Optimizing queries which are part of standard graphene.ObjectType

[cramshaw]

Hi! First off thanks for this module, it's really great, and for the release yesterday, I'd just come across the prefetch/select related bug and it was fixed before I could report it!

So I have been playing extensively and want to use this alongside some sort of pagination. Currently I have a query that looks like:

  wi4(page: 2) {
    totalCount
    page
    results {
      idInstance
      image {
         idImage
      }
    }
  }

Where results is a Django queryset.

To make this work I have a resolver which does:

class Query:
    wi_4 = graphene.Field(
        PaginatedInstanceType, page=graphene.Int())

    def resolve_wi_1(self, info, page=1, **kwargs):
        qs = gql_optimizer.query(
            Instance.objects.all(), info)
        return get_paginator(qs, PaginatedInstanceType, page=page)

and the type:

class PaginatedInterface(graphene.Interface):
    page = graphene.Int()
    pages = graphene.Int()
    has_next = graphene.Boolean()
    has_prev = graphene.Boolean()
    total_count = graphene.Int()
    results = None

class PaginatedInstanceType(graphene.ObjectType):
    results = graphene.List(InstanceType)

    class Meta:
        interfaces = (PaginatedInterface,)

Where InstanceType is a standard django-graphene DjangoObjectType.

Unfortunately this seems to result in the optimizer being disregarded and it runs n+1 queries. I've tried the recommended advanced usages but they all seem to rely on the ObjectType being an actual DjangoObjectType, which doesn't work here as I don't have a model at the top level.

I have found that if I 'trick' it by settings the results selection as the main selection, the field_name to results and the parent type to the GraphQLObject of the PaginatedInstanceType (passed in here as object_type) then it all works beautifully.

       results = filter(
            lambda x: x.name.value == 'results', info.field_asts[0].selection_set.selections)
        info.field_asts = list(results)

        info.parent_type = info.schema.get_type(
            str(object_type))

        info.field_name = 'results'

Before I start trying to make any changes, I wanted to check that this isn't currently supported and I haven't just missed how to go about it?

Thanks!

[Instrumedley]

no updates for this I guess :/