自定义超链接URL字段，用于DRF序列化程序中的多个查找字段 [英] Custom Hyperlinked URL field for more than one lookup field in a serializer of DRF

问题描述

我正在使用 Django Rest Framework 为我的项目开发Web api.在我的项目中，我需要像这样构建嵌套的api的端点:

I am using Django Rest Framework for developing web api for my project. As in my project i need to build nested api's endpoint like this:

   /users/ - to get all users
   /users/<user_pk> - to get details of a particular user
   /users/<user_pk>/mails/ - to get all mails sent by a user
   /users/<user_pk>/mails/<pk> - to get details of a mail sent by a user

因此，我正在使用 drf-nested-routers 来简化编写和编写代码.维护这些嵌套的资源.

So, i am using drf-nested-routers for ease of writing & maintaing these nested resources.

我希望所有端点的输出都具有超链接，以获取每个嵌套资源的详细信息以及诸如此类的其他详细信息:

I want output of all my endpoints have hyperlink for getting details of each nested resource alongwith other details like this:

[
    {
        "url" : "http://localhost:8000/users/1",
        "first_name" : "Name1",
        "last_name": "Lastname"
        "email" : "[e-mail]",
        "mails": [
            {
                 "url": "http://localhost:8000/users/1/mails/1",
                 "extra_data": "This is a extra data",
                 "mail":{
                     "url": "http://localhost:8000/mails/3"
                     "to" : "[e-mail]",
                     "from": "[e-mail]",
                     "subject": "This is a subject text",
                     "message": "This is a message text"
                 }
            },
            {
             ..........
            }
           ..........
         ]
    }
    .........
]

为此，我根据DRF docs通过继承HyperlinkedModelSerializer编写了序列化器，该序列化器会在序列化过程中自动添加一个url字段作为响应.

To do this, i write my serializers by inherit HyperlinkedModelSerializer as per DRF docs, which automatically adds a url field in response during serialization.

但是，默认情况下，DRF序列化程序不支持为上述嵌套资源生成url，或者我们可以说的不仅仅是单个查找字段.为了处理这种情况，他们建议创建自定义超链接字段.

我遵循了该文档，并编写了用于处理嵌套资源的url生成的自定义代码.我的代码段如下:

I followed this doc, and write custom code for handling url generation of nested resource. My code snippets are as follows:

from django.contrib.auth.models import AbstractUser
from django.db import models

# User model
class User(models.AbstractUser):
    mails = models.ManyToManyField('Mail', through='UserMail', 
                                     through_fields=('user', 'mail'))

# Mail model
class Mail(models.Model):
    to = models.EmailField()
    from = models.EmailField()
    subject = models.CharField()
    message = models.CharField()

# User Mail model
class UserMail(models.Model):
    user = models.ForeignKey('User')
    mail = models.ForeignKey('Mail')
    extra_data = models.CharField()

serializers.py

from rest_framework import serializers
from .models import User, Mail, UserMail
from .serializers_fields import UserMailHyperlink

# Mail Serializer
class MailSerializer(serializers.HyperlinkedModelSerializer):
    class Meta:
        model = Mail
        fields = ('url', 'to', 'from', 'subject', 'message' )

# User Mail Serializer
class UserMailSerializer(serializers.HyperlinkedModelSerializer):
    url = UserMailHyperlink()
    mail = MailSerializer()

    class Meta:
        model = UserMail
        fields = ('url', 'extra_data', 'mail')  


# User Serializer
class UserSerializer(serializers.HyperlinkedModelSerializer):
    mails = UserMailSerializer(source='usermail_set', many=True)

    class Meta:
        model = User
        fields = ('url', 'first_name', 'last_name', 'email', 'mails')

serializers_fields.py

from rest_framework import serializers
from rest_framework.reverse import reverse
from .models import UserMail

class UserMailHyperlink(serializers.HyperlinkedRelatedField):
    view_name = 'user-mail-detail'
    queryset = UserMail.objects.all()

    def get_url(self, obj, view_name, request, format):
        url_kwargs = {
            'user_pk' : obj.user.pk,
            'pk' : obj.pk
        }
        return reverse(view_name, kwargs=url_kwargs, request=request, 
                          format=format)

