DRF add computed field based on requesting user

Question

I have a model which has specific many to many fields to the user model. Now, to prevent information leaking, I do not want to return the whole related field though the rest framework. But, I want to create some kind of computed field, such that it return True if the requesting user is in the related field, and False otherwise. Is there a way to make this work?

For example, as it stands now, the rest framework will list the users for "user_like" and "user_bookmark", which I dont want to happen, hence I want to exclude them from the serialized. But I want to have a field, say, named is_liked, which will be true if request.user is in user_like, and false otherwise.

My current setup:

model

class Post(models.Model):
    title = models.CharField(max_length=100)
    user = models.ForeignKey(User, on_delete=models.CASCADE)
    image = models.ImageField(upload_to='dream_photos')

    description = models.TextField(max_length=500)
    date_added = models.DateTimeField(auto_now_add=True)

    user_like = models.ManyToManyField(User, related_name='likes', blank=True)
    user_bookmark = models.ManyToManyField(
        User, related_name='bookmarks', blank=True)
    total_likes = models.PositiveIntegerField(db_index=True, default=0)

    tags = TaggableManager()

serialiser

class PostSerializer(TaggitSerializer, serializers.ModelSerializer):
    tags = TagListSerializerField()

    class Meta:
        model = Dream
        fields = ('title','user', 'image','description','date_added', 'tags', 'total_likes' )

views

class PostViewSet(viewsets.ModelViewSet):
    queryset = Post.objects.prefetch_related('user').all()
    serializer_class = PostSerializer
    permission_classes = [permissions.IsAuthenticated]

    @action(detail=False, methods=['get'], url_path='current-profile', url_name='current-profile')
    def current_user_posts(self, request):
        # I expected this to add the extra field I required
        # But it does not seem to work as expected
        queryset = self.get_queryset().filter(user=request.user).annotate(
            bookmark=(request.user in "user_bookmark"))

        serializer = self.get_serializer(queryset, many=True)
        return Response(serializer.data)

Expected behavior when requesting:

{
        "id": 1,
        "tags": [
            "test"
        ],
        "title": "Tets",
        "image": "http://127.0.0.1:8000/media/post_photos/photo1648638314.jpeg",
    
        "description": "TEst",
        "date_added": "2022-05-20T17:47:55.739431Z",
        "total_likes": 0,
        "user": 1,
        "like": true, // true if current user is in user_like, false otherwise
        "bookmark": true // true if current user is in user_bookmark, false otherwise
    }

Actual behavior

TypeError: 'in ' requires string as left operand, not SimpleLazyObject

Edit 1: The answer from here seems to help to resolve the error. Unfortunately, the annotated field does not seem to be returned by the serializer

the edited view:

class PostViewSet(viewsets.ModelViewSet):
    queryset = Post.objects.prefetch_related('user').all()
    serializer_class = PostSerializer
    permission_classes = [permissions.IsAuthenticated]

    @action(detail=False, methods=['get'], url_path='current-profile', url_name='current-profile')
    def current_user_posts(self, request):
        queryset = self.get_queryset().filter(user=request.user).annotate(
            bookmark=Exists(Post.user_bookmark.through.objects.filter(
                post_id=OuterRef('pk'), user_id=request.user.id))
        )

        serializer = self.get_serializer(queryset, many=True)
        return Response(serializer.data)

Answer 1

Job logs in GitLab CI/CD have a limited length. It's designed for operational reasons, as the jobs can create arbitrary outputs so one could - by mistake or by purpose - output even gigabytes or terabytes of text.

But if you scroll up in the output, you have the abillity to view the full logs by pressing 'Complete Raw'.