为什么要用序列化组件

当我们做前后端分离的项目~~我们前后端交互一般都选择JSON数据格式，JSON是一个轻量级的数据交互格式。

那么我们给前端数据的时候都要转成json格式，那就需要对我们从数据库拿到的数据进行序列化。

接下来我们看下django序列化和rest_framework序列化的对比~~

Django的序列化方法

class BooksView(View):
    def get(self, request):
        book_list = Book.objects.values("id", "title", "chapter", "pub_time", "publisher")
        book_list = list(book_list)
        # 如果我们需要取外键关联的字段信息 需要循环获取外键 再去数据库查然后拼接成我们想要的
        ret = []
        for book in book_list:
            pub_dict = {}
            pub_obj = Publish.objects.filter(pk=book["publisher"]).first()
            pub_dict["id"] = pub_obj.pk
            pub_dict["title"] = pub_obj.title
            book["publisher"] = pub_dict
            ret.append(book)
        ret = json.dumps(book_list, ensure_ascii=False, cls=MyJson)
        return HttpResponse(ret)


# json.JSONEncoder.default()
# 解决json不能序列化时间字段的问题
class MyJson(json.JSONEncoder):
    def default(self, field):
        if isinstance(field, datetime.datetime):
            return field.strftime(‘%Y-%m-%d %H:%M:%S‘)
        elif isinstance(field, datetime.date):
            return field.strftime(‘%Y-%m-%d‘)
        else:
            return json.JSONEncoder.default(self, field)

.values 序列化结果

from django.core import serializers


# 能够得到我们要的效果 结构有点复杂
class BooksView(View):
    def get(self, request):
        book_list = Book.objects.all()
        ret = serializers.serialize("json", book_list)
        return HttpResponse(ret)

django serializers

DRF序列化的方法

首先，我们要用DRF的序列化，就要遵循人家框架的一些标准，

　　-- Django我们CBV继承类是View，现在DRF我们要用APIView

　　-- Django中返回的时候我们用HTTPResponse，JsonResponse，render ，DRF我们用Response

为什么这么用~我们之后会详细讲~~我们继续来看序列化~~

序列化

class BookSerializer(serializers.Serializer):
    id = serializers.IntegerField()
    title = serializers.CharField(max_length=32)
    CHOICES = ((1, "Linux"), (2, "Django"), (3, "Python"))
    chapter = serializers.ChoiceField(choices=CHOICES, source="get_chapter_display")
    pub_time = serializers.DateField()

第一步声明序列化类

from rest_framework.views import APIView
from rest_framework.response import Response

class BookView(APIView):
    def get(self, request):
        book_list = Book.objects.all()
        ret = BookSerializer(book_list, many=True)
        return Response(ret.data)

第二步序列化对象

外键关系的序列化

# by gaoxin
from rest_framework import serializers
from .models import Book


class PublisherSerializer(serializers.Serializer):
    id = serializers.IntegerField(read_only=True)
    title = serializers.CharField(max_length=32)


class UserSerializer(serializers.Serializer):
    id = serializers.IntegerField(read_only=True)
    name = serializers.CharField(max_length=32)
    age = serializers.IntegerField()


class BookSerializer(serializers.Serializer):
    id = serializers.IntegerField(read_only=True)
    title = serializers.CharField(max_length=32)
    CHOICES = ((1, "Linux"), (2, "Django"), (3, "Python"))
    chapter = serializers.ChoiceField(choices=CHOICES, source="get_chapter_display", read_only=True)
    pub_time = serializers.DateField()

    publisher = PublisherSerializer(read_only=True)
    user = UserSerializer(many=True, read_only=True)

