我正尝试根据父模型的某些条件（而不是外键）向现有序列化程序添加嵌套序列化程序。用例是返回一个“Research”对象，该对象带有一个“ResearchTemplate”对象数组，这些对象通过在Postgres ArrayField上筛选来标识。

型号

class Research(TimeStampedModel):
    category = models.CharField(max_length=100, choices=RESEARCH_TEMPLATE_CATEGORIES, default='quote')
    body = models.CharField(max_length=1000, blank=True, default='') #The body of text comprising the nugget
    additionaldata = JSONField(null=True) # all the varying values to be merged into a wrapper

    def __str__(self):
        return self.body

    class Meta:
        ordering = ('created',)

class ResearchTemplate(TimeStampedModel):
    template = models.TextField(blank=True, default='')
    category = models.CharField(max_length=100, choices=RESEARCH_TEMPLATE_CATEGORIES, default='quote')
    mergefields = ArrayField(models.CharField(max_length=200), blank=True)

    def save(self, *args,**kwargs):
        merges = re.findall("{{(.*?)}}", self.template) #get all the template names from within the mustaches
        self.mergefields = list(set(merges)) #TODO: Make Unique
        super(TimeStampedModel, self).save(*args,**kwargs)

    def __str__(self):
        return self.wrapper

    class Meta:
        ordering = ('created',)

序列化程序

class ResearchSerializer(serializers.ModelSerializer):
        templates = ResearchTemplateSerializer(many=True)

        class Meta:
            model = Research
            fields = ('id', 'created', 'speaker', 'body', 'templates')

class ResearchTemplateSerializer(serializers.RelatedField):
        def get_queryset(self, values):
            return ResearchTemplate.objects.filter(mergefields__contained_by=['django']) #This must an array of keys from the Research object's JSON field

        class Meta:
            model = ResearchTemplate
            fields = ('id', 'template')

当有一个外键Map序列化程序时，我已经能够嵌套它们，但是我不能用自定义查询集这样做。也许我没有正确地考虑这个问题，我需要在研究模型上有某种形式的“关系”字段。
如何嵌套从具有从父模型指定的值的筛选器返回的所有行的序列化列表？