JSON Question
Serializing deeply nested associations with active_model_serializers
I'm using
Rails 4.2.1active_model_serializers 0.10.0.rc2I'm new to API's and chose
active_model_serializersRABLThe problem I'm having is that I can't seem to include various attributes in multi-level relationships. For instance I have:
Projects
class ProjectSerializer < ActiveModel::Serializer
attributes :id,
:name,
:updated_at
has_many :estimates, include_nested_associations: true
end
and Estimates
class EstimateSerializer < ActiveModel::Serializer
attributes :id,
:name,
:release_version,
:exchange_rate,
:updated_at,
:project_id,
:project_code_id,
:tax_type_id
belongs_to :project
belongs_to :project_code
belongs_to :tax_type
has_many :proposals
end
Proposals
class ProposalSerializer < ActiveModel::Serializer
attributes :id,
:name,
:updated_at,
:estimate_id
belongs_to :estimate
end
When I hit the
/projects/1{
"id": 1,
"name": "123 Park Ave.",
"updated_at": "2015-08-09T02:36:23.950Z",
"estimates": [
{
"id": 1,
"name": "E1",
"release_version": "v1.0",
"exchange_rate": "0.0",
"updated_at": "2015-08-12T04:23:38.183Z",
"project_id": 1,
"project_code_id": 8,
"tax_type_id": 1
}
]
}
However, what I'd like it to produce is:
{
"id": 1,
"name": "123 Park Ave.",
"updated_at": "2015-08-09T02:36:23.950Z",
"estimates": [
{
"id": 1,
"name": "E1",
"release_version": "v1.0",
"exchange_rate": "0.0",
"updated_at": "2015-08-12T04:23:38.183Z",
"project": {
"id": 1,
"name": "123 Park Ave."
},
"project_code": {
"id": 8,
"valuation": 30
},
"tax_type": {
"id": 1,
"name": "no-tax"
},
"proposals": [
{
"id": 1,
"name": "P1",
"updated_at": "2015-08-12T04:23:38.183Z"
},
{
"id": 2,
"name": "P2",
"updated_at": "2015-10-12T04:23:38.183Z"
}
]
}
]
}
Ideally, I'd also like to be able to specify which attributes, associations, and attributes of those associations that are included in each serializer.
I've been looking through the AMS issues, and there does seem to be some back and forth on how this should be handled (or if this kind of functionality is even actually supported) but I'm having difficulty figuring out exactly what the current state is.
- https://github.com/rails-api/active_model_serializers/issues/835
- https://github.com/rails-api/active_model_serializers/issues/968
- https://github.com/rails-api/active_model_serializers/issues/414
- https://github.com/rails-api/active_model_serializers/issues/444
One of the proposed solutions was to override the attribute with a method to call the nested attributes, but that seems to be regarded as a hack so I wanted to avoid it if possible.
Anyway, an example of what of how to go about this or general API advice would be much appreciated.
Answer Source
So this my not be the best or even a good answer, but this is working how I need it to.
While including nested and side-loaded attributes appears to be supported when using the json_api adapter with AMS, I needed support for flat json. In addition, this method worked well because each serializer is specifically generating exactly what I need it to independent of any other serializer and without having to do anything in the controller.
Comments / alternate methods are always welcome.
Project Model
class Project < ActiveRecord::Base
has_many :estimates, autosave: true, dependent: :destroy
end
ProjectsController
def index
@projects = Project.all
render json: @projects
end
ProjectSerializer
class ProjectSerializer < ActiveModel::Serializer
attributes :id,
:name,
:updated_at,
# has_many
:estimates
def estimates
customized_estimates = []
object.estimates.each do |estimate|
# Assign object attributes (returns a hash)
# ===========================================================
custom_estimate = estimate.attributes
# Custom nested and side-loaded attributes
# ===========================================================
# belongs_to
custom_estimate[:project] = estimate.project.slice(:id, :name) # get only :id and :name for the project
custom_estimate[:project_code] = estimate.project_code
custom_estimate[:tax_type] = estimate.tax_type
# has_many w/only specified attributes
custom_estimate[:proposals] = estimate.proposals.collect{|proposal| proposal.slice(:id, :name, :updated_at)}
# ===========================================================
customized_estimates.push(custom_estimate)
end
return customized_estimates
end
end
Result
[
{
"id": 1,
"name": "123 Park Ave.",
"updated_at": "2015-08-09T02:36:23.950Z",
"estimates": [
{
"id": 1,
"name": "E1",
"release_version": "v1.0",
"exchange_rate": "0.0",
"created_at": "2015-08-12T04:23:38.183Z",
"updated_at": "2015-08-12T04:23:38.183Z",
"project": {
"id": 1,
"name": "123 Park Ave."
},
"project_code": {
"id": 8,
"valuation": 30,
"created_at": "2015-08-09T18:02:42.079Z",
"updated_at": "2015-08-09T18:02:42.079Z"
},
"tax_type": {
"id": 1,
"name": "No Tax",
"created_at": "2015-08-09T18:02:42.079Z",
"updated_at": "2015-08-09T18:02:42.079Z"
},
"proposals": [
{
"id": 1,
"name": "P1",
"updated_at": "2015-08-12T04:23:38.183Z"
},
{
"id": 2,
"name": "P2",
"updated_at": "2015-10-12T04:23:38.183Z"
}
]
}
]
}
]
I basically disregarded trying to implement any has_many or belongs_to associations in the serializers and just customized the behavior. I used slice to select specific attributes. Hopefully a more elegant solution will be forth coming.