外键关系的序列化

反序列化

当前端给我们发post的请求的时候~前端给我们传过来的数据~我们要进行一些校验然后保存到数据库~

这些校验以及保存工作，DRF的Serializer也给我们提供了一些方法了~~

首先~我们要写反序列化用的一些字段~有些字段要跟序列化区分开~~

Serializer提供了.is_valid() 和.save()方法~~

# serializers.py 文件
class BookSerializer(serializers.Serializer):
    id = serializers.IntegerField(read_only=True)
    title = serializers.CharField(max_length=32)
    CHOICES = ((1, "Linux"), (2, "Django"), (3, "Python"))
    chapter = serializers.ChoiceField(choices=CHOICES, source="get_chapter_display", read_only=True)
    w_chapter = serializers.IntegerField(write_only=True)
    pub_time = serializers.DateField()

    publisher = PublisherSerializer(read_only=True)
    user = UserSerializer(many=True, read_only=True)

    users = serializers.ListField(write_only=True)
    publisher_id = serializers.IntegerField(write_only=True)

    def create(self, validated_data):
        book = Book.objects.create(title=validated_data["title"], chapter=validated_data["w_chapter"], pub_time=validated_data["pub_time"],                                  publisher_id=validated_data["publisher_id"])
        book.user.add(*validated_data["users"])
        return book

反序列化serializer.py

class BookView(APIView):
    def get(self, request):
        book_list = Book.objects.all()
        ret = BookSerializer(book_list, many=True)
        return Response(ret.data)

    def post(self, request):
        # book_obj = request.data
        print(request.data)
        serializer = BookSerializer(data=request.data)
        if serializer.is_valid():
            print(12341253)
            serializer.save()
            return Response(serializer.validated_data)
        else:
            return Response(serializer.errors)

反序列化views.py

当前端给我们发送patch请求的时候，前端传给我们用户要更新的数据，我们要对数据进行部分验证~~

class BookSerializer(serializers.Serializer):
    id = serializers.IntegerField(read_only=True)
    title = serializers.CharField(max_length=32)
    CHOICES = ((1, "Linux"), (2, "Django"), (3, "Python"))
    chapter = serializers.ChoiceField(choices=CHOICES, source="get_chapter_display", read_only=True)
    w_chapter = serializers.IntegerField(write_only=True)
    pub_time = serializers.DateField()

    publisher = PublisherSerializer(read_only=True)
    user = UserSerializer(many=True, read_only=True)

    users = serializers.ListField(write_only=True)
    publisher_id = serializers.IntegerField(write_only=True)

    def create(self, validated_data):
        book = Book.objects.create(title=validated_data["title"], chapter=validated_data["w_chapter"], pub_time=validated_data["pub_time"],
                                   publisher_id=validated_data["publisher_id"])
        book.user.add(*validated_data["users"])
        return book

    def update(self, instance, validated_data):
        instance.title = validated_data.get("title", instance.title)
        instance.chapter = validated_data.get("w_chapter", instance.chapter)
        instance.pub_time = validated_data.get("pub_time", instance.pub_time)
        instance.publisher_id = validated_data.get("publisher_id", instance.publisher_id)
        if validated_data.get("users"):
            instance.user.set(validated_data.get("users"))
        instance.save()
        return instance

PATCH请求serializers.py

class BookView(APIView):
     def patch(self, request):
        print(request.data)
        book_id = request.data["id"]
        book_info = request.data["book_info"]
        book_obj = Book.objects.filter(pk=book_id).first()
        serializer = BookSerializer(book_obj, data=book_info, partial=True)
        if serializer.is_valid():
            serializer.save()
            return Response(serializer.data)
        else:
            return Response(serializer.errors)

PATCH请求views.py

验证

如果我们需要对一些字段进行自定义的验证~DRF也给我们提供了钩子方法~~

class BookSerializer(serializers.Serializer):
    id = serializers.IntegerField(read_only=True)
    title = serializers.CharField(max_length=32)
    # 省略了一些字段 跟上面代码里一样的
    # 。。。。。
     def validate_title(self, value):
        if "python" not in value.lower():
            raise serializers.ValidationError("标题必须含有Python")
        return value

