[英]Using Python Dataclass in Django Models
My goal is to create a struct-like object in Django PostgreSQL, such as:我的目标是在 Django PostgreSQL 中创建一个类似 struct 的对象,例如:
specs.coordinate.x
specs.coordinate.y
specs.coordinate.z
x , y , z should therefore be subclasses of coordinate. x , y , z因此应该是坐标的子类。 They should be me mutable as well, wherefore named tuples cannot be used.它们也应该是可变的,因此不能使用命名元组。
I have tried it using the new dataclass:我已经尝试使用新的数据类:
from django.db import models
from dataclasses import dataclass
class Specs(models.Model):
name = models.CharField(max_length=80)
age = models.IntegerField()
@dataclass
class coordinate:
x: float
y: float
z: float
However the coordinate x,y,z items are not visible in pgAdmin:但是坐标 x,y,z项在 pgAdmin 中不可见:
What am I doing wrong?我究竟做错了什么? Is there any better approach for this than with dataclass?有没有比数据类更好的方法?
If your goal is simply to use the syntax: specs.coordinate.x
in your datasheet you can simply use a OneToOne
Relation.如果您的目标只是在数据表中使用语法: specs.coordinate.x
,您可以简单地使用OneToOne
关系。
This also allows for the special case, that either all Coordinates values have to be set or none.这也允许特殊情况,即必须设置所有坐标值或不设置。
from django.db import models
class Coordinate(models.Model):
x = models.FloatField()
y = models.FloatField()
z = models.FloatField()
class Specs(models.Model):
name = models.CharField(max_length=80)
age = models.IntegerField()
coordinate = models.OneToOneField(
Coordinate,
null=True,
on_delete=models.SET_NULL
)
Now you can use Specs just like the a Dataclass.现在您可以像使用 Dataclass 一样使用 Specs。
This solution uses dacite library to achieve support to nested dataclasses.该解决方案使用英安岩库来实现对嵌套数据类的支持。
from dacite import from_dict
from django.db import models
from dataclasses import dataclass, asdict
import json
"""Field that maps dataclass to django model fields."""
class DataClassField(models.CharField):
description = "Map python dataclasses to model."
def __init__(self, dataClass, *args, **kwargs):
self.dataClass = dataClass
if not kwargs['max_length']
kwargs['max_length'] = 1024
super().__init__(*args, **kwargs)
def deconstruct(self):
name, path, args, kwargs = super().deconstruct()
kwargs['dataClass'] = self.dataClass
return name, path, args, kwargs
def from_db_value(self, value, expression, connection):
if value is None:
return value
obj = json.loads(value)
return from_dict(data_class=self.dataClass, data=obj)
def to_python(self, value):
if isinstance(value, self.dataClass):
return value
if value is None:
return value
obj = json.loads(value)
return from_dict(data_class=self.dataClass, data=obj)
def get_prep_value(self, value):
if value is None:
return value
return json.dumps(asdict(value)
@dataclass
class MyDataclass1:
foo: str
@dataclass
class MyDataclass2:
bar: MyDataclass1
class MyModel(models.Model)
dc1 = DataClassField(dataClass=MyDataclass1, null=True)
dc2 = DataClassField(dataClass=MyDataclass2, null=True)
instance = MyModel(dc1=MyDataclass1(foo="foo"))
instance.save()
instance.dc1.foo # Should return "foo"
instance.dc1 # Should return <MyDataclass1 .... at ....>
To add support for custom datatypes (for example datetimes) you need to specify JSON decoding and encoding for these types as represented below:要添加对自定义数据类型(例如日期时间)的支持,您需要为这些类型指定 JSON 解码和编码,如下所示:
def decode_datetime(dt: datetime) -> str:
return dt.strftime('%Y-%m-%dT%H:%M:%S%z')
def encode_datetime(dt: str) -> datetime:
return datetime.strptime(dt, '%Y-%m-%dT%H:%M:%S%z')
def default(o):
if isinstance(o, (datetime.datetime)):
return decode_datetime(o)
else:
print(o)
class DataclassField(models.CharField):
...
def from_db_value(self, value, expression, connection):
...
obj = json.loads(value)
return from_dict(data_class=self.data_class, data=obj,
config=Config(type_hooks={datetime.datetime: encode_datetime}))
def to_python(self, value):
...
return from_dict(data_class=self.data_class, data=obj,
config=Config(type_hooks={datetime.datetime: encode_datetime}))
def get_prep_value(self, value):
...
return json.dumps(asdict(value), sort_keys=True, indent=1, default=default)
I would use django's built-in PointField .我会使用 django 的内置PointField 。 It stores Point objects.它存储 Point 对象。
so you would use所以你会使用
from django.contrib.gis.db import models
class Specs(models.Model):
coordinate = models.PointField
In django you could access it as you would with a normal django model field:在 django 中,您可以像使用普通 django 模型字段一样访问它:
spec.coordinate
And since is a Point
object, you could also do:由于是Point
对象,您还可以执行以下操作:
spec.coordinate.x
spec.coordinate.y
spec.coordinate.z
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.