Awesome
DRF Writable Nested
This is a writable nested model serializer for Django REST Framework which allows you to create/update your models with related nested data.
The following relations are supported:
- OneToOne (direct/reverse)
- ForeignKey (direct/reverse)
- ManyToMany (direct/reverse excluding m2m relations with through model)
- GenericRelation (this is always only reverse)
Requirements
- Python (3.8, 3.9, 3.10, 3.11, 3.12)
- Django (4.2, 5.0)
- djangorestframework (3.14+)
Installation
pip install drf-writable-nested
Usage
For example, for the following model structure:
from django.db import models
class Site(models.Model):
url = models.CharField(max_length=100)
class User(models.Model):
username = models.CharField(max_length=100)
class AccessKey(models.Model):
key = models.CharField(max_length=100)
class Profile(models.Model):
sites = models.ManyToManyField(Site)
user = models.OneToOneField(User, on_delete=models.CASCADE)
access_key = models.ForeignKey(AccessKey, null=True, on_delete=models.CASCADE)
class Avatar(models.Model):
image = models.CharField(max_length=100)
profile = models.ForeignKey(Profile, related_name='avatars', on_delete=models.CASCADE)
We should create the following list of serializers:
from rest_framework import serializers
from drf_writable_nested.serializers import WritableNestedModelSerializer
class AvatarSerializer(serializers.ModelSerializer):
image = serializers.CharField()
class Meta:
model = Avatar
fields = ('pk', 'image',)
class SiteSerializer(serializers.ModelSerializer):
url = serializers.CharField()
class Meta:
model = Site
fields = ('pk', 'url',)
class AccessKeySerializer(serializers.ModelSerializer):
class Meta:
model = AccessKey
fields = ('pk', 'key',)
class ProfileSerializer(WritableNestedModelSerializer):
# Direct ManyToMany relation
sites = SiteSerializer(many=True)
# Reverse FK relation
avatars = AvatarSerializer(many=True)
# Direct FK relation
access_key = AccessKeySerializer(allow_null=True)
class Meta:
model = Profile
fields = ('pk', 'sites', 'avatars', 'access_key',)
class UserSerializer(WritableNestedModelSerializer):
# Reverse OneToOne relation
profile = ProfileSerializer()
class Meta:
model = User
fields = ('pk', 'profile', 'username',)
Also, you can use NestedCreateMixin
or NestedUpdateMixin
from this package
if you want to support only create or update logic.
For example, we can pass the following data with related nested fields to our main serializer:
data = {
'username': 'test',
'profile': {
'access_key': {
'key': 'key',
},
'sites': [
{
'url': 'http://google.com',
},
{
'url': 'http://yahoo.com',
},
],
'avatars': [
{
'image': 'image-1.png',
},
{
'image': 'image-2.png',
},
],
},
}
user_serializer = UserSerializer(data=data)
user_serializer.is_valid(raise_exception=True)
user = user_serializer.save()
This serializer will automatically create all nested relations and we receive a complete instance with filled data.
user_serializer = UserSerializer(instance=user)
print(user_serializer.data)
{
'pk': 1,
'username': 'test',
'profile': {
'pk': 1,
'access_key': {
'pk': 1,
'key': 'key'
},
'sites': [
{
'pk': 1,
'url': 'http://google.com',
},
{
'pk': 2,
'url': 'http://yahoo.com',
},
],
'avatars': [
{
'pk': 1,
'image': 'image-1.png',
},
{
'pk': 2,
'image': 'image-2.png',
},
],
},
}
It is also possible to pass through values to nested serializers from the call
to the base serializer's save
method. These kwargs
must be of type dict
. E g:
# user_serializer created with 'data' as above
user = user_serializer.save(
profile={
'access_key': {'key': 'key2'},
},
)
print(user.profile.access_key.key)
'key2'
Note: The same value will be used for all nested instances like default value but with higher priority.
Testing
To run unit tests, run:
# Setup the virtual environment
python3 -m venv envname
source envname/bin/activate
pip install django
pip install django-rest-framework
pip install -r requirements.txt
# Run tests
py.test
Known problems with solutions
Validation problem for nested serializers with unique fields on update
We have a special mixin UniqueFieldsMixin
which solves this problem.
The mixin moves UniqueValidator
's from the validation stage to the save stage.
