I am using Django Rest Framework for a project and I have a nested serializer like this:
class TopSerializer(serializers.ModelSerializer):
contact = (something goes here)
email = (something goes here)
For POST, PATCH, PUT, DELETE I want to specify these values with a slug. Suppose each class (Contact, Email) has a member called resource_id and that is my slug. For these methods I could use:
class TopSerializer(serializers.ModelSerializer):
contact = serializers.SlugRelatedField(read_only=False, slug_field='resource_id')
email = serializers.SlugRelatedField(read_only=False, slug_field='resource_id')
However, for GET I want to return the embedded objects too, so I could use:
class TopSerializer(serializers.ModelSerializer):
contact = ContactSerializer(read_only=True)
email = EmailSerializers(read_only=True)
So how do I specify in my serializer that contact can be either a slug or a serialized object? Is there a way to do this with just one serializer or must I have two different serializers and use the request.method in the view to select which serializer I use?
Or, should I use something like this:
class TopSerializer(serializers.ModelSerializer):
contact = ContactSerializer(read_only=True)
email = EmailSerializers(read_only=True)
contact_rid = serializers.SlugRelatedField(read_only=False,slug_field=resource_id,queryset=models.Contact.objects.all())
email_rid = serializers.SlugRelatedField(read_only=False,slug_field=resource_id,queryset=models.Email.objects.all())
This way I can use contact_rid and email_rid for POST/PATCH/PUT/DELETE and get contact and email fields back in GET.
Am I on the right track? Other suggestions?
Check out custom fields https://www.django-rest-framework.org/api-guide/fields/#custom-fields
You could define a custom serializer fields that overrides serializers.Field and overrride to_representation to return the fully serialized object and to_internal_value to mimic the behavior of a slugRelatedField.
You are on the right track!
Use one related field for write and another to read the full object is a good approach if you need more details for related objects.
You can also add to the slug field the flag write_only=True if you want the field is used only for write. However, checking this option will not hint selected objects when you are under an update route in Browseable API
Check this anwser
Related
Hiy!
I want all objects(rows in Test Type) with ModelService
So I could iterate through collection and update a Single row (object)'s attribute with new value
I see getModelService.create(TestModel.class) and getModelService.save()
but will they not create a new object/row rather than update a existing object?right
I don't want to create a new one rather selecting one of the existing matching my criteria and update one attribute of that
can somebody help with List<TestModel> testModels = getModelService.get(TestModel.class) will that return me all rows (collection) of Test Type/Table?
unfortunately I can't test it so need help
Actually I am in validateInterceptor ... and on the basis of this intercepted model changed attribute value I have to update another model attribute value...
thanks
ModelService.create(new TestModel.class) will create a single instance of the specified type and attach it to the modelservice's context.
But it will only be saved to the persistence store when you call modelService.save(newInstance)
ModelService.get() returns a model object but expects a Jalo object as input, (Jalo being the legacy persistence layer of hybris) so that won't work for you.
To retrieve objects you can either write your own queries using the FlexibleSearchService or you can have a look at the DefaultGenericDao which has a bunch of simple find() type of methods.
Typically you would inject the dao like e.g.:
private GenericDao<TestModel> dao;
[...]
public void myMethod()
{
List<TestModel> allTestModels = dao.find();
[...]
}
There are a lot more methods with which you can create WHERE type of statements to restrict your result.
Regarding ValidateInterceptor:
Have a look at the wiki page for the lifecycle of interceptors:
https://wiki.hybris.com/display/release5/Interceptors
It's not a good idea to modify 'all' objects of a type while being an interceptor of that type.
So if you're in an interceptor declared for the Test item type, then don't try to modify the items there.
If you happen to be in a different interceptor and want to modify items of a different type:
E.g. you have Type1 which has a list of Type2 objects in it and in the interceptor for Type1 you want to modify all Type2 objects.
For those scenarios you would have to add the instances of Type2 that you modify to the interceptor context so that those changes will be persisted.
That would be something like:
void onValidate(Test1 model, InterceptorContext ctx) throws InterceptorException
{
...
