Django rest APITestCase 客户端在单元测试中将 null 布尔值转换为 false

Question

I'm having troubles testing an endpoint where there is a non null value:

class Status(models.Model):
    """Class to represent the status of an Item. """

    name = models.TextField()

    description = models.TextField(null=True)

    color = models.TextField(null=True)

    in_possession = models.BooleanField()

    class Meta: # pylint: disable=too-few-public-methods
        """Class to represent metadata of the object."""
        ordering = ['pk']

    def __str__(self):
        """String for representing the Model object."""
        return str(self.name)

Which I serialize and add a view with the common ModelView and ModelSerializer :

class StatusSerializer(serializers.ModelSerializer):
    """Serializer for Status."""

    class Meta: # pylint: disable=too-few-public-methods
        """Class to represent metadata of the object."""
        model = Status
        fields = [ 'id', 'name', 'description', 'color', 'in_possession']

class StatusViewset(viewsets.ModelViewSet): # pylint: disable=too-many-ancestors
    """API Endpoint to return the list of status"""
    queryset = Status.objects.all()
    serializer_class = StatusSerializer
    permission_classes = (IsAuthenticated, IsFullUserOrReadOnly)
    pagination_class = None

And now I want to unit test the mandatory fields, so I created this subroutine:

   def get_new_status(self, seed):
        """ This method returns the first status in the fixture"""
        return {
            "name": "name" + str(seed),
            "description": "description."  + str(seed),
            "color": "#32a852",
            "in_possession": True
        }
    def test_mandatory_fields(self):
        """
        Test that the user can not create a status if a mandatory field is missing.
        """

        tests = ['name', 'in_possession']
        for test in tests:
            self.client.force_login(user=self.full_user)
            data = self.get_new_status('test_fields')
            expected = {
                test: [
                    "This field is required."
                ]
            }
            del data[test]
            response = self.client.post(STATUS_PATH, data=data,
                                    HTTP_AUTHORIZATION="Token " + self.token_full_user.key)
            print(response.status_code)
            print(response.data)
            self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST)
            self.assertEqual(response.data, expected)

I loop through an array and it works as expected, I make prints and I can check the boolean field has been deleted from data before sending the request, but in the boolean scenario the Status is created and I get:

AssertionError: 201 != 400

Am I missing something? If I do the same request with a insomnia (like postman) I get the right validation:

{
  "in_possession": [
    "This field is required."
  ]
}

So this makes me think there is something wrong with the client but I have no idea how can I further investigate this.

UPDATE: I tried to simplify it even more:

    def test_mandatory_fields(self):
        """
        Test that the user can not create a status if a mandatory field is missing.
        """


        self.client.force_login(user=self.full_user)
        data = self.get_new_status('test_fields')
        expected = {
            "in_possession": [
                "This field is required."
            ]
        }
        seed = 1
        data = {
        "name": "NewStatus" + str(seed),
        "description": "This status represents items that are in possession and still in use or there is no intention of getting rid of them."  + str(seed),
        "color": "#32a852"
        # "in_possession": True
        }
        print("DATA TO SEND")
        print(data)
        response = self.client.post(STATUS_PATH, data=data,
                                HTTP_AUTHORIZATION="Token " + self.token_full_user.key)
        print(response.status_code)
        print(response.data)
        self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST)
        self.assertEqual(response.data, expected)

Still the same result.

Update 2 : Further debugging following hints from @bdbd:

I have been some hours printing and overwriting methods with prints from serializers.ModelSerializer . It seems to follows this steps:

1. is_valid
2. run_validation
3. to_internal_value So this method with some prints:

    def to_internal_value(self, data):
        """
        Dict of native values <- Dict of primitive datatypes.
        """
        if not isinstance(data, Mapping):
            message = self.error_messages['invalid'].format(
                datatype=type(data).__name__
            )
            raise ValidationError({
                APISettings.NON_FIELD_ERRORS_KEY: [message]
            }, code='invalid')

        ret = OrderedDict()
        errors = OrderedDict()
        fields = self._writable_fields

        for field in fields:
            validate_method = getattr(self, 'validate_' + field.field_name, None)
            print(field.field_name)
            primitive_value = field.get_value(data)
            print('primitive_value is...')
            print(primitive_value)
            try:
                validated_value = field.run_validation(primitive_value)
                print('validated_value is..')
                print(validated_value)
                if validate_method is not None:
                    print('Is not none...')
                    validated_value = validate_method(validated_value)
                    print('validation_value is...')
                    print(validated_value)
            except ValidationError as exc:
                print('except1')
                errors[field.field_name] = exc.detail
            except DjangoValidationError as exc:
                print('except2')
                errors[field.field_name] = get_error_detail(exc)
            except SkipField:
                print('except3')
                pass
            else:
                print('except4')
                print('source_attrs')
                print(field.source_attrs)
                print('validated_value')
                print(validated_value)
                print('before_set_value')
                print(ret)
                set_value(ret, field.source_attrs, validated_value)
                print('after_set_value')
                print(ret)

        if errors:
            print('ValidationError...')
            raise ValidationError(errors)

        print('returning ret...')
        print(ret)
        print('...')
        return ret

So this is the output when I execute the REST query with Insomina:

...
OrderedDict([('name', 'Active_3'), ('description', None)])
color
primitive_value is...
#00ff00
validated_value is..
#00ff00
except4
source_attrs
['color']
validated_value
#00ff00
before_set_value
OrderedDict([('name', 'Active_3'), ('description', None)])
after_set_value
OrderedDict([('name', 'Active_3'), ('description', None), ('color', '#00ff00')])
in_possession
primitive_value is...
<class 'rest_framework.fields.empty'>
except1
ValidationError...
Bad Request: /api/v1/status/
[05/Jun/2021 01:08:56] "POST /api/v1/status/ HTTP/1.1" 400 45

And when I execute it through the Test:

primitive_value is...
#32a852
validated_value is..
#32a852
except4
source_attrs
['color']
validated_value
#32a852
before_set_value
OrderedDict([('name', 'NewStatustest_fields'), ('description', 'This status represents .test_fields')])        
after_set_value
OrderedDict([('name', 'NewStatustest_fields'), ('description', 'This status represents .test_fields'), ('color', '#32a852')])
in_possession
primitive_value is...
False
validated_value is..
False
except4
source_attrs
['in_possession']
validated_value
False
before_set_value
OrderedDict([('name', 'NewStatustest_fields'), ('description', 'This status represents .test_fields'), ('color', '#32a852')])
after_set_value
OrderedDict([('name', 'NewStatustest_fields'), ('description', 'This status represents .test_fields'), ('color', '#32a852'), ('in_possession', False)])
returning ret...
OrderedDict([('name', 'NewStatustest_fields'), ('description', 'This status represents ....test_fields'), ('color', '#32a852'), ('in_possession', False)])
...
value is...
OrderedDict([('name', 'NewStatustest_fields'), ('description', 'This status represents '), ('color', '#32a852'), ('in_possession', False)])
after run validation...
OrderedDict([('name', 'NewStatustest_fields'), ('description', 'This status represents .test_fields'), ('color', '#32a852'), ('in_possession', False)])
OrderedDict([('name', 'NewStatustest_fields'), ('description', 'This status represents .test_fields'), ('color', '#32a852'), ('in_possession', False)])
FAIL
test_update_status (test_status.StatusAPI)

As you can see after the print primitive_value is... I get different results from some strange reason, I have no idea how to continue from there.

Answer 1

I recommend you either define a default or set null=True on your models BooleanField . That might solve the weird behavior.

You most likely use a HTML checkbox for the boolean value on the frontend. The browser will omit all check boxes that are unchecked, leaving you with just the true parameters. Therefor the backend validation might consider all non-existent parameters to be false.

I haven't tested the above recommendation, but I guess you'll improve your code if you just set the default or handle empty boolean parameters yourself instead calling a missing boolean in a form a bad request.

    in_possession = models.BooleanField(default=False)