单个字段的验证

class BookSerializer(serializers.Serializer):
    id = serializers.IntegerField(read_only=True)
    title = serializers.CharField(max_length=32)
    CHOICES = ((1, "Linux"), (2, "Django"), (3, "Python"))
    chapter = serializers.ChoiceField(choices=CHOICES, source="get_chapter_display", read_only=True)
    w_chapter = serializers.IntegerField(write_only=True)
    pub_time = serializers.DateField()
    date_added = serializers.DateField(write_only=True)
    # 新增了一个上架时间字段  
    # 省略一些字段。。都是在原基础代码上增加的
    # 。。。。。。

    # 对多个字段进行验证 要求上架日期不能早于出版日期 上架日期要大
    def validate(self, attrs):
        if attrs["pub_time"] > attrs["date_added"]:
            raise serializers.ValidationError("上架日期不能早于出版日期")
        return attrs

多个字段的验证

def my_validate(value):
    if "敏感词汇" in value.lower:
        raise serializers.ValidationError("包含敏感词汇，请重新提交")
    return value


class BookSerializer(serializers.Serializer):
    id = serializers.IntegerField(read_only=True)
    title = serializers.CharField(max_length=32, validators=[my_validate])
    # 。。。。。。

验证器 validators

ModelSerializer

现在我们已经清楚了Serializer的用法，会发现我们所有的序列化跟我们的模型都紧密相关~

那么，DRF也给我们提供了跟模型紧密相关的序列化器~~ModelSerializer~~

　　-- 它会根据模型自动生成一组字段

　　-- 它简单的默认实现了.update()以及.create()方法

定义一个ModelSerializer序列化器

class BookSerializer(serializers.ModelSerializer):
    class Meta:
        model = Book
        fields = "__all__"
        # fields = ["id", "title", "pub_time"]
        # exclude = ["user"]
        # 分别是所有字段 包含某些字段 排除某些字段

定义ModelSerializer

外键关系的序列化

注意：当序列化类MATE中定义了depth时，这个序列化类中引用字段（外键）则自动变为只读

class BookSerializer(serializers.ModelSerializer):
    class Meta:
        model = Book
        fields = "__all__"
        # fields = ["id", "title", "pub_time"]
        # exclude = ["user"]
        # 分别是所有字段 包含某些字段 排除某些字段
        depth = 1
# depth 代表找嵌套关系的第几层

外键关系序列化

自定义字段

我们可以声明一些字段来覆盖默认字段，来进行自定制~

比如我们的选择字段，默认显示的是选择的key，我们要给用户展示的是value。

class BookSerializer(serializers.ModelSerializer):
    chapter = serializers.CharField(source="get_chapter_display", read_only=True)
    
    class Meta:
        model = Book
        fields = "__all__"
        # fields = ["id", "title", "pub_time"]
        # exclude = ["user"]
        # 分别是所有字段 包含某些字段 排除某些字段
        depth = 1

自定义字段

Meta中其它关键字参数

class BookSerializer(serializers.ModelSerializer):
    chapter = serializers.CharField(source="get_chapter_display", read_only=True)

    class Meta:
        model = Book
        fields = "__all__"
        # fields = ["id", "title", "pub_time"]
        # exclude = ["user"]
        # 分别是所有字段 包含某些字段 排除某些字段
        depth = 1
        read_only_fields = ["id"]
        extra_kwargs = {"title": {"validators": [my_validate,]}}

Meta中参数

post以及patch请求

由于depth会让我们外键变成只读，所以我们再定义一个序列化的类，其实只要去掉depth就可以了~~

class BookSerializer(serializers.ModelSerializer):
    chapter = serializers.CharField(source="get_chapter_display", read_only=True)

    class Meta:
        model = Book
        fields = "__all__"
        # fields = ["id", "title", "pub_time"]
        # exclude = ["user"]
        # 分别是所有字段 包含某些字段 排除某些字段
        read_only_fields = ["id"]
        extra_kwargs = {"title": {"validators": [my_validate,]}}