List<Type2> type2s = dao.find();
for (Type2 type2 : type2s)
{
// do something with it
// then make sure to persist that change
ctx.registerElementFor(type2, PersistenceOperation.SAVE);
[...]
}
}
First of all - i think it's not a good idea, to create/update models in any interceptor, especially in 'validation' one.
Regarding your question:
ModelService in most of the cases works with single model, and
designed for create/update/delete operations.
To retreive all models of certain type, you have to use FlexibleSearchService
Then to update each retrieved TestType model, you can use ModelService's save method.
A query to retreive all TestType models will look like:
SELECT PK FROM {TestType}
You could simply use the Flexible Search Service search by example method, and the model service to save them all. Here is an example using Groovy script, with all products :
import java.util.List
import de.hybris.platform.core.model.product.ProductModel
import de.hybris.platform.servicelayer.search.FlexibleSearchService
import de.hybris.platform.servicelayer.model.ModelService
FlexibleSearchService fsq = spring.getBean("flexibleSearchService")
ModelService ms = spring.getBean("modelService")
ProductModel prd = ms.create(ProductModel.class)
List<ProductModel> products = fsq.getModelsByExample(prd)
//Do Whatever you want with the objects in the List
ms.saveAll(products)
I want to get my object index as a "resource_uri" instead id
I take the usual way I make a model , views , serializers :
class User(BaseModel):
user = models.ForeignKey(settings.AUTH_USER_MODEL)
models.CharField()
class UserSerailizers(serializers.ModelSerializer):
class Meta:
model = User
fields = ('id','user','formatted_address')
classclass UserList(generics.ListCreateAPIView):
queryset = Image.objects.all()
serializer_class = UserSerializer
when i call < my_domain/user/ > I get this response
{
id:1,
name:'toto'
}
but I want to have an answer to this form:
{
'url': my_domain/user/1/
'name': 'toto'
}
Any thoughts?
If you want a hyperlink instead of a primary key in your model representations, you have to use either HyperlinkedModelSerializer or more generic Serializer along with HyperlinkedIdentityField and/or HyperlinkedRelatedField. The former is probably what you are looking for.
The HyperlinkedModelSerializer class is similar to the ModelSerializer class except that it uses hyperlinks to represent relationships, rather than primary keys.
See Django REST framework documentation for more details.
As already commented, you need to use the HyperlinkedModelSerializer as you've shown.
The lookup_field attribute should be inside the Meta class.
And the latest and this is a guess: You just have a ListView for your User model. In order to show the detail for the user, you need also the retrieve method. I would recommend you using the ModelViewset so it automatically implements all methods.
Context
I have the following data structure:
class Birthday(ndb.Model):
day = ndb.IntegerProperty()
month = ndb.IntegerProperty()
year = ndb.IntegerProperty()
class User(ndb.Model):
name = ndb.StringProperty()
birthday = ndb.StructuredProperty(Birthday)
# ... other properties
Problem
When I try to use the populate() method on an instance of User, it gives an error: expecting a Birthday instance instead of a dictionary of params.
If I remove the birthday property, everything works fine: the User instance is populated with the dictionary of params.
Shouldn't the populate() method recognize structured properties and automatically populate them as well?
Any clues?
Thanks
PS: The populate method could also use a forgiving mode on which it ignores unknown properties for which there are references on the params dictionary.
>>Added comments
I'm using a generic REST Handler which is extended for accessing and changing several data types. The extension has to define a method getModel() that returns the model class to access/manipulate. The model class has to implement a few methods, namely create(cls, params).
The POST handler parses params (sent by AngularJS using $resouce -- link below) the following way:
# inside the generic REST Handler
params = json.loads(self.request.body, object_hook=self.datetime_decoder) # parse json params
...
self.getModel().create(params) # invokes the create method of the
The model class implements the create method the following way:
#classmethod
def create(cls, params = None):
obj = cls()
if params:
obj.update(**params)
obj.put()
return True, obj
return False, None
The contents of the JSON dict are:
{"name":"Ana Matos","email":"ana.matos#nvd.com","phone":"+35196983465671","birthday":{"day":1,"month":0,"year":1980},"gender":"FEMALE","groups":["2012/2013"],"serviceProviderId":206133}
JSON contens -- firefox screenshot
AngularJS $resource
Are you reporting a bug or requesting a feature? The populate() method requires its parameter types to match the declared type of the property, which in this case is a Birthday instance.
