DRF HyperlinkedRelatedField跨OneToOne关联报无属性错误

开发基于Django/DRF的业务系统时，设计了办公室（Office）、HR用户两类模型：HR用户通过OneToOneField与HRProfile模型关联，需求为支持将HR用户分配至指定办公室。当前在OfficeSerializer中通过HyperlinkedRelatedField配置关联对象查找逻辑时，始终无法跨关联访问User模型的email字段，抛出'HRProfile' object has no attribute 'related_user.email'异常。

注：直接在HRProfile的__str__方法中调用self.related_user.email可正常获取值，排除模型关联本身的配置错误。

模型层代码 #

class User(AbstractBaseUser, PermissionsMixin):
    ...
    email = models.EmailField(unique=True)
    ...

from polymorphic.models import PolymorphicModel
class Profile(PolymorphicModel):
    ...    
    related_user = models.OneToOneField(
        settings.AUTH_USER_MODEL,
        on_delete=models.SET_NULL,
        related_name="profile",
    )
    ...

class HRProfile(Profile):
    some_hr_specific_field = models.CharField()    
    def __str__(self) -> str:
        return self.related_user.email

class Office(models.Model):
    ...
    assigned_hrs = models.ManyToManyField(
        "users.HRProfile", 
        related_name="offices", 
        blank=True
    )
    ...

视图层代码 #

class UserViewSet(GenericViewSet):
    serializer_class = UserSerializer
    queryset = User.objects.all()
    lookup_field = "email"
    lookup_url_kwarg = "email"
    lookup_value_regex = "[\\w@.]+"
    ...

# OfficeViewSet仅配置了queryset和serializer两个必填参数

存在问题的序列化器代码 #

class OfficeSerializer(serializers.HyperlinkedModelSerializer):
    ...
    assigned_hrs = serializers.HyperlinkedRelatedField(
        queryset=HRProfile.objects.all(),
        view_name="api:user-detail",
        lookup_field="related_user.email",
        many=True,
    )
    ...

• 配置参数source='assigned_hrs.related_user', lookup_field='email'
• 将HyperlinkedRelatedField替换为SlugRelatedField测试，设置slug_field="related_user.email"
• 使用双下划线替代点号分隔关联字段路径
  当前仅当设置lookup_field='some_hr_specific_field'时字段可以正常解析，但该字段不符合业务需求。

DRF内置的RelatedField系列字段（包括HyperlinkedRelatedField、SlugRelatedField）的lookup_field、slug_field参数不支持跨关联关系的链式属性查找：框架会直接在当前序列化的模型实例（此处为HRProfile实例）上查找与参数值完全同名的属性，传入related_user.email时，框架会直接寻找名为related_user.email的属性，自然会抛出属性不存在的异常。
__str__方法中self.related_user.email可以正常访问，是因为这是Python原生的链式属性调用逻辑，和DRF字段的参数解析逻辑完全无关，二者没有可对比性。

根据业务需求选以下两种方案即可：

方案1：保留超链接形式（指向User详情接口） #

自定义支持链式属性查找的关联字段，重写属性获取、反序列化逻辑：

from rest_framework.relations import HyperlinkedRelatedField
from rest_framework.exceptions import ValidationError
from django.core.exceptions import ObjectDoesNotExist

class ChainedLookupHyperlinkedField(HyperlinkedRelatedField):
    def get_attribute(self, instance):
        """序列化输出时，按点号分割路径逐层取属性值"""
        current_attr = instance
        for field_part in self.lookup_field.split('.'):
            try:
                current_attr = getattr(current_attr, field_part)
            except ObjectDoesNotExist:
                return None
        return current_attr

    def to_internal_value(self, data):
        """反序列化输入时，根据传入的email值匹配对应的HRProfile实例"""
        try:
            user = User.objects.get(**{self.lookup_url_kwarg: data})
            # 适配django-polymorphic的多态查询，返回真实的HRProfile实例
            return user.profile.get_real_instance()
        except (User.DoesNotExist, ObjectDoesNotExist):
            raise ValidationError(f"对应邮箱为{data}的HR用户不存在")

替换序列化器中原有字段配置即可：

assigned_hrs = ChainedLookupHyperlinkedField(
    queryset=HRProfile.objects.all(),
    view_name="api:user-detail",
    lookup_field="related_user.email",
    lookup_url_kwarg="email",
    many=True,
)

方案2：仅需返回邮箱值、不需要超链接 #

直接使用ListField配合序列化、反序列化逻辑重写实现，代码更简单易维护：

class OfficeSerializer(serializers.HyperlinkedModelSerializer):
    assigned_hrs = serializers.ListField(
        child=serializers.EmailField(),
        required=False
    )

    def to_representation(self, instance):
        rep = super().to_representation(instance)
        # 序列化输出时替换为HR的邮箱列表
        rep['assigned_hrs'] = [hr.related_user.email for hr in instance.assigned_hrs.all()]
        return rep

    def create(self, validated_data):
        hr_emails = validated_data.pop('assigned_hrs', [])
        office = super().create(validated_data)
        # 根据邮箱列表匹配对应的HRProfile
        matched_hrs = HRProfile.objects.filter(related_user__email__in=hr_emails)
        office.assigned_hrs.set(matched_hrs)
        return office

    def update(self, instance, validated_data):
        hr_emails = validated_data.pop('assigned_hrs', None)
        office = super().update(instance, validated_data)
        if hr_emails is not None:
            matched_hrs = HRProfile.objects.filter(related_user__email__in=hr_emails)
            office.assigned_hrs.set(matched_hrs)
        return office

内容的提问来源于stack exchange，提问作者aSadLamp