If you want more details, you can read related issues and articles: https://github.com/beda-software/drf-writable-nested/issues/1 http://www.django-rest-framework.org/api-guide/validators/#updating-nested-serializers
Example of usage:
class Child(models.Model):
field = models.CharField(unique=True)
class Parent(models.Model):
child = models.ForeignKey('Child')
class ChildSerializer(UniqueFieldsMixin, serializers.ModelSerializer):
class Meta:
model = Child
class ParentSerializer(NestedUpdateMixin, serializers.ModelSerializer):
child = ChildSerializer()
class Meta:
model = Parent
Note: UniqueFieldsMixin
must be applied only on serializer
which has unique fields.
Mixin ordering
When you are using both mixins
(UniqueFieldsMixin
and NestedCreateMixin
or NestedUpdateMixin
)
you should put UniqueFieldsMixin
ahead.
For example:
class ChildSerializer(UniqueFieldsMixin, NestedUpdateMixin,
serializers.ModelSerializer):
Update problem for nested fields with form-data in PATCH
and PUT
methods
There is a special problem while we try to update any model object with nested fields
within it via PUT
or PATCH
using form-data we can not update it. And it complains
about fields not provided. So far, we came to know that this is also a problem in DRF.
But we can follow a tricky way to solve it at least for now.
See the below solution about the problem
If you want more details, you can read related issues and articles: https://github.com/beda-software/drf-writable-nested/issues/106 https://github.com/encode/django-rest-framework/issues/7262#issuecomment-737364846
Example:
# Models
class Voucher(models.Model):
voucher_number = models.CharField(verbose_name="voucher number", max_length=10, default='')
image = models.ImageField(upload_to="vouchers/images/", null=True, blank=True)
class VoucherRow(models.Model):
voucher = models.ForeignKey(to='voucher.Voucher', on_delete=models.PROTECT, verbose_name='voucher',
related_name='voucherrows', null=True)
account = models.CharField(verbose_name="fortnox account number", max_length=255)
debit = models.DecimalField(verbose_name="amount", decimal_places=2, default=0.00, max_digits=12)
credit = models.DecimalField(verbose_name="amount", decimal_places=2, default=0.00, max_digits=12)
description = models.CharField(verbose_name="description", max_length=100, null=True, blank=True)
# Serializers for these models
class VoucherRowSerializer(WritableNestedModelSerializer):
class Meta:
model = VoucherRow
fields = ('id', 'account', 'debit', 'credit', 'description',)
class VoucherSerializer(serializers.ModelSerializer):
voucherrows = VoucherRowSerializer(many=True, required=False, read_only=True)
class Meta:
model = Voucher
fields = ('id', 'participants', 'voucher_number', 'voucherrows', 'image')
Now if you want to update Voucher
with VoucherRow
and voucher image then you need to do it
using form-data via PUT
or PATCH
request where your voucherrows
fields are nested field.
With the current implementation of the drf-writable-nested
doesn't update it. Because it does
not support something like-
voucherrows[1].account=1120
voucherrows[1].debit=1000.00
voucherrows[1].credit=0.00
voucherrows[1].description='Debited from Bank Account'
voucherrows[2].account=1130
voucherrows[2].debit=0.00
voucherrows[2].credit=1000.00
voucherrows[2].description='Credited to Cash Account'
This is not supported at least for now. So, we can achieve the result in a different way.
Instead of sending the array fields separately in this way we can convert the whole fields
along with values in a json
string like below and set it as value to the field voucherrows
.
"[{\"account\": 1120, \"debit\": 1000.00, \"credit\": 0.00, \"description\": \"Debited from Bank Account\"}, {\"account\": 1130, \"debit\": 0.00, \"credit\": 1000.00, \"description\": \"Credited to Cash Account\"}]"
Now it'll be actually sent as a single field value to the application for the field voucherrows
.
From your views
you need to parse it like below before sending it to the serializer-
class VoucherViewSet(viewsets.ModelViewSet):
serializer_class = VoucherSerializer
queryset = serializer_class.Meta.model.objects.all().order_by('-created_at')
def update(self, request, *args, **kwargs):
request.data.update({'voucherrows': json.loads(request.data.pop('voucherrows', None))})
return super().update(request, *args, **kwargs)
Now, you'll get the voucherrows
field with data in the right format in your serializers.
Similar approach will be also applicable for generic views for django rest framework
Authors
2014-2022, beda.software