[UPDATE] - All changes from the fork have been merged to the upstream. This repo has been archived in favour of the upstream with the recent movement of the project to an external organisation and addition of external contributor(s) to ensure the project remains alive.
Please change your local remote to pull from this repository:
git remote set-url [previous-remote-name] [email protected]:procore-oss/blueprinter.git
to see the previous upstream remote name, run:
git remote -v
Blueprinter is a JSON Object Presenter for Ruby that takes business objects and breaks them down into simple hashes and serializes them to JSON. It can be used in Rails in place of other serializers (like JBuilder or ActiveModelSerializers). It is designed to be simple, direct, and performant.
It heavily relies on the idea of views
which, similar to Rails views, are ways of predefining output for data in different contexts.
Docs can be found here.
Basic
If you have an object you would like serialized, simply create a blueprint. Say, for example, you have a User record with the following attributes [:uuid, :email, :first_name, :last_name, :password, :address]
.
You may define a simple blueprint like so:
class UserBlueprint < Blueprinter::Base
identifier :uuid
fields :first_name, :last_name, :email
end
and then, in your code:
puts UserBlueprint.render(user) # Output is a JSON string
And the output would look like:
{
"uuid": "733f0758-8f21-4719-875f-262c3ec743af",
"email": "[email protected]",
"first_name": "John",
"last_name": "Doe"
}
Collections
You can also pass a collection object or an array to the render method.
puts UserBlueprint.render(User.all)
This will result in JSON that looks something like this:
[
{
"uuid": "733f0758-8f21-4719-875f-262c3ec743af",
"email": "[email protected]",
"first_name": "John",
"last_name": "Doe"
},
{
"uuid": "733f0758-8f21-4719-875f-743af262c3ec",
"email": "[email protected]",
"first_name": "John",
"last_name": "Doe 2"
}
]
You can also configure other classes to be treated like collections. For example, if you are using Mongoid, you can configure it to treat Mongoid::Criteria
objects as collections:
Blueprinter.configure do |config|
config.custom_array_like_classes = [Mongoid::Criteria]
end
Or if you wanted it to treat the Set
class as a collection:
Blueprinter.configure do |config|
config.custom_array_like_classes = [Set]
end
Renaming
You can rename the resulting JSON keys in both fields and associations by using the name
option.
class UserBlueprint < Blueprinter::Base
identifier :uuid
field :email, name: :login
association :user_projects, name: :projects
end
This will result in JSON that looks something like this:
{
"uuid": "92a5c732-2874-41e4-98fc-4123cd6cfa86",
"login": "[email protected]",
"projects": []
}
Views
You may define different outputs by utilizing views:
class UserBlueprint < Blueprinter::Base
identifier :uuid
field :email, name: :login
view :normal do
fields :first_name, :last_name
end
view :extended do
include_view :normal
field :address
association :projects
end
end
A view can include fields from another view by utilizing include_view
and include_views
.
Usage:
puts UserBlueprint.render(user, view: :extended)
Output:
{
"uuid": "733f0758-8f21-4719-875f-262c3ec743af",
"address": "123 Fake St.",
"first_name": "John",
"last_name": "Doe",
"login": "[email protected]"
}
Identifiers
identifier
s are used to specify a field or method name used as an identifier. Usually, this is something like :id
.
Example:
class UserBlueprint < Blueprinter::Base
identifier :uuid
end
Blueprinter identifier
s have a few properties that set them apart from field
s.
- Identifiers are always rendered and considered their own view (the
:identifier
view). - When rendering, identifier fields are always sorted first, before other fields.
If either of the above two developer conveniences are not desired, you can simply create your identifier fields as regular field
s.
Root
You can also optionally pass in a root key to wrap your resulting json in:
class UserBlueprint < Blueprinter::Base
identifier :uuid
field :email, name: :login
view :normal do
fields :first_name, :last_name
end
end
Usage:
puts UserBlueprint.render(user, view: :normal, root: :user)
Output:
{
"user": {
"uuid": "733f0758-8f21-4719-875f-262c3ec743af",
"first_name": "John",
"last_name": "Doe",
"login": "[email protected]"
}
}
Meta Attributes
You can additionally add meta-data to the json as well:
class UserBlueprint < Blueprinter::Base
identifier :uuid
field :email, name: :login
view :normal do
fields :first_name, :last_name
end
end
Usage:
json = UserBlueprint.render(user, view: :normal, root: :user, meta: {links: [
'https://app.mydomain.com',
'https://alternate.mydomain.com'
]})
puts json
Output:
{
"user": {
"uuid": "733f0758-8f21-4719-875f-262c3ec743af",
"first_name": "John",
"last_name": "Doe",
"login": "[email protected]"
},
"meta": {
"links": [
"https://app.mydomain.com",
"https://alternate.mydomain.com"
]
}
}
NOTE: For meta attributes, a root is mandatory.
