如何通过 DRF APIView 从非常大的查询集中获取随机对象？

Question

我对 django 和 DRF 比较陌生。我正在尝试为几个显示资源或资源和另一个单词（标记）的游戏做一个 API。

我正在尝试在其中一个游戏的视图中获取随机选择的标签对象以及随机选择的资源。目前，我正在获得一个随机资源，其中包含该资源的标签列表（根据 SuggestionsSerializer）。但是，我似乎无法过滤掉随机选择的标签。

这两种方法我都试过了：

order_by("?").first()

以及使用我在下面列出的控制器类中的方法（在过滤掉特定资源的标签之后）。

我怀疑 Tagging 表中的对象太多了，也许我的 get_random_object() 方法对 Tagging 表不够有效。这是问题还是我错过了什么？我还能如何解决这个问题？

提前谢谢你。

models.py

class Resource(models.Model):
    id = models.PositiveIntegerField(null=False, primary_key=True)
    hash_id = models.CharField(max_length=256)
    creators = models.ManyToManyField(Creator)
    titles = models.ManyToManyField(Title)
    created_start = models.DateField(null=True)
    created_end = models.DateField(null=True)
    location = models.CharField(max_length=512, null=True)
    institution_source = models.CharField(max_length=512, blank=True)
    institution = models.CharField(max_length=512, blank=True)
    origin = models.URLField(max_length=256, null=True)
    enabled = models.BooleanField(default=True)
    media_type = models.CharField(max_length=256, default='picture')

    objects = models.Manager()

    def __str__(self):
        return self.hash_id or ''

    @property
    def tags(self):
        tags = self.taggings.values('tag').annotate(count=Count('tag'))

        return tags.values('tag_id', 'tag__name', 'tag__language', 'count')

class Tagging(models.Model):
    user = models.ForeignKey(CustomUser, on_delete=models.SET_NULL, null=True)
    gameround = models.ForeignKey(Gameround, on_delete=models.CASCADE, related_name='taggings')
    resource = models.ForeignKey(Resource, on_delete=models.CASCADE, related_name='taggings')
    tag = models.ForeignKey(Tag, on_delete=models.CASCADE)
    created = models.DateTimeField(editable=False)
    score = models.PositiveIntegerField(default=0)
    origin = models.URLField(max_length=256, blank=True, default='')

    objects = models.Manager()

    def __str__(self):
        return str(self.tag) or ''

serializers.py

class TaggingSerializer(serializers.ModelSerializer):
  tag = TagSerializer(read_only=True)
  resource = ResourceSerializer(read_only=True)
  gameround = GameroundSerializer(read_only=True)

  class Meta:
    model = Tagging
    fields = ('id', 'tag', 'gameround', 'created', 'score', 'resource')

  def create(self, validated_data):
    return Tagging.objects.create(**validated_data)

  def to_representation(self, data):
    data = super().to_representation(data)
    return data

class SuggestionsSerializer(serializers.ModelSerializer):
  creators = CreatorSerializer(many=True)
  titles = TitleSerializer(many=True)
  suggestions = serializers.SerializerMethodField('get_suggestions')

  class Meta:
    model = Resource
    fields = ['id', 'hash_id', 'titles', 'creators', 'suggestions']
    read_only_fields = ['titles', 'creators', 'institution']

  def get_suggestions(self, res):
    suggestions = res.tags
    return suggestions

  def to_representation(self, data):
    data = super().to_representation(data)
    return data

views.py

class GameView(APIView):
    """
    API endpoint that retrieves the game
    """

    def get(self, request, *args, **kwargs):
        
        controller = GameViewController()

        resource_suggestions = Resource.objects.all().filter(id=controller.get_random_object(Resource))
        suggestions_serializer = SuggestionsSerializer(resource_suggestions, many=True)

        # first approach
        tag = Tagging.objects.all().filter(resource=resource_suggestions, id=controller.get_random_object(Tagging))

        # second approach
tag = Tagging.objects.all().filter(resource=resource_suggestions).order_by('?').first()
        tagging_serializer = TaggingSerializer(tag)

return Response({
            'tag': tagging_serializer.data,
            'resource and suggestions': suggestions_serializer.data,
        })

views.py 中的 GameViewController 类

class GameViewController:
...
    def get_random_object(self, MyModel):
        random_object = None
        object_count = MyModel.objects.all().count() + 1
        while not MyModel.objects.all().filter(id=random_object).exists():
            for obj in range(object_count):
                n = random.randint(1, object_count)
                if MyModel.objects.all().filter(id=n).exists():
                    random_object = n
                    return random_object

• 我得到一个“精确查找的 QuerySet 值必须限制为使用切片的一个结果。” order_by('?).first() 方法出错，而 get_random_object() 根本没有结果，因为查询集可能太大，这种方法无法有效工作。

Answer 1

问题出在这行代码
tag=Tagging.objects.all().filter(resource=resource_suggestions).order_by('?').first() resource_suggestions 似乎是一个查询集，至少，根据错误，它不止一个对象。如果您想通过此查询集中的任何内容过滤您的标记对象，您需要这样做：

.filter(resource__in=resource_suggestions)

__in 运算符将根据查询集资源建议中的任何内容进行过滤。