It would help if you showed the contents of the JSON dict that you are passing to populate() (and exactly how you are passing it).
Possibly the solution is as simple as getting the 'birthday' value from the JSON dict and using it to create a Birthday instance. But I would have to see your code to know for sure.
I have a requirement to design a RESTful Service using RESTEasy. Clients can call this common service with any number of Query Parameters they would want to. My REST code should be able to read these Query Params in some way. For example if I have a book search service, clients can make the following calls.
http://domain.com/context/rest/books/searchBook?bookName=someBookName
http://domain.com/context/rest/books/searchBook?authorName=someAuthor& pubName=somePublisher
http://domain.com/context/rest/books/searchBook?isbn=213243
http://domain.com/context/rest/books/searchBook?authorName=someAuthor
I have to write a service class like below to handle this.
#Path("/books")
public class BookRestService{
// this is what I currently have, I want to change this method to in-take all the
// dynamic parameters that can come
#GET
#Path("/searchBook")
public Response searchBook(#QueryParam("bookName") String bookName,#QueryParam("isbn") String isbn) {
// fetch all such params
// create a search array and pass to backend
}
#POST
#Path("/addBook")
public Response addBook(......) {
//....
}
}
Sorry for the bad format (I couldn't get how code formatting works in this editor!). As you can see, I need to change the method searchBook() so that it will take any number of query parameters.
I saw a similar post here, but couldn't find the right solution.
How to design a RESTful URL for search with optional parameters?
Could any one throw some light on this please?
The best thing to do in this case would be using a DTO containing all the fields of your search criteria. For example, you mentioned 4 distinct parameters.
Book Name (bookName)
Author Name (authorName)
Publisher Name (pubName)
ISBN (isbn)
Create a DTO containing the fields having the following annotations for every property you want to map the parameters to:
public class CriteriaDTO{
#QueryParam("isbn")
private String isbn;
.
.
Other getter and setters of other properties
}
Here is a method doing that for your reference:
#GET
#Produces("application/json")
#Path("/searchBooks")
public ResultDTO search(#Form CriteriaDTO dto){
}
using following URL will populate the CriteriaDTO's property isbn automatically:
your.server.ip:port/URL/Mapping/searchBooks?isbn=123456789&pubName=testing
A similar question was asked here: How do you map multiple query parameters to the fields of a bean on Jersey GET request?
I went with kensen john's answer (UriInfo) instead. It allowed to just iterate through a set to check which parameters were passed.
I am using classed based views with django 1.3 and am trying to figure out how to create an object without using the form. I do not need any user input to create the object but I am still getting an error message that the template is missing. Below is my current view where I have tried to subclass the form_valid method but its not working. Any help would be appreciated.
class ReviewerCreateView(CreateView):
model = Reviewer
def form_valid(self, form):
self.object = form.save(commit=False)
self.object.user = self.request.user
self.object.role = 2
self.object.save()
return HttpResponseRedirect(self.get_success_url())
A CreateView is a specialized view whose purpose is to display a form on GET and validate the form data and create a new object based on the form data on POST.
Since you don't need to display a form and process the form data, a CreateView is not the tool for your job.
You either need a plain old function-based view, or, if you prefer to use a class-based view, derive from View and override get() or post(). For example, adapting your sample code:
class ReviewerCreator(View):
def get(self, request, *args, **kwargs):
Reviewer(user=request.user, role=2).save()
return HttpResponseRedirect('/your_success_url/')
I don't believe a view needs to do anything explicit with a form if it does not need one.
You can instantiate a Reviewer object. It's just a python object.
class ReviewerCreateView(CreateView):
model = Reviewer
self.object.user = self.request.user
self.object.role = 2
self.object.save()
return HttpResponseRedirect(self.get_success_url())