In the serializer used to create a model I want to rename my model field to (field_name)_id so it's clearer for API consumers that this field is an ID field. The model also has a unique_together constraint on some fields. However when validation runs in the serializer, it fails with a KeyError that the field does not exist:
...rest_framework/utils/serializer_helpers.py", line 148, in __getitem__
return self.fields[key]
KeyError: 'question'
Is there a simple way to get this to work? Minimal example code below.
Model
class MyModel(Model):
question = ForeignKey('uppley.Question', null=False, on_delete=PROTECT)
user = ForeignKey('catalystlab.User', null=False, on_delete=PROTECT)
class Meta:
unique_together = ('question', 'user',)
Serializer
class MyCreateSerializer(ModelSerializer):
question_id = PrimaryKeyRelatedField(
write_only=True,
source='question',
queryset=Question.objects.all(),
)
user_id = PrimaryKeyRelatedField(
write_only=True,
source='user',
queryset=User.objects.all(),
)
class Meta:
model = MyModel
fields = ('question_id', 'user_id',)
test.py - test for demonstration purposes
question = QuestionFactory()
user = UserFactory()
data = {
'question_id': question.id,
'user_id': user.id,
}
serializer = MyCreateSerializer(data=data, write_only=True)
is_valid = serializer.is_valid(raise_exception=True) #KeyError exception raised here.
Previously with DRF 3.10.3 this all worked fine, however with 3.11.0 this now throws a KeyError as mentioned above.
What I have tried
Removing the source field on PrimaryKeyRelatedField for user_id and question_id in the Serializer actually results in bypassing the unique_together validation in DRF and the KeyError is avoided. However the validated data is not mapped back to the original field names (user and question). In this case we have to manually change the keys back to their original names before we can create an instance of the Model from the validated data.
Is there a better way to do this?
You can make a custom serializer like :-
class MyCreateSerializer(serializers.Serializer):
question_id = serializers.PrimaryKeyRelatedField(
write_only=True,
queryset=Question.objects.all(),
)
user_id = PrimaryKeyRelatedField(
write_only=True,
queryset=User.objects.all(),
)
and make custom create function in it for creating object. like :-
def create(self, validated_data):
try:
question = validated_data.get('question_id')
user = validated_data.get('user_id')
instance = MyModel.objects.create(question=question, user=user)
except TypeError:
raise TypeError("Something went wrong while creating objects")
return instance
Related
I am utilizing PrimaryKeyRelatedField to retrieve and write M2M data.
My models.py:
class Task(MP_Node):
...
linked_messages = models.ManyToManyField('workdesk.Message', blank=True, related_name='supported_tasks')
(MP_Node is an abstraction of models.Model from django-treebeard).
My serializers.py:
class TaskSerializer(serializers.ModelSerializer):
...
linked_messages = serializers.PrimaryKeyRelatedField(many=True, required=False, allow_null=True, queryset=Message.objects.all())
class Meta:
model = Task
fields = [..., 'linked_messages']
My api.py:
class TaskViewSet(ModelViewSet):
queryset = Task.objects.all()
serializer_class = TaskSerializer
def create(self, request):
serializer = self.get_serializer(data=request.data)
if serializer.is_valid(raise_exception=True):
print(serializer.data)
With other fields, if the field is set to null=True in the models, or required=False on the serializer, I don't need to include them in the data to instantiate the serializer. However, these fields do not seem to work this way, instead returning KeyError: 'linked_messages' when serializer.data is called.
As a workaround I tried adding setting allow_null, as indicated by the docs, and then manually feed it a null value:
request.data['linked_messages'] = None
but this returns as 404:
"linked_messages":["This field may not be null."]
If I set it to a blank string:
"resources":["Expected a list of items but got type \"str\"."]
If I set it to an empty list, serializer.data again gives me an error:
`TypeError: unhashable type: 'list'`
It seems to have me any way I turn. What am I not understanding about this field?
