top of Python graphene lib https://www.apollographql.com/docs/apollo-server/federation/federation-spec/

Use at your own risk

Based on discussion: https://github.com/graphql-python/graphene/issues/953#issuecomment-508481652

Supports now:

• sdl (_service fields) # make possible to add schema in federation (as is)
• @key decorator (entity support) # to perform Queries across service boundaries
  • You can use multiple @key per each ObjectType

      @key('id')
      @key('email')
      class User(ObjectType):
          id = Int(required=True)
          email = String()
      
          def __resolve_reference(self, info, **kwargs):
              if self.id is not None:
                  return User(id=self.id, email=f'name_{self.id}@gmail.com')
              return User(id=123, email=self.email)              
  

• extend # extend remote types
• external # mark field as external

Todo implement:

• @requires
• @provides

import graphene
from graphene_federation import build_schema, key

@key(fields='id')  # mark File as Entity and add in EntityUnion https://www.apollographql.com/docs/apollo-server/federation/federation-spec/#key
class File(graphene.ObjectType):
    id = graphene.Int(required=True)
    name = graphene.String()

    def resolve_id(self, info, **kwargs):
        return 1

    def resolve_name(self, info, **kwargs):
        return self.name

    def __resolve_reference(self, info, **kwargs):  # https://www.apollographql.com/docs/apollo-server/api/apollo-federation/#__resolvereference
        return get_file_by_id(self.id)

import graphene
from graphene_federation import build_schema


class Query(graphene.ObjectType):
    ...
    pass

schema = build_schema(Query)  # add _service{sdl} field in Query

import graphene
from graphene_federation import external, extend

@extend(fields='id')
class Message(graphene.ObjectType):
    id = external(graphene.Int(required=True))

    def resolve_id(self, **kwargs):
        return 1

__resolve_reference

• Each type which is decorated with @key or @extend is added to _Entity union
• __resolve_reference method can be defined for each type that is an entity. This method is called whenever an entity is requested as part of the fulfilling a query plan. If not explicitly defined, default resolver is used. Default resolver just creates instance of type with passed fieldset as kwargs, see entity.get_entity_query for more details
• You should define __resolve_reference, if you need to extract object before passing it to fields resolvers (example: FileNode)
• You should not define __resolve_reference, if fileds resolvers need only data passed in fieldset (example: FunnyText)
• read more in official documentation

---

For more details see examples

Or better check integration_tests

Also cool example of integration with Mongoengine

For contribution:

Run tests:

• make test
• if you've changed Dockerfile or requirements run make build before make test