I need to get at query string values in some viewset logic (in this case, derived from ModelViewSet). Everything I've read, including the Django REST Framework doc, says that request is an attribute of the viewset. But when I actually try to refer to it in code, no matter how I do so, I am shown the runtime error 'AddressViewSet' object has no attribute 'request' . Here's one simplified version of the class definition that triggers the error:
class AddressViewSet(viewsets.ModelViewSet):
def __init__(self, suffix, basename, detail):
attendee = ""
if self.request.query_params.get('attendee'):
attendee = self.request.query_params.get('attendee')
self.serializer_class = AddressSerializer
self.queryset = Address.objects.all()
How does one read request properties in viewset logic in DRF?
You are overriding the incorrect method. The ViewSet like all class based views in Django (All DRF views inherit from django.views.generic.View down the line) is instantiated (The pattern usually seen as View.as_view() internally creates an instance of the class) before the request is even received. This instance is used in the url patterns and when a request matching the url pattern is found a dynamically created function is called for the view which then calls dispatch.
Coming back to the point __init__ is not the correct method to override, if you want to filter the queryset you should instead be overriding get_queryset:
class AddressViewSet(viewsets.ModelViewSet):
queryset = Address.objects.all()
serializer_class = AddressSerializer
def get_queryset(self):
queryset = super().get_queryset()
attendee = ""
if self.request.query_params.get('attendee'):
attendee = self.request.query_params.get('attendee')
# Filter queryset here
return queryset
What I did not understand is that you can override a function get_queryset() but if you do that you do not assign queryset directly - that happens automatically. Once that overridden function is called the request is available as a property of the viewset, so it can be referred to there.
I am using Django detailview. initially, I used the URL pattern
url(r'^todo/details/(?P<pk>[\d]+)', views.todoDetailView.as_view(), name='detail_todo'),
my view is
class todoDetailView(DetailView):
model = models.todo
It worked fine.
In the second case, my URL is
url(r'^todo/details/(?P<id>[\d]+)', views.todoDetailView.as_view(), name='detail_todo'),
this time, I modified my view to
class todoDetailView(DetailView):
model = models.todo
# context_object_name = 'todo_detail'
def get_object(self, **kwargs):
print(kwargs)
return models.todo.objects.get(id=self.kwargs['id'])
It worked fine, I modified the second case to
class todoDetailView(DetailView):
model = models.todo
# context_object_name = 'todo_detail'
def get_queryset(self):
return models.todo.objects.get(id=self.kwargs['id'])
then I get an error,
Generic detail view todoDetailView must be called with either an object pk or a slug.
I know that there is no proper slug or pk provided. So, initially I added get_object() (it worked) but get_queryset() is not working. What is the difference in their working ??
And also if a user is getting details only based on the slug, I read on StackOverflow that
this can be used
slug_field = 'param_name'
slug_url_kwarg = 'param_name'
link - Generic detail view ProfileView must be called with either an object pk or a slug
Can anyone explain me actual working of get_object() and get_queryset() (also get_slug_field() if possible)
Along with the terms slug_field and slug_url_kwarg
Thanks in advance
get_object returns an object (an instance of your model), while get_queryset returns a QuerySet object mapping to a set of potentially multiple instances of your model. In the case of the DetailView (or in fact any class that inherits from the SingleObjectMixin, the purpose of the get_queryset is to restrict the set of objects from which you'll try to fetch your instance.
If you want to show details of an instance, you have to somehow tell Django how to fetch that instance. By default, as the error message indicates, Django calls the get_object method which looks for a pk or slug parameter in the URL. In your first example, where you had pk in the URL, Django managed to fetch your instance automatically, so everything worked fine. In your second example, you overrode the get_object method and manually used the id passed as parameter to fetch the object, which also worked. In the third example, however, you didn't provide a get_object method, so Django executed the default one. SingleObjectMixin's default get_object method didn't find either a pk or a slug, so it failed.
There are multiple ways to fix it:
1. Use pk in the URL
The simplest one is to simply use the code you provided in your first example. I don't know why you were unsatisfied with that, it is perfectly fine. If you're not happy, please explain why in more detail.
2. Override get_object
This is the second solution you provided. It is overkill because if you properly configured your view with the correct options (as you will see in the following alternatives), Django would take care of fetching the object for you.
3. Provide the pk_url_kwarg option
If you really want to use id in the URL for some reason, you can indicate that in your view by specifying the pk_url_kwarg option:
class todoDetailView(DetailView):
model = models.todo
pk_url_kwarg = 'id'
4. Provide the slug_field and slug_url_kwarg options [DON'T DO THIS]
This is a terrible solution because you are not really using a slug, but an id, but it should in theory work. You will basically "fool" Django into using the id field as if it was a slug. I am only mentioning it because you explicitly asked about these options in your question.
class todoDetailView(DetailView):
model = models.todo
slug_field = 'id'
slug_url_kwarg = 'id'
Regarding your get_queryset method: in your example, it doesn't even get to be executed, but in any case it is broken because it returns an individual object instead of a queryset (that's what objects.get does). My guess is you probably don't need a custom get_queryset method at all. This would be useful for example if you had a complex permission system in which different users can only access a different subset of todo objects, which I assume is not your case. Currently, if you provide this get_queryset method, even if everything else is configured properly, you will get an error. Probably an AttributeError saying that the queryset object has no attribute filter (because it will actually be a todo object and not a QuerySet object as Django expects).