Use default argument -
linked_messages = serializers.PrimaryKeyRelatedField(
many=True,
queryset=Message.objects.all(),
default=[]
)
# print(serializer.data)
# {'linked_messages': []}
I am running into an issue while deserializing JSON data. One of the field is the customerID and i cannot find a way to user a Serializer class properly.
Here is my code:
class UserProfileData(models.Model):
user = models.ForeignKey(get_user_model(), on_delete=models.CASCADE)
captureDateTime = models.CharField(_('Capture datetime'), blank=True, null=True, max_length=100)
class UserProfileDataSerializer(serializers.ModelSerializer):
class Meta:
model = UserProfileData
fields = "__all__"
The JSON i receive is the following:
{ "customerID": "someUUID", "captureDateTime": "..." }
Here is the current state of my view:
#api_view(['POST'])
def register_profile(request):
data = JSONParser().parse(request)
serializer = UserProfileDataSerializer(data=data)
if serializer.is_valid():
serializer.save()
return JsonResponse(serializer.data, status=201)
return JsonResponse(serializer.errors, status=400)
It fails with the following error:
{'user': [ErrorDetail(string='This field is required.', code='required')]}
I understand i am missing something here, but can't figure out what... Also, i almost forgot to mention the User object has a customerId field. Thanks for your help.
You can pass the source argument to a field to map it to an attribute with a different name. Something like this should work
class UserProfileDataSerializer(serializers.ModelSerializer):
# May want to use a UUIDField based on your question
# either 'user_id' or 'user' as source
customerID = serializers.CharField(source='user') # 'user'
class Meta:
model = UserProfileData
fields = ('captureDateTime', 'customerID')
Here is the working version:
class UserProfileDataSerializer(serializers.ModelSerializer):
customerId = serializers.SlugRelatedField(source='user', queryset=get_user_model().objects.all(), many=False, slug_field='customerId')
class Meta:
model = UserProfileData
fields = ('captureDateTime', 'customerId')
I am new to django rest framework, and I am trying to build api for a todo app using which i can add collaborator to a todo my todo model is as follows:
class Todo(models.Model):
creator = models.ForeignKey(User, on_delete=models.CASCADE)
title = models.CharField(max_length=255)
collaborators = models.ManyToManyField(User,related_name='collaborators')
def __str__(self):
return self.title
Serializer.py
class TodoCreateSerializer(serializers.ModelSerializer):
def save(self, **kwargs):
data = self.validated_data
user = self.context['request'].user
title = data['title']
todo = Todo.objects.create(creator=user, title=title)
return todo
class Meta:
model = Todo
fields = ('id', 'title',)
class UserObjectSerializer(serializers.ModelSerializer):
class Meta:
model = User
fields = ['username']
class TodoSerializer(serializers.ModelSerializer):
collaborators = UserObjectSerializer(many=True, required = False)
class Meta:
model = Todo
fields = ['id', 'title', 'collaborators']
class CollaboratorSerializer(serializers.ModelSerializer):
collaborators = UserObjectSerializer(many=True)
class Meta:
model = Todo
fields = ['collaborators']
and when i try to add collaborator object to my collaborator field using the following method in the viewset
class CollaboratorViewSet(viewsets.ModelViewSet):
serializer_class = CollaboratorSerializer
queryset=Todo.objects.all()
#action(methods=['put'], detail=True, permission_classes=[permissions.IsAuthenticated, ], url_path='add-collaborator', url_name='add_collaborator')
def add_collaborator(self, request, pk=None):
try:
todo = Todo.objects.get(pk=pk)
except Exception:
return Response({'Error':'Invalid Id'})
else:
if request.user == todo.creator:
try:
collborators = request.data["collaborators"]
except KeyError:
return Response({'Error':'Collaborator field not specified in request'}, status = status.HTTP_400_BAD_REQUEST)
else:
collab_list = []
for collab in collborators:
try:
collab_name = collab['username']
except KeyError:
return Response({'Error':'No username provided'}, status = status.HTTP_400_BAD_REQUEST)
else:
new_collab = User.objects.filter(username=collab_name)
if new_collab:
collab_list.append(new_collab)
else:
return Response({'detail':'No user with provided username exists'}, status = status.HTTP_400_BAD_REQUEST)
todo.collaborators.add(*collab_list)
return Response({'Success':'Added collaborators successfully'}, status=status.HTTP_200_OK)
else:
raise PermissionDenied("You are not the creator of this Todo.")
