Django rest framework 如何使用自定义逻辑创建序列化程序？

输入只有两个参数：

param1=Int，可选

param2=Char，需要

在输出时-

json

，包含4个元素

key1=Char

键2=Int

键3=Int

键4=Int

一切都可以是空的

现在我已经在

get方法的视图中实现了它，但是我想为它创建一个
序列化器
，并将所有的逻辑放在那里，但是我真的不知道在哪里传输逻辑-
验证
，
到_表示

现在看起来像这样

视图.py

class ExampleRetrieveApiView(APIView):
    def get(self, request, *args, **kwargs):
        # A lot of logic related with input parameters.
        return Response({
            'key1': value1,
            'key2': value2,
            'key3': value3,
            'key4': value4
        })


class ExampleRetrieveApiView(APIView):
    serializer_class = ExampleSerializer

    def get(self, request):
        serializer = ExampleSerializer(???)
        return Response(serializer.data)

class ExampleSerializer(serializers.Serializer):
    ???

我需要做些像这样的事情

视图.py

class ExampleRetrieveApiView(APIView):
    def get(self, request, *args, **kwargs):
        # A lot of logic related with input parameters.
        return Response({
            'key1': value1,
            'key2': value2,
            'key3': value3,
            'key4': value4
        })


class ExampleRetrieveApiView(APIView):
    serializer_class = ExampleSerializer

    def get(self, request):
        serializer = ExampleSerializer(???)
        return Response(serializer.data)

class ExampleSerializer(serializers.Serializer):
    ???

序列化程序.py

class ExampleRetrieveApiView(APIView):
    def get(self, request, *args, **kwargs):
        # A lot of logic related with input parameters.
        return Response({
            'key1': value1,
            'key2': value2,
            'key3': value3,
            'key4': value4
        })


class ExampleRetrieveApiView(APIView):
    serializer_class = ExampleSerializer

    def get(self, request):
        serializer = ExampleSerializer(???)
        return Response(serializer.data)

class ExampleSerializer(serializers.Serializer):
    ???

将输入参数作为dict传递。然后，应使用SerializerMethodField（）进行此操作：

from rest_framework import serializers

class ExampleSerializer(serializers.Serializer):
  key1 = serializers.SerializerMethodField()
  key2 = serializers.SerializerMethodField()
  key3 = serializers.SerializerMethodField()
  key4 = serializers.SerializerMethodField()

  def get_key1(self, obj):
    # do some calculations, let's say we want to return input1 multiplied by 2
    # I'm accessing obj data and if it's empty assigning 0, you can add your own check instead
    return obj.get('input1', 0)*2

  def get_key2(self, obj):
    # same deal, do whatever calculations you want with the input data
    return obj.get('input2')

  def get_key3(self, obj):
    # your logic
    return

  def get_key4(self, obj):
    return

对于声明为SerializerMethodField的每个字段，都应该有一个对应的get\u field方法

如何获取输入参数取决于您，您可以通过查询参数或主体数据来实现。
在视图中像这样调用序列化程序，例如使用body参数：

类示例示例RetrieveApiView（APIView）：

视图

class ExampleAPIView(APIView):
    serializer_class = ExampleSerializer

    def get(self, request):
        serializer = ExampleSerializer(data=request.query_params)
        if serializer.is_valid(raise_exception=True):
            return(Response(serializer.data))


class ExampleSerializer(serializers.Serializer):
    param1 = serializers.IntegerField()
    param2 = serializers.CharField()

    def validate(self, data):
        #custom logic
        data['param3'] = 'text'
        #another params
        return data

    def to_representation(self, data):
        return {
            'key1': data['param3']
            'key2': data['param4']
            #another key for representation
        }

序列化程序

class ExampleAPIView(APIView):
    serializer_class = ExampleSerializer

    def get(self, request):
        serializer = ExampleSerializer(data=request.query_params)
        if serializer.is_valid(raise_exception=True):
            return(Response(serializer.data))


class ExampleSerializer(serializers.Serializer):
    param1 = serializers.IntegerField()
    param2 = serializers.CharField()

    def validate(self, data):
        #custom logic
        data['param3'] = 'text'
        #another params
        return data

    def to_representation(self, data):
        return {
            'key1': data['param3']
            'key2': data['param4']
            #another key for representation
        }

这也是可行的，但就我个人而言，我不会在validate中实现实际用于计算/转换输入的逻辑。验证的目的是以某种非标准的方式验证输入。在你的例子中，你实际上是在把输入转换成不同的东西。