DRF Custom Related Field

This is a Django REST Framework's PrimaryKeyRelatedField like field which allows you to pass custom fields (instead of default pk) to serialize relation.

Requirements

• Python 3.6+
• Django 2+
• djangorestframework 3+

Installation

pip install drf-custom-related-field

Usage

For example, we have following model structure:

from django.db import models


class Company(models.Model):
    name = models.CharField(max_length=100)
    country = models.CharField(max_length=100)

    def upper_name(self):
        return self.name.upper()


class Address(models.Model):
    street = models.CharField(max_length=255)
    building = models.CharField(max_length=255)

    def full_address(self):
        return f'{self.street}, {self.building}'


class WorkingBuilding(models.Model):
    capacity = models.IntegerField(default=0)
    address = models.ForeignKey(Address, on_delete=models.CASCADE)


class Employee(models.Model):
    username = models.CharField(max_length=100)
    company = models.ForeignKey(Company, on_delete=models.CASCADE)
    workplace = models.ForeignKey(WorkingBuilding, null=True, on_delete=models.CASCADE, related_name='employees')

And we have following instances:

work_address = Address.objects.create(street='Main st.', building='10')
workplace = WorkingBuilding.objects.create(capacity=200, address=work_address)
company = Company.objects.create(name='Great Inc.', country='US', )
employee = Employee.objects.create(username='ckkz', company=company, workplace=workplace)

Now we ready to use our field. To specify which field we want to map for relation, pass required field_name argument to the CustomRelatedField. It can be either str or simple callable (with no arguments) that returns str.

For example:

CustomRelatedField(field_name='name', read_only=True)

def get_field_name():
    return 'some_field'

CustomRelatedField(field_name=get_field_name, read_only=True)

Usage examples:

1. Map custom field for read only.

class EmployeeSerializer(serializers.ModelSerializer):
   company = CustomRelatedField(queryset=Company.objects, field_name='name')

   class Meta:
       model = Employee
       fields = ('username', 'company')

serializer = EmployeeSerializer(employee)
assert serializer.data['company'] == company.name

{
"username": "ckkz",
"company": "Great Inc."
} 

2. Assign new value by custom field (name in this case)

class EmployeeSerializer(serializers.ModelSerializer):
   company = CustomRelatedField(queryset=Company.objects, field_name='name')

   class Meta:
       model = Employee
       fields = ('username', 'company')

new_company = Company.objects.create(name='New Company', country='RU')
serializer = EmployeeSerializer(employee, data={'company': new_company.name}, partial=True)
serializer.is_valid(raise_exception=True)
serializer.save()
employee.refresh_from_db()
assert employee.company_id == new_company.id

{
"username": "ckkz",
"company": "New Company"
} 

3. Use many=True

class WorkingBuildingSerializer(serializers.ModelSerializer):
   employees = CustomRelatedField(field_name='username', many=True, read_only=True)

   class Meta:
       model = WorkingBuilding
       fields = ('capacity', 'employees')

serializer = WorkingBuildingSerializer(workplace)
assert len(serializer.data['employees']) == workplace.employees.count()

{
 "capacity": 200,
 "employees": ["ckkz"]
}

4. Use nested (dotted) relations and callable model fields

class EmployeeSerializer(serializers.ModelSerializer):
   workplace = CustomRelatedField(source='workplace.address', field_name='full_address', read_only=True)

   class Meta:
       model = Employee
       fields = ('username', 'workplace')

serializer = EmployeeSerializer(employee)
assert serializer.data['workplace'] == employee.workplace.address.full_address() 

{
 "username": "ckkz",
 "workplace": "Main st., 10"
}