我在使用Ember自反模型连接Grape(使用Grape实体序列化)的Json有效负载时遇到问题。模型看起来像这样:
Category = DS.Model.extend {
name: DS.attr 'string',
children: DS.hasMany 'category', inverse: 'parent',
parent: DS.belongsTo 'category', inverse 'children'
}
所以你可以看到我试图在这里建模类别 - 子类别关系。来自端点的示例json响应是:
{
"category": {
"id": 1,
"name": "Sport",
"child_ids": [
5,
6,
8,
7
]
},
"children": [
{
"id": 5,
"name": "Basketball",
"parent_id": 1
},
{
"id": 6,
"name": "Football",
"parent_id": 1
},
{
"id": 8,
"name": "Running",
"parent_id": 1
},
{
"id": 7,
"name": "Volleyball",
"parent_id": 1
}
]
}
有警告信息:
WARNING: Encountered "children" in payload, but no model was found for model name "child" (resolved model name using mateby-front@serializer:-active-model:.modelNameFromPayloadKey("children"))
正如在ActiveModelAdapter的文档中所说,它应该为嵌套资源提供id,并为带有sideloaded数据的父资源提供id。问题是在余烬检查员中有一个类别记录,但没有相关的孩子。
我还尝试制作两个模型:category和subcategory,并提供与其他id / id namings相同的有效负载,我可以看到在这种情况下,所有需要的Subcategory记录,但不是hasMany for Category(两个单独的数据) )。
答案 0 :(得分:0)
问题可能来自于您忘记在第二个反面前放置双点的事实:
更改
Category = DS.Model.extend {
name: DS.attr 'string',
children: DS.hasMany 'category', inverse: 'parent',
parent: DS.belongsTo 'category', inverse 'children' #here
}
到
Category = DS.Model.extend {
name: DS.attr 'string',
children: DS.hasMany 'category', inverse: 'parent',
parent: DS.belongsTo 'category', inverse: 'children' #here
}