#action(methods=['patch'], detail=True, permission_classes=[permissions.IsAuthenticated, ], url_path='delete-collaborator', url_name='delete_collaborator')
I get an error
django.db.utils.OperationalError: sub-select returns 11 columns - expected 1
but when i use todo.collaborators.set(collaborator_object)
collaborator gets added but this way i am able to add only one collaborator and no effect occurs when i try to add multiple collaborators.
I know that we can add multiple object to m2m relation fields using .add() method but that's not working when i try to add even one object, i also went through the documentation for .add() method but could not find any help.
Sorry for such a long question and thanks for helping in advance!
I managed to solve the error, actually I was was trying to access the collaborators as
new_collab = User.objects.filter(username=collab_name)
which will return a queryset even if there is a unique user with given username
and then i was trying to add collaborator directly to collaborators field which caused this error as it expected an User instance but I was providing a queryset
to avoid that I just replaced the above line as
new_collab = User.objects.get(username=collab_name)
which will return a User object instance if it exists so no error is raised and I am able to add multiple as well as single collaborators.
I'm implementing some voting functionality in an application, where a logged-in user specifies a post that they would like to vote for using a payload like this:
{
"post": 1,
"value": 1
}
As you can tell, the a user field is absent - this is because it gets set in my viewset's perform_create method. I've done this to ensure the vote's user gets set server side. This is what the viewset looks like:
class CreateVoteView(generics.CreateAPIView):
permission_classes = (permissions.IsAuthenticated,)
serializer_class = VoteSerializer
def perform_create(self, serializer):
serializer.save(user=self.request.user)
Here is what the model looks like:
class Vote(models.Model):
post = models.ForeignKey(Post, on_delete=models.CASCADE, related_name='votes', null=False)
user = models.ForeignKey(settings.AUTH_USER_MODEL, on_delete=models.CASCADE, related_name='votes', null=False)
class Values(models.IntegerChoices):
UP = 1, _('Up')
DOWN = -1, _('Down')
value = models.IntegerField(choices=Values.choices, null=False)
class Meta:
unique_together = ('post', 'user')
and finally, the serializer:
class VoteSerializer(serializers.ModelSerializer):
class Meta:
model = Vote
fields = ['post', 'value']
From what I understand, in order for DRF to enforce a unique together validation, both fields (in my case, user and post) must be included in the serializer's fields. As I've mentioned, I'd like to avoid this. Is there any other way of implementing this type of validation logic?
EDIT:
To clarify: the records do not save - I receive this error:
django.db.utils.IntegrityError: (1062, "Duplicate entry '1-3' for key 'api_vote.api_vote_post_id_user_id_73614533_uniq'")
However, my goal is to return a Bad Request instead of an Internal Server Error much like I would when traditionally using a DRF serializer and excluding required fields from a payload.
To output a custom error message due to the IntegrityError, you can override the create method in your serializer:
from django.db import IntegrityError
class VoteSerializer(serializers.ModelSerializer):
class Meta:
model = Vote
fields = ['post', 'value']
def create(self, validated_data):
try:
validated_data['user'] = self.context['request'].user
return super().create(validated_data)
except IntegrityError:
error_msg = {'error': 'IntegrityError message'}
raise serializers.ValidationError(error_msg)
You can try this on your views
try:
MoviesWatchList.objects.create(user=request.user, content=movie)
return response.Response({'message': f'{movie} added in watchlist.'}, status=status.HTTP_201_CREATED)
except:
return response.Response({'message': f'{movie} already added to watchlist.'}, status=status.HTTP_304_NOT_MODIFIED)
I would like tastypie to create a UserProfileResource as a result of me POSTing to a UserResource.