    def get_object(self, view_name, view_args, view_kwargs):
        lookup_kwargs = {
           'user_pk': view_kwargs['user_pk'],
           'pk': view_kwargs['pk']
        }
        return self.get_queryset().get(**lookup_kwargs)

views.py

from rest_framework import viewsets
from rest_framework.response import Response
from django.shortcuts import get_object_or_404
from .models import User, UserMail
from .serializers import UserSerializer, MailSerializer

class UserViewSet(viewsets.ModelViewSet):
    queryset = User.objects.all()
    serializer_class = UserSerializer

class UserMailViewSet(viewsets.ViewSet):
    queryset = UserMail.objects.all()
    serializer_class = UserMailSerializer

    def list(self, request, user_pk=None):
        mails = self.queryset.filter(user=user_pk)
        serializer = self.serializer_class(mails, many=True,
            context={'request': request}
        )
        return Response(serializer.data)

    def retrieve(self, request, pk=None, user_pk=None):
        queryset = self.queryset.filter(pk=pk, user=user_pk)
        mail = get_object_or_404(queryset, pk=pk)
        serializer = self.serializer_class(mail,
            context={'request': request}
        )
        return Response(serializer.data)

urls.py

from rest_framework.routers import DefaultRouter
from rest_framework_nested import routers
from django.conf.urls import include, url
import views

router = DefaultRouter()
router.register(r'users', views.UserViewSet, base_name='user')

user_router = routers.NestedSimpleRouter(router, r'users',
    lookup='user'
)
user_router.register(r'mails', views.UserMailViewSet,
    base_name='user-mail'
)


urlpatterns = [
    url(r'^', include(router.urls)),
    url(r'^', include(user_router.urls)), 
]

现在，在执行项目并ping /users/ api端点时执行此操作后，出现此错误:

Now, after doing this when i run a project and ping /users/ api endpoint, i got this error:

AttributeError :'UserMail'对象没有属性'url'

AttributeError : 'UserMail' object has no attribute 'url'

我不明白为什么会出现此错误，因为在UserMailSerializer中我添加了url字段作为此序列化程序的属性，因此当它必须进行序列化时，为什么将url字段作为模型. 请帮助我摆脱这个问题.

I couldn't understand why this error came, because in UserMailSerializer i added url field as a attribute of this serializer, so when it has to serialize why it takes url field as a attribute of UserMail model. Please help me out to get away from this problem.

PS::请不要在模型中提出任何重构建议.就像，在这里我只是用user& mail的东西.因此，将此作为测试用例并建议我一个解决方案.

P.S: Please don't suggest any refactoring in models. As, here i just disguised my project real idea with user & mail thing. So, take this as test case and suggest me a solution.

推荐答案

最近我只需要做类似的事情.我的解决方案最终创建了一个自定义关系字段.为了节省空间，病态会(无耻地)指向

I just needed to do something similar lately. My solution ended up making a custom relations field. To save space, Ill simply (and shamelessly) will point to the source code. The most important part is adding lookup_fields and lookup_url_kwargs class attributes which are used internally to both lookup objects and construct the URIs:

class MultiplePKsHyperlinkedIdentityField(HyperlinkedIdentityField):
    lookup_fields = ['pk']
    def __init__(self, view_name=None, **kwargs):
        self.lookup_fields = kwargs.pop('lookup_fields', self.lookup_fields)
        self.lookup_url_kwargs = kwargs.pop('lookup_url_kwargs', self.lookup_fields)
        ...

这反过来又允许这样的用法:

That in turn allows the usage like:

class MySerializer(serializers.ModelSerializer):
    url = MultiplePKsHyperlinkedIdentityField(
        view_name='api:my-resource-detail',
        lookup_fields=['form_id', 'pk'],
        lookup_url_kwargs=['form_pk', 'pk']
    )

这也是我的使用方式源代码.

希望可以帮助您入门.

这篇关于自定义超链接URL字段，用于DRF序列化程序中的多个查找字段的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！