Django Rest 框架 : PUT/POST fails in case of blank optional fields in nested serializer

Question

我正在使用 Django 1.7 版和 Django Rest 框架 3.1.3 版。

我有两个相互关联的模型，如下所示:

class Person(models.Model):
    name = models.CharField(max_length=200)     
    GENDER_CHOICES = (
        ('M', 'Male'),
        ('F', 'Female'),
        ('T', 'Transgender'),
    )
    gender = models.CharField(max_length=1, choices=GENDER_CHOICES)
    address = models.CharField(max_length=1000)
    work_phone = models.CharField(max_length=15, blank=True, null=True,
                        validators=[RegexValidator(regex='^[0-9]+$', 
                        message='Phone number must contain only digits', 
                        code='nomatch')])
    home_phone = models.CharField(max_length=15, blank=True, null=True,
                        validators=[RegexValidator(regex='^[0-9]+$', 
                        message='Phone number must contain only digits', 
                        code='nomatch')])
    mobile = models.CharField(max_length=15, blank=True, null=True,
                        validators=[RegexValidator(regex='^[0-9]+$', 
                        message='Phone number must contain only digits', 
                        code='nomatch')])
    email = models.EmailField(blank=True)
    date_of_birth = models.DateField()
    user = OneToOneField(User, null=True, blank=True)   
    DOCUMENT_CHOICES = (
        ('DL', 'Driving License'),
        ('RC', 'Ration card'),
        ('PP', 'Passport'),
        ('NO',  'None'),                 
    )
    additional_document = models.CharField(max_length=2, choices=DOCUMENT_CHOICES)
    document_number = models.CharField(max_length=10, blank=True, null=True)
    document_scan = models.FileField(max_length=1000, blank=True, null=True)
    photo = models.ImageField(blank=True,null=True)

class Admin(models.Model):
    person = OneToOneField(Person)
    # More fields will be added later

因此，Person 包含一个指向 User(Django 的 auth_user 模型)的 OneToOne 字段，而 Admin 包含一个指向 Person 的 OneToOne 字段。但是，我想为最终用户提供一个屏幕来一次性更新所有这些细节(类似于在一个模板中组合多个 ModelForm)。

我已经创建了与这些模型中的每一个相对应的显式序列化程序，它们具有如下嵌套关系。

class UserSerializer(serializers.Serializer):
    id = serializers.IntegerField(label='ID', read_only=True)
    username = serializers.CharField(max_length=30)
    password = serializers.CharField(max_length=128, style={'input_type': 'password'})

class PersonSerializer(serializers.Serializer):
    id = serializers.IntegerField(label="ID", read_only=True)
    user = UserSerializer()
    name = serializers.CharField(max_length=200)
    gender = serializers.ChoiceField(choices=['M', 'F', 'T'])
    address = serializers.CharField(max_length=1000)
    work_phone = serializers.CharField(max_length=15, required=False, allow_null=True)
    home_phone = serializers.CharField(max_length=15, required=False, allow_null=True)
    mobile = serializers.CharField(max_length=15, required=False, allow_null=True)
    email = serializers.EmailField(required=False, allow_blank=True)
    date_of_birth = serializers.DateField()
    additional_document = serializers.ChoiceField(choices=['NO','DL', 'RC', 'PP'])
    document_number = serializers.CharField(max_length=10, required=False, allow_null=True)
    document_scan = serializers.FileField(max_length=1000, required=False, allow_null=True)
    photo = serializers.ImageField(required=False, allow_null=True)


class AdminSerializer(serializers.Serializer):
    id = serializers.IntegerField(label="ID", read_only=True)
    person = PersonSerializer()
    # More fields will be added later

每个序列化程序都定义了一个显式的 create() 和 update() 方法，为了简洁起见，我省略了这些方法。我可以看到这些没有被调用，所以在验证过程中可能会遇到错误。

当我执行 GET 操作时，Person 或 Admin api 端点正确返回所有嵌套对象。

当我对 Person 执行 PUT 或 POST 操作时，它接受可选字段中的空白输入。但是，如果我在对 Admin 对象执行 PUT 或 POST 操作时将相同的字段保留为空白，则会如下标记错误。

我的 PUT 请求是:

{
    "id": 1,
    "person": {
        "id": 1,
        "user": {
            "id": 2,
            "username": "nina.sinha",
            "password": "pbkdf2_sha256$12000$CjmifxpP1DDv$1622S7enIwnT+NcwhMKIEWrq8Uy7gQQJPQQueML21Sg="
        },
        "name": "Nina Sinha",
        "gender": "F",
        "address": "1234, 6th Cross, 7th Main Street",
        "work_phone": "42011567",
        "home_phone": "",
        "mobile": "9854609658",
        "email": "nina.sinha@yahoo.com",
        "date_of_birth": "1985-04-03",
        "additional_document": "NO",
        "document_number": "",
        "document_scan": null,
        "photo": null
    }
}

错误:

HTTP 400 Bad Request
Content-Type: application/json
Vary: Accept
Allow: GET, PUT, PATCH, HEAD, OPTIONS

{
    "person": {
        "document_scan": [
            "The submitted data was not a file. Check the encoding type on the form."
        ],
        "photo": [
            "The submitted data was not a file. Check the encoding type on the form."
        ],
        "document_number": [
            "This field may not be blank."
        ],
        "user": [
            "This field is required."
        ],
        "home_phone": [
            "This field may not be blank."
        ]
    }
}

字段 document_scan、document_number 和 home_phone 留空，但它们是可选字段，AdminSerializer 应该接受这一点。

已经输入了与用户相关的字段，但我仍然收到错误“此字段是必需的”，好像它们丢失了一样。

我在这里缺少什么？如果我使用 ModelSerializers，我会面临同样的问题。

Answer 1

查看您的代码为什么不使用模型序列化器而不是序列化器。在序列化程序级别，您可以根据要求更改字段。您可以在序列化程序中将必填字段更改为非必填字段，并且在模型级别不会出现问题。

序列化程序.py

class UserSerializer(serializers.ModelSerializer):
    class Meta:
        model = User
        fields = ('id', 'username', 'password')

class PersonSerializer(serializers.ModelSerializer):
    user = UserSerializer()
    class Meta:
        model = Person
        fields = ('id', 'user', # specify other fields)

class AdminSerializer(serializers.ModelSerializer):
    person = PersonSerializer()
    class Meta:
        model = Admin
        fields = ('id', 'person', # specify other fields)

如果要将 PersonSerializer 中的用户字段更改为非必需字段，请按如下所示更改代码。

user = UserSerializer(required=False)