models.py:
class UserProfile(models.Model):
home_address = models.TextField()
user = models.ForeignKey(User, unique=True)
resources.py
class UserProfileResource(ModelResource):
home_address = fields.CharField(attribute='home_address')
class Meta:
queryset = UserProfile.objects.all()
resource_name = 'profile'
excludes = ['id']
include_resource_uri = False
class UserResource(ModelResource):
profile = fields.ToOneField(UserProfileResource, 'profile', full=True)
class Meta:
queryset = User.objects.all()
resource_name = 'user'
allowed_methods = ['get', 'post', 'delete', 'put']
fields = ['username']
filtering = {
'username': ALL,
}
curl command:
curl -v -H "Content-Type: application/json" -X POST --data '{"username":"me", "password":"blahblah", "profile":{"home_address":"somewhere"}}' http://127.0.0.1:8000/api/user/
But I am getting:
Django Version: 1.4
Exception Type: IntegrityError
Exception Value:
null value in column "user_id" violates not-null constraint
It seems like a chicken and egg scenario. I need the user_id to create the UserProfileResource and I need the profile to create the UserResource. Obviously I am doing something very silly.
Can anyone out there shine a light?
Many thanks
johnoc
I modified my code as Pablo suggested below.
class UserProfileResource(StssRessource):
home_address = fields.CharField(attribute='home_address')
user = fields.ToOneField('resources.UserResource', attribute='user', related_name='profile')
class Meta:
queryset = UserProfile.objects.all()
resource_name = 'profile'
class UserResource(ModelResource):
profile = fields.ToOneField('resources.UserProfileResource', attribute='profile', related_name = 'user', full=True)
class Meta:
queryset = User.objects.all()
resource_name = 'user'
But am getting :
Django Version: 1.4
Exception Type: DoesNotExist
Which relates to trying to access the User resource in the ORM and it not existing while its creating the related_objects UserProfileResource. Which is correct. The User ORM isnt created until after the related_objects have been created.
Anyone else seen this??
After 2 days I finally managed to save related resources, the problem was that you have to specify both sides of the relation and their related names, in your case it would be something like that:
class UserProfileResource(ModelResource):
home_address = fields.CharField(attribute='home_address')
user = fields.ToOneField('path.to.api.UserResource', attribute='user', related_name='profile')
#in my case it was a toManyField, I don't know if toOneField works here, you can try toManyField.
class UserResource(ModelResource):
profile = fields.ToOneField(UserProfileResource, 'profile', related_name='user', full=True)
EDIT #2: Finally figured out how to fix things, but unfortunately it requires a bit of subclassing and overrides. Here's how I got it working:
First, create a new field subclass - I called my RelatedToOneField:
from tastypie.bundle import Bundle
from django.core.exceptions import ObjectDoesNotExist, MultipleObjectsReturned
from tastypie.exceptions import ApiFieldError, NotFound
class RelatedToOneField(fields.RelatedField):
"""
Provides access to related data via foreign key.
This subclass requires Django's ORM layer to work properly.
"""
help_text = 'A single related resource. Can be either a URI or set of nested resource data.'
