django rest 框架 - 将相关对象字段序列化为平面 json dict

Question

The DRF docs describe a way to create nested serializers that can produce dicts like this:

{
    "field1": "val1",
    "field2": "val2",
    "related_obj": {
        "related_obj_field_1": "val1",
        "related_obj_field_2": "val2",
    }
}

but what if I want to create a flat dict that would include all related object fields at the same level as the parent object fields? Like so:

{
    "field1": "val1",
    "field2": "val2",
    "related_obj_field_1": "val1",
    "related_obj_field_2": "val2",

}

Currently I am achieving this using SerializerMethodField :

class SomeSerializer(serializers.ModelSerializer):

    related_obj_field = serializers.SerializerMethodField()

    @staticmethod
    def get_related_obj_field(obj):
        return obj.related_obj.field

But I hope there's a cleaner way to do this.

Answer 1

Well I don't guess there is some in-built Python functionality to flatten a dictionary, but you can write one by using the concept of recursion. The idea is to iterate over all the key, value pairs and see if the value is a dictionary, then call the recursive method, else just update the flatten dictionary with the key, value

def flatten_dict(dictionary, flatten_dictionary):
    for k, v in dictionary.iteritems():
        if not isinstance(v, dict):
            flatten_dictionary[k] = v
        else:
            flatten_dict(v, flatten_dictionary)

f_dict = {}
flatten_dict(d, f_dict)
print f_dict
>>> {'field2': 'val2', 'related_obj_field_1': 'val1', 'related_obj_field_2': 'val2', 'field1': 'val1'}

Answer 2

Clear way for achieving it is using to_representation() method of Serializer. It helps modify serializer's data before converting to json:

class OrganizationUsersSerializer(serializers.ModelSerializer):
    user = UserSerializer()

    class Meta:
        model = UserOrganizations
        fields = ['organization_name', 'user']

    def to_representation(self, instance):
        data = super(OrganizationUsersSerializer, self).to_representation(instance)
        profile = data.pop('user')
        for key, val in profile.items():
            data.update({key: val})
        return data

So, you remove nested object and add its fields(specified in its own Serializer) to the same level.