Exclude Fields
You can specifically choose to exclude certain fields for specific views
class UserBlueprint < Blueprinter::Base
identifier :uuid
field :email, name: :login
view :normal do
fields :first_name, :last_name
end
view :extended do
include_view :normal
field :address
exclude :last_name
end
end
Usage:
puts UserBlueprint.render(user, view: :extended)
Output:
{
"uuid": "733f0758-8f21-4719-875f-262c3ec743af",
"address": "123 Fake St.",
"first_name": "John",
"login": "[email protected]"
}
Use excludes
to exclude multiple fields at once inline.
class UserBlueprint < Blueprinter::Base
identifier :uuid
field :email, name: :login
view :normal do
fields :age, :first_name, :last_name,
end
view :extended do
include_view :normal
field :address
excludes :age, :last_name
end
end
Associations
You may include associated objects. Say for example, a user has projects:
class ProjectBlueprint < Blueprinter::Base
identifier :uuid
field :name
end
class UserBlueprint < Blueprinter::Base
identifier :uuid
field :email, name: :login
view :normal do
fields :first_name, :last_name
association :projects, blueprint: ProjectBlueprint
end
end
Usage:
puts UserBlueprint.render(user, view: :normal)
Output:
{
"uuid": "733f0758-8f21-4719-875f-262c3ec743af",
"first_name": "John",
"last_name": "Doe",
"login": "[email protected]",
"projects": [
{
"uuid": "dca94051-4195-42bc-a9aa-eb99f7723c82",
"name": "Beach Cleanup"
},
{
"uuid": "eb881bb5-9a51-4d27-8a29-b264c30e6160",
"name": "Storefront Revamp"
}
]
}
It is also possible to pass options from one Blueprint to another via an association. For example:
class VehicleBlueprint < Blueprinter::Base
identifier :uuid
field :full_name do |vehicle, options|
"#{vehicle.model} #{options[:trim]}"
end
end
class DriverBlueprint < Blueprinter::Base
identifier :uuid
view :normal do
fields :first_name, :last_name
association :vehicles, blueprint: VehicleBlueprint, options: { trim: 'LX' }
end
end
Default Association/Field Option
By default, an association or field that evaluates to nil
is serialized as nil
. A default serialized value can be specified as an option on the association or field for cases when the association/field could potentially evaluate to nil
. You can also specify a global field_default
or association_default
in the Blueprinter config which will be used for all fields/associations that evaluate to nil.
Blueprinter.configure do |config|
config.field_default = "N/A"
config.association_default = {}
end
class UserBlueprint < Blueprinter::Base
identifier :uuid
view :normal do
field :first_name, default: "N/A"
association :company, blueprint: CompanyBlueprint, default: {}
end
end
default_if
Sometimes, you may want certain "empty" values to pass through to the default value.
Blueprinter provides the ability to treat the following empty types as the default value (or nil
if no default provided).
An empty array or empty active record collection.
An empty hash.
An empty string or symbol.
class UserBlueprint < Blueprinter::Base
identifier :uuid
view :normal do
# If first_name is an empty string, it will become "N/A"
field :first_name, default_if: Blueprinter::EMPTY_STRING, default: "N/A"
# If the projects association collection is empty, it will become nil
association :projects, blueprint: ProjectBlueprint, default_if: Blueprinter::EMPTY_COLLECTION
end
end
Supporting Dynamic Blueprints For Associations
When defining an association, we can dynamically evaluate the blueprint. This comes in handy when adding polymorphic associations, by allowing reuse of existing blueprints.
class Task < ActiveRecord::Base
belongs_to :taskable, polymorphic: true
end
class Project < ActiveRecord::Base
has_many :tasks, as: :taskable
def blueprint
ProjectBlueprint
end
end
class TaskBlueprint < Blueprinter::Base
identifier :uuid
view :normal do
field :title, default: "N/A"
association :taskable, blueprint: ->(taskable) {taskable.blueprint}, default: {}
end
end
NOTE: taskable.blueprint
should return a valid Blueprint class. Currently, has_many
is not supported because of the very nature of polymorphic associations.