post/patch请求序列化类

SerializerMethodField

外键关联的对象有很多字段我们是用不到的~都传给前端会有数据冗余~就需要我们自己去定制序列化外键对象的哪些字段~~

class BookSerializer(serializers.ModelSerializer):
    chapter = serializers.CharField(source="get_chapter_display", read_only=True)
    user = serializers.SerializerMethodField()
    publisher = serializers.SerializerMethodField()

    def get_user(self, obj):
        # obj是当前序列化的book对象
        users_query_set = obj.user.all()
        return [{"id": user_obj.pk, "name": user_obj.name} for user_obj in users_query_set]

    def get_publisher(self, obj):
        publisher_obj = obj.publisher
        return {"id": publisher_obj.pk, "title": publisher_obj.title}

    class Meta:
        model = Book
        fields = "__all__"
        # fields = ["id", "title", "pub_time"]
        # exclude = ["user"]
        # 分别是所有字段 包含某些字段 排除某些字段
        read_only_fields = ["id"]
        extra_kwargs = {"title": {"validators": [my_validate,]}}

SerializerMethodField

用ModelSerializer改进上面Serializer的完整版

class BookSerializer(serializers.ModelSerializer):
    dis_chapter = serializers.SerializerMethodField(read_only=True)
    users = serializers.SerializerMethodField(read_only=True)
    publishers = serializers.SerializerMethodField(read_only=True)

    def get_users(self, obj):
        # obj是当前序列化的book对象
        users_query_set = obj.user.all()
        return [{"id": user_obj.pk, "name": user_obj.name} for user_obj in users_query_set]

    def get_publishers(self, obj):
        publisher_obj = obj.publisher
        return {"id": publisher_obj.pk, "title": publisher_obj.title}

    def get_dis_chapter(self, obj):
        return obj.get_chapter_display()

    class Meta:
        model = Book
        # fields = "__all__"
        # 字段是有序的
        fields = ["id", "title","dis_chapter", "pub_time", "publishers", "users","chapter", "user", "publisher"]
        # exclude = ["user"]
        # 分别是所有字段 包含某些字段 排除某些字段
        read_only_fields = ["id", "dis_chapter", "users", "publishers"]
        extra_kwargs = {"title": {"validators": [my_validate,]}, "user": {"write_only": True}, "publisher": {"write_only": True},
                        "chapter": {"write_only": True}}

ModelSerializer

Django Rest Framework 视图和路由

DRF的视图

APIView

我们django中写CBV的时候继承的是View，rest_framework继承的是APIView，那么他们两个有什么不同呢~~~

urlpatterns = [
    url(r‘^book$‘, BookView.as_view()),
    url(r‘^book/(?P<id>\d+)$‘, BookEditView.as_view()),
]

我们可以看到，不管是View还是APIView最开始调用的都是as_view()方法~~那我们走进源码看看~~

技术图片

我们能看到，APIView继承了View, 并且执行了View中的as_view()方法，最后把view返回了，用csrf_exempt()方法包裹后去掉了csrf的认证。

那我们看看View中的as_view()方法做了什么~

技术图片

我们看到了~在View中的as_view方法返回了view函数，而view函数执行了self.dispatch()方法~~但是这里的dispatch方法应该是我们APIView中的~~

技术图片

我们去initialize_request中看下把什么赋值给了request，并且赋值给了self.request, 也就是我们在视图中用的request.xxx到底是什么~~

技术图片

我们看到，这个方法返回的是Request这个类的实例对象~~我们注意我们看下这个Request类中的第一个参数request，是我们走我们django的时候的原来的request~

技术图片

我们看到了，这个Request类把原来的request赋值给了self._request, 也就是说以后_request是我们老的request，新的request是我们这个Request类~~