def __init__(self, to, attribute, related_name=None, default=fields.NOT_PROVIDED,
null=False, blank=False, readonly=False, full=False,
unique=False, help_text=None):
super(RelatedToOneField, self).__init__(
to, attribute, related_name=related_name, default=default,
null=null, blank=blank, readonly=readonly, full=full,
unique=unique, help_text=help_text
)
self.fk_resource = None
def dehydrate(self, bundle):
try:
foreign_obj = getattr(bundle.obj, self.attribute)
except ObjectDoesNotExist:
foreign_obj = None
if not foreign_obj:
if not self.null:
raise ApiFieldError("The model '%r' has an empty attribute '%s' and doesn't allow a null value." % (bundle.obj, self.attribute))
return None
self.fk_resource = self.get_related_resource(foreign_obj)
fk_bundle = Bundle(obj=foreign_obj, request=bundle.request)
return self.dehydrate_related(fk_bundle, self.fk_resource)
def hydrate(self, bundle):
value = super(RelatedToOneField, self).hydrate(bundle)
if value is None:
return value
# START OF MODIFIED CONTENT
kwargs = {
'request': bundle.request,
}
if self.related_name:
kwargs['related_obj'] = bundle.obj
kwargs['related_name'] = self.related_name
return self.build_related_resource(value, **kwargs)
#return self.build_related_resource(value, request=bundle.request)
#END OF MODIFIED CONTENT
Then override the obj_create & save_related functions in your "top" model, or in this case, UserResource. Here's the relevant overrides:
def obj_create(self, bundle, request=None, **kwargs):
"""
A ORM-specific implementation of ``obj_create``.
"""
bundle.obj = self._meta.object_class()
for key, value in kwargs.items():
setattr(bundle.obj, key, value)
bundle = self.full_hydrate(bundle)
# Save the main object.
# THIS HAS BEEN MOVED ABOVE self.save_related().
bundle.obj.save()
# Save FKs just in case.
self.save_related(bundle)
# Now pick up the M2M bits.
m2m_bundle = self.hydrate_m2m(bundle)
self.save_m2m(m2m_bundle)
return bundle
def save_related(self, bundle):
"""
Handles the saving of related non-M2M data.
Calling assigning ``child.parent = parent`` & then calling
``Child.save`` isn't good enough to make sure the ``parent``
is saved.
To get around this, we go through all our related fields &
call ``save`` on them if they have related, non-M2M data.
M2M data is handled by the ``ModelResource.save_m2m`` method.
"""
for field_name, field_object in self.fields.items():
if not getattr(field_object, 'is_related', False):
continue
if getattr(field_object, 'is_m2m', False):
continue
if not field_object.attribute:
continue
# Get the object.
# THIS HAS BEEN MOVED ABOVE the field_object.blank CHECK
try:
related_obj = getattr(bundle.obj, field_object.attribute)
except ObjectDoesNotExist:
related_obj = None
# THE 'not related_obj' CHECK HAS BEEN ADDED
if field_object.blank and not related_obj: # ADDED
continue
# Because sometimes it's ``None`` & that's OK.
if related_obj:
# THIS HAS BEEN ADDED
setattr(related_obj, field_object.related_name, bundle.obj) # ADDED
related_obj.save()
setattr(bundle.obj, field_object.attribute, related_obj)
After you add those to your API, everything should work (At least on 0.9.11). The primary part of the fix is the related_obj's weren't be added properly for ToOneField's. My RelatedToOneField subclass implements this check into the field hydrate code.
EDIT: I was wrong again, ToOneField's still don't work in 0.9.12. My gotcha was that there was already a UserProfileResource with the same data I was trying to post in the database. It just grabbed that row and modified it instead of creating something new.
After also spending way too much time on this, it seems that there was a bug for ToOneField's that was fixed in version 0.9.12 (see comments in Pablo's accepted answer for relevant discussion).
If django-tastypie >= 0.9.12, the following should work:
class UserResource(ModelResource):
profile = fields.ToOneField('path.to.api.UserProfileResource', 'profile', related_name='user', full=True)
class UserProfileResource(ModelResource):
home_address = fields.CharField(attribute='home_address')
user = fields.ToOneField(UserResource, attribute='user', related_name='profile')
if django-tastypie <0.9.12, you'll need to do the following:
class UserResource(ModelResource):
profile = fields.ToOneField('path.to.api.UserProfileResource', 'profile', related_name='user', full=True)
class UserProfileResource(ModelResource):
home_address = fields.CharField(attribute='home_address')
user = fields.ToManyField(UserResource, attribute='user', related_name='profile')
Note: switched the order of UserResource & UserProfileResource since that made more sense for my mental model.