The default get_object for DetailView tries to fetch the object using pk or slug from the URL. The simplest thing for you to do is to use (?P<pk>[\d]+) in the URL pattern.
When you override get_object, you are replacing this default behaviour, so you don't get any errors.
When you override get_queryset, Django first runs your get_queryset method an fetches the queryset. It then tries to fetch the object from that queryset using pk or slug, and you get an error because you are not using either of them.
The slug_field and slug_url_kwarg parameters are both defined in the docs. The slug_fields is the name of the field in the model used to fetch the item, and slug_url_kwarg is the name of the parameter in the URL pattern. In your case, you are fetching the object using the primary key (pk/id), so you shouldn't use either of these options.
For your URL pattern with (?P<id>[\d]+), you could use pk_url_kwarg = 'id'. That would tell Django to fetch the object using id from the URL. However, it's much simpler to use your first URL pattern with (?P<pk>[\d]+), then you don't have to override any of the methods/attributes above.
get_object() majorly uses with the generic views which takes pk or id
like: DetailView, UpdateView, DeleteView
where get_queryset() uses with the ListView where we are expecting more objects
Also, get_object() or self.get_object() uses pk as default lookup field or can use slug Field
a little peek at get_object()
if queryset is None:
queryset = self.get_queryset()
# Next, try looking up by primary key.
pk = self.kwargs.get(self.pk_url_kwarg)
slug = self.kwargs.get(self.slug_url_kwarg)
if pk is not None:
queryset = queryset.filter(pk=pk)
# Next, try looking up by slug.
if slug is not None and (pk is None or self.query_pk_and_slug):
slug_field = self.get_slug_field()
queryset = queryset.filter(**{slug_field: slug})
I can't help you with what the error message explicitly means, but get_queryset is used in listviews by get multiple objects, while get_object is used to get a single object (ie DetailView).
If you have a pk you can use to get an object, you don't need to specify a slug field. Slug field are used to filter out objects when you don't have or can't show the primary key publicly. This gives a better explanation of a slug field.
How do I get the current user related object using django's generic class DetailView? Using function based views, I can obtain the object like this:
def foo(request):
friendProfile = get_object_or_404(request.user.profile.friends,username="admin")
What is the equivalent using detail view? I'm guessing it's something related to get_object or get_context_data but I can't fully understand the documents.
Thank you.
request is an attribute of class based views. To get the current user you should use self.request.user.
On a DetailView overload the get_queryset to edit the queryset used to get the object.
I don't know precisely how your friend model is defined but let's assume it has a foreign key pointing to your profile with a related_name set to friend_of. Your view could be:
class FriendProfileDetail(DetailView):
model = Friend
def get_queryset(self):
return self.model.objects.filter(friend_of=self.request.user.profile)
how can I change model in UpdateView for different type of users? I have Student and Teacher inherited from AbstractBaseUser, and I need edit for for them
class EditUser(UpdateView):
success_url = '/success/'
template_name = 'edit-profile.html'
model = Teacher (I need to choose this Teacher or Student)
I know about get_template_names(self) method or get_success_url(self), but can not find any get_model method.
I need somethng like:
def get_model_name(self):
if self.request.user.user_type == 'teacher':
return Teacher
if self.request.user.user_type == 'student':
return Studend
Thank you.
According to Django documentation:
model
The model that this view will display data for. Specifying model = Foo
is effectively the same as specifying queryset = Foo.objects.all(),
where objects stands for Foo’s default manager.
queryset
A QuerySet that represents the objects. If provided, the value of
queryset supersedes the value provided for model.
get_queryset()
Returns the queryset that will be used to retrieve the object that
this view will display. By default, get_queryset() returns the value
of the queryset attribute if it is set, otherwise it constructs a
QuerySet by calling the all() method on the model attribute’s default
manager.
So, all you need is redefine get_queryset method
I'm following the tutorial laid out here to create generic class based views for my API - however, I have run into a small problem. I would like to update the model behind the view partially. I used to be able to do this by using the partial property when I created the serializer. However, it seems that once I start using generic class based views I lose the ability to set whether or not I can allow partial updates to the model. How can I override the partial property of a ModelSerializer? My code is quite simple:
class DejavuUserDetail(generics.RetrieveUpdateAPIView):
'''
Get a user or update a user
'''
lookup_field = "email"
queryset = DejavuUser.objects.all()
serializer_class = UserSerializer
class UserSerializer(serializers.ModelSerializer):
class Meta:
model = DejavuUser
partial = True
def restore_object(self, attrs, instance=None):
"""
Given a dictionary of deserialized field values, either update
an existing model instance, or create a new model instance.
"""
if instance is not None:
#set the required fields and return the instance
I'm trying to access the API via PUT
For partial updates use PATCH.
Also note that partial isn't an option on the serializer metaclass, but instead set on instantiating the serializer.