Defining A Field Directly In The Blueprint
You can define a field directly in the Blueprint by passing it a block. This is especially useful if the object does not already have such an attribute or method defined, and you want to define it specifically for use with the Blueprint. This is done by passing field
a block. The block also yields the object and any options that were passed from render
. For example:
class UserBlueprint < Blueprinter::Base
identifier :uuid
field :full_name do |user, options|
"#{options[:title_prefix]} #{user.first_name} #{user.last_name}"
end
end
Usage:
puts UserBlueprint.render(user, title_prefix: "Mr")
Output:
{
"uuid": "733f0758-8f21-4719-875f-262c3ec743af",
"full_name": "Mr John Doe"
}
Defining An Identifier Directly In The Blueprint
You can also pass a block to an identifier:
class UserBlueprint < Blueprinter::Base
identifier :uuid do |user, options|
options[:current_user].anonymize(user.uuid)
end
end
Usage:
puts UserBlueprint.render(user, current_user: current_user)
Output:
{
"uuid": "733f0758-8f21-4719-875f-262c3ec743af",
}
Defining An Association Directly In The Blueprint
You can also pass a block to an association:
class ProjectBlueprint < Blueprinter::Base
identifier :uuid
field :name
end
class UserBlueprint < Blueprinter::Base
identifier :uuid
association :projects, blueprint: ProjectBlueprint do |user, options|
user.projects + options[:draft_projects]
end
end
Usage:
puts UserBlueprint.render(user, draft_projects: Project.where(draft: true))
Output:
{
"uuid": "733f0758-8f21-4719-875f-262c3ec743af",
"projects": [
{"uuid": "b426a1e6-ac41-45ab-bfef-970b9a0b4289", "name": "query-console"},
{"uuid": "5bd84d6c-4fd2-4e36-ae31-c137e39be542", "name": "blueprinter"},
{"uuid": "785f5cd4-7d8d-4779-a6dd-ec5eab440eff", "name": "uncontrollable"}
]
}
Passing Additional Properties To #render
render
takes an options hash which you can pass additional properties, allowing you to utilize those additional properties in the field
block. For example:
class UserBlueprint < Blueprinter::Base
identifier :uuid
field(:company_name) do |_user, options|
options[:company].name
end
end
Usage:
puts UserBlueprint.render(user, company: company)
Output:
{
"uuid": "733f0758-8f21-4719-875f-262c3ec743af",
"company_name": "My Company LLC"
}
Conditional Fields
Both the field
and the global Blueprinter Configuration supports :if
and :unless
options that can be used to serialize fields conditionally.
Blueprinter.configure do |config|
config.if = ->(field_name, obj, _options) { !obj[field_name].nil? }
config.unless = ->(field_name, obj, _options) { obj[field_name].nil? }
end
class UserBlueprint < Blueprinter::Base
identifier :uuid
field :last_name, if: ->(_field_name, user, options) { user.first_name != options[:first_name] }
field :age, unless: ->(_field_name, user, _options) { user.age < 18 }
end
NOTE: The field-level setting overrides the global config setting (for the field) if both are set.
Custom Formatting for Dates and Times
To define a custom format for a Date or DateTime field, include the option datetime_format
.
This global or field-level option can be either a string representing the associated strftime
format,
or a Proc which receives the original Date/DateTime object and returns the formatted value.
When using a Proc, it is the Proc's responsibility to handle any errors in formatting.
If a global datetime_format is set (either as a string format or a Proc), this option will be
invoked and used to format all fields that respond to strftime
.
Blueprinter.configure do |config|
config.datetime_format = ->(datetime) { datetime.nil? ? datetime : datetime.strftime("%s").to_i }
end
Usage (String Option):
class UserBlueprint < Blueprinter::Base
identifier :name
field :birthday, datetime_format: "%m/%d/%Y"
end
Output:
{
"name": "John Doe",
"birthday": "03/04/1994"
}
Usage (Proc Option):
class UserBlueprint < Blueprinter::Base
identifier :name
field :birthday, datetime_format: ->(datetime) { datetime.nil? ? datetime : datetime.strftime("%s").to_i }
end
Output:
{
"name": "John Doe",
"birthday": 762739200
}
NOTE: The field-level setting overrides the global config setting (for the field) if both are set.
Transform Classes
Blueprinter provides the ability to specify transform
s on views, which enable further
processing and transforming of resulting view field hashes prior to serialization.
Use transform
to specify one transformer to be included for serialization.
A transformer is a class, extending Blueprinter::Transformer
and implementing the transform
method.
Whatever is returned from this transform
method will end up being the resulting hash passed to serialization.
Create a Transform class extending from Blueprinter::Transformer
class DynamicFieldTransformer < Blueprinter::Transformer
def transform(hash, object, options)
hash.merge!(object.dynamic_fields)
end
end
class User
def custom_columns
self.dynamic_fields #which is an array of some columns
end
def custom_fields
custom_columns.each_with_object({}){|col,result| result[col] = self.send(col)}
end
end
Then specify the transform to use for the view.
class UserBlueprint < Blueprinter::Base
fields :first_name, :last_name
transform DynamicTransformer
end
You can also specify global default transformers. Create one or more transformer classes extending from Blueprinter::Transformer
and set the default_transformers
configuration
class LowerCamelTransformer < Blueprinter::Transformer
def transform(hash, _object, _options)
hash.transform_keys! { |key| key.to_s.camelize(:lower).to_sym }
end
end
Blueprinter.configure do |config|
config.default_transformers = [LowerCamelTransformer]
end
Note: Any transforms specified on a per-blueprint or per-view level will override the default_transformers
in the configuration.
Configurable Extractors
Blueprinter gets a given objects' values from the fields definitions using extractor classes. You can substitute your own extractor class globally or per-field.
For a specific kind of field, create an extractor class extending from Blueprinter::Extractor
class MyFieldExtractor < Blueprinter::Extractor
def extract(_field_name, _object, _local_options, _options={})
# process your obscure_object
_object.clarified
end
end
class MysteryBlueprint < Blueprinter::Base
field :obscure_object, extractor: MyFieldExtractor
end
For a global default, create an extractor class extending from Blueprinter::AutoExtractor
and set the extractor_default
configuration
class MyAutoExtractor < Blueprinter::AutoExtractor
def initialize
super
@my_field_extractor = MyFieldExtractor