那我们继承APIView之后请求来的数据都在哪呢~~

技术图片

我们用了rest_framework框架以后，我们的request是重新封装的Request类~

request.query_params 存放的是我们get请求的参数

request.data 存放的是我们所有的数据，包括post请求的以及put，patch请求~~~

相比原来的django的request，我们现在的request更加精简，清晰了~~~

现在我们知道了APIView和View的一些区别~~当然还有~~后面我们还会说~~

我们写的视图可能对多个表进行增删改查，就导致我们的视图特别多重复的代码~~

那么我们尝试着来进行封装一下~~

第一次封装

class BookView(APIView):

    def get(self, request):
        query_set = Book.objects.all()
        book_ser = BookSerializer(query_set, many=True)
        return Response(book_ser.data)

    def post(self, request):
        query_set = request.data
        book_ser = BookSerializer(data=query_set)
        if book_ser.is_valid():
            book_ser.save()
            return Response(book_ser.validated_data)
        else:
            return Response(book_ser.errors)


class BookEditView(APIView):

    def get(self, request, id):
        query_set = Book.objects.filter(id=id).first()
        book_ser = BookSerializer(query_set)
        return Response(book_ser.data)

    def patch(self, request, id):
        query_set = Book.objects.filter(id=id).first()
        book_ser = BookSerializer(query_set, data=request.data, partial=True)
        if book_ser.is_valid():
            book_ser.save()
            return Response(book_ser.validated_data)
        else:
            return Response(book_ser.errors)

    def delete(self, request, id):
        query_set = Book.objects.filter(id=id).first()
        if query_set:
            query_set.delete()
            return Response("")
        else:
            return Response("删除的书籍不存在")

APIView视图

class GenericAPIView(APIView):
    queryset = None
    serializer_class = None

    def get_queryset(self):
        return self.queryset.all()

    def get_serializer(self, *args, **kwargs):
        return self.serializer_class(*args, **kwargs)


class ListModelMixin(object):
    def list(self, request, *args, **kwargs):
        queryset = self.get_queryset()
        serializer = self.get_serializer(queryset, many=True)
        return Response(serializer.data)


class CreateModelMixin(object):
    def create(self, request, *args, **kwargs):
        serializer = self.get_serializer(data=request.data)
        if serializer.is_valid():
            serializer.save()
            return Response(serializer.validated_data)
        else:
            return Response(serializer.errors)


class RetrieveModelMixin(object):
    def retrieve(self, request, id, *args, **kwargs):
        book_obj = self.get_queryset().filter(pk=id).first()
        book_ser = self.get_serializer(book_obj)
        return Response(book_ser.data)


class UpdateModelMixin(object):
    def update(self, request, id, *args, **kwargs):
        book_obj = self.get_queryset().filter(pk=id).first()
        book_ser = self.get_serializer(book_obj, data=request.data, partial=True)
        if book_ser.is_valid():
            book_ser.save()
            return Response(book_ser.validated_data)
        else:
            return Response(book_ser.errors)


class DestroyModelMixin(object):
    def destroy(self, request, id, *args, **kwargs):
        queryset = self.get_queryset()
        try:
            queryset.get(pk=id).delete()
            return Response("")
        except Exception as e:
            return Response("信息有误")
# 我们把公共的部分抽出来 这样不管写多少表的增删改查都变的很简单
# 这样封装后我们的视图会变成这样

class BookView(GenericAPIView, ListModelMixin, CreateModelMixin):
    queryset = Book.objects.all()
    serializer_class = BookSerializer

    def get(self, request, *args, **kwargs):
        return self.list(request, *args, **kwargs)

    def post(self, request, *args, **kwargs):
        return self.create(request, *args, **kwargs)


class BookEditView(GenericAPIView, RetrieveModelMixin, UpdateModelMixin, DestroyModelMixin):
    queryset = Book.objects.all()
    serializer_class = BookSerializer

    def get(self, request, id, *args, **kwargs):
        return self.retrieve(request, id, *args, **kwargs)

    def patch(self, request, id, *args, **kwargs):
        return self.update(request, id, *args, **kwargs)

    def destroy(self, request, id, *args, **kwargs):
        return self.delete(request, id, *args, **kwargs)

