我有一个带有Relay和过滤器的Graphene接口。它工作得很好,但我想添加order_by选项。我的对象看起来像:
class FooGQLType(DjangoObjectType):
class Meta:
model = Foo
exclude_fields = ('internal_id',)
interfaces = (graphene.relay.Node,)
filter_fields = {
"id": ["exact"],
"code": ["exact", "icontains"],
}
connection_class = ExtendedConnection
class Query(graphene.ObjectType):
foo = DjangoFilterConnectionField(FooGQLType)
ExtendedConnection不应该相关,但是:
class ExtendedConnection(graphene.Connection):
class Meta:
abstract = True
total_count = graphene.Int()
def resolve_total_count(root, info, **kwargs):
return root.length
这使我可以像foo(code_Icontains:"bar")
一样进行查询。
根据{{3}},我应该为此在FilterSet中使用OrderingFilter。我觉得这有点烦人,因为过滤器应该是自动的,但是如果我这样做的话:
class FooGQLFilter(FilterSet):
class Meta:
model = Foo
order_by = OrderingFilter(
fields=(
('code', 'code'),
('lastName', 'last_name'),
('otherNames', 'other_names'),
)
)
我收到需要提供fields
或exclude
的错误消息:
AssertionError: Setting 'Meta.model' without either 'Meta.fields' or 'Meta.exclude' has been deprecated since 0.15.0 and is now disallowed. Add an explicit 'Meta.fields' or 'Meta.exclude' to the FooGQLFilter class.
因此,如果我添加一个fields = []
使其静音,则它将编译。
但是,当我在以下位置使用它时:
foo = DjangoFilterConnectionField(FooGQLType, filterset_class=FooGQLFilter)
我的常规过滤器(例如code_Icontains
)消失了。我可以在那儿再次添加它们,但这很愚蠢。快速查看源代码,看起来Relay或django-filters已经创建了一个FilterSet类(很有意义),用这种方式覆盖它显然是一个糟糕的主意。
如何在Graphene Relay过滤的对象上添加orderBy过滤器?我觉得这应该很简单,但是我正在努力弄清楚。
我还看到了用DjangoFilterConnectionField
子类化connection_resolver
的示例,该示例以某种方式注入了order_by,但告诉我没有orderBy参数。
答案 0 :(得分:4)
Eric的解决方案不适用于当前的graphene-django版本(2.9.1)或更高版本的graphene-django 2.6.0版本。
DjangoFilterConnectionField方法在2.7.0版本中进行了更改。 有关更多详细信息,您可以检查更改日志here
使用Eric的解决方案,它将产生错误,
connection_resolver() missing 1 required positional argument: 'info’
所以我已经修改了解决方案,并且效果很好。
from graphene_django.filter import DjangoFilterConnectionField
from graphene.utils.str_converters import to_snake_case
class OrderedDjangoFilterConnectionField(DjangoFilterConnectionField):
@classmethod
def resolve_queryset(
cls, connection, iterable, info, args, filtering_args, filterset_class
):
qs = super(DjangoFilterConnectionField, cls).resolve_queryset(
connection, iterable, info, args
)
filter_kwargs = {k: v for k, v in args.items() if k in filtering_args}
qs = filterset_class(data=filter_kwargs, queryset=qs, request=info.context).qs
order = args.get('orderBy', None)
if order:
if type(order) is str:
snake_order = to_snake_case(order)
else:
snake_order = [to_snake_case(o) for o in order]
qs = qs.order_by(*snake_order)
return qs
答案 1 :(得分:1)
我已经针对这个主题改编了a GitHub issue的解决方案:
from graphene_django.filter import DjangoFilterConnectionField
from graphene.utils.str_converters import to_snake_case
class OrderedDjangoFilterConnectionField(DjangoFilterConnectionField):
"""
Adapted from https://github.com/graphql-python/graphene/issues/251
Substituting:
`claims = DjangoFilterConnectionField(ClaimsGraphQLType)`
with:
```
claims = OrderedDjangoFilterConnectionField(ClaimsGraphQLType,
orderBy=graphene.List(of_type=graphene.String))
```
"""
@classmethod
def connection_resolver(cls, resolver, connection, default_manager, max_limit,
enforce_first_or_last, filterset_class, filtering_args,
root, info, **args):
filter_kwargs = {k: v for k, v in args.items() if k in filtering_args}
qs = filterset_class(
data=filter_kwargs,
queryset=default_manager.get_queryset(),
request=info.context
).qs
order = args.get('orderBy', None)
if order:
if type(order) is str:
snake_order = to_snake_case(order)
else:
snake_order = [to_snake_case(o) for o in order]
qs = qs.order_by(*snake_order)
return super(DjangoFilterConnectionField, cls).connection_resolver(
resolver,
connection,
qs,
max_limit,
enforce_first_or_last,
root,
info,
**args
)
要使用它,只需修改以下查询:
claims = DjangoFilterConnectionField(ClaimsGraphQLType)
到
claims = OrderedDjangoFilterConnectionField(ClaimsGraphQLType,
orderBy=graphene.List(of_type=graphene.String))
然后您可以查询:
{ claims(status: 2, orderBy: "-id") { id } }
或
{ claims(status: 2, orderBy: ["creationDate", "lastName"]) { id } }
答案 2 :(得分:0)
嘿,我相信我有一个简单易行的答案来处理 orderBy,尤其是作为石墨烯中的列表。我回答的原因不是因为我相信我有最聪明的解决方案,我只是想知道如果我在不知不觉中造成伤害有什么区别。我实际上不了解 graphene_django 中的连接内容,因此我从观看有关 GraphQL 中排序对象的讨论中获得启发,采取了不同的路线。
https://www.youtube.com/watch?v=dDxUu-K2qdE
首先使用您的 FooType
,我已将字段添加为枚举
class FooGQLType(DjangoObjectType):
class Meta:
model = Foo
class FooFields(graphene.Enum):
CODE = "code"
LAST_NAME = "last_name"
OTHER_NAMES = "other_names"
class Directions(graphene.Enum):
ASC = "asc"
DESC = "desc"
然后输入以在 Query
class FooOrderByInput(graphene.InputObjectType):
order_by = FooFields()
direction = Directions()
我不确定您要过滤的内容,但由于问题主要是关于 order_by
,所以我暂时假设它是一个随机字符串。我觉得有一种方法可以发出这一行 db 请求,但我不确定如何。
ORDER_BY = 'order_by'
DIRECTION = 'direction'
class Query(graphene.ObjectType):
foo = graphene.Field(FooGQLType, input=graphene.List(FooOrderByInput), filter_opt=graphene.String())
def resolve_foo(self, info, input):
qs = Foo.objects.filter(filter_opt=filter_opt)
for obj in input:
order_by = obj.get(ORDER_BY)
direction = "-" if obj.get(DIRECTION) == Directions.DESC else ""
qs = qs.order_by(f"{direction}{order_by}")
return qs
这方面最好的事情(无论如何对我来说)是它允许对多个字段进行查询,例如
query ($input: [FooOrderByInput]) {
foo (input: $input) {
id
code
lastName
otherNames
}
}
VARIABLES
{
"input": [
{
"orderBy": "LAST_NAME",
"direction": "DESC"
},
{
"orderBy": "CODE",
"direction": "ASC"
}
]
}
而且除了那些拼写正确的关键字外,它也不允许任何其他内容。它也不必处理camelCase到snake_case。
如果有更好的解决方案或者我的很烂,请告诉我。我很高兴这似乎已经解决了!