第一次封装

我们封装的GenericAPIView，包括封装每个方法的类，其实框架都帮我们封装好了~~

我们可以直接继承这些类~~来实现上面的视图~~可是还有没有更简单的方法呢~我们再次封装一下~~

第二次封装

# 上面我们写的继承类太长了~~我们再改改

class ListCreateAPIView(GenericAPIView, ListModelMixin, CreateModelMixin):
    pass


class RetrieveUpdateDestroyAPIView(GenericAPIView, RetrieveModelMixin, UpdateModelMixin, DestroyModelMixin):
    pass


class BookView(ListCreateAPIView):
    queryset = Book.objects.all()
    serializer_class = BookSerializer

    def get(self, request, *args, **kwargs):
        return self.list(request, *args, **kwargs)

    def post(self, request, *args, **kwargs):
        return self.create(request, *args, **kwargs)


class BookEditView(RetrieveUpdateDestroyAPIView):
    queryset = Book.objects.all()
    serializer_class = BookSerializer

    def get(self, request, id, *args, **kwargs):
        return self.retrieve(request, id, *args, **kwargs)

    def patch(self, request, id, *args, **kwargs):
        return self.update(request, id, *args, **kwargs)

    def delete(self, request, id, *args, **kwargs):
        return self.delete(request, id, *args, **kwargs)

第二次封装

这次我们只是让继承变的简单了一点而已，好像并没有什么大的进步~~

我们可不可以把这两个视图合并成一个视图呢~~~框架给我们提供了一个路由传参的方法~~

我们看下ViewSetMixin

技术图片

actions这个默认参数其实就是我们路由可以进行传参了~~~

下面这个循环~可以看出~我们要传的参数是一个字段~key应该是我们的请求方式，value应该对应我们处理的方法~

这样我们每个视图就不用在写函数了~因为已经和内部实现的函数相对应了~

第三次封装

urlpatterns = [
    # url(r‘^book$‘, BookView.as_view()),
    # url(r‘^book/(?P<id>\d+)$‘, BookEditView.as_view()),
    url(r‘^book$‘, BookView.as_view({"get": "list", "post": "create"})),
    url(r‘^book/(?P<pk>\d+)$‘, BookView.as_view({"get": "retrieve", "patch": "update", "delete": "destroy"})),
]

路由urls.py

from rest_framework.viewsets import ViewSetMixin


# class BookView(ViewSetMixin, ListCreateAPIView, RetrieveUpdateDestroyAPIView):
#     queryset = Book.objects.all()
#     serializer_class = BookSerializer
    
    
# 如果我们再定义一个类
class ModelViewSet(ViewSetMixin, ListCreateAPIView, RetrieveUpdateDestroyAPIView):
    pass


class BookView(ModelViewSet):
    queryset = Book.objects.all()
    serializer_class = BookSerializer

第三次封装

我们现在的视图就只要写两行就可以了~~~

其实我们写的所有的视图~框架都帮我们封装好了~

注意一点~~用框架封装的视图~我们url上的那个关键字参数要用pk~~系统默认的~~

奉献一张图来看下我们的继承顺序~~~

技术图片

DRF的路由

我们上面的路由传参写的特别多~~框架也帮我们封装好了~

from .views import BookView
from rest_framework.routers import DefaultRouter

router = DefaultRouter()
router.register(r"book", BookView)

urlpatterns = [
    # url(r‘^book$‘, BookView.as_view()),
    # url(r‘^book/(?P<id>\d+)$‘, BookEditView.as_view()),
    # url(r‘^book$‘, BookView.as_view({"get": "list", "post": "create"})),
    # url(r‘^book/(?P<pk>\d+)$‘, BookView.as_view({"get": "retrieve", "patch": "update", "delete": "destroy"})),

]
urlpatterns += router.urls

路由

我们可以看到~~通过框架我们可以把路由视图都变的非常简单~~

但是需要自定制的时候还是需要我们自己用APIView写~~当不需要那么多路由的时候~也不要用这种路由注册~~

总之~~一切按照业务需要去用~~~

DRF框架（1）

标签：time 轻量模型简单 post请求用户 ica chapter b16

原文地址：https://www.cnblogs.com/abdm-989/p/11872826.html

踩

(0)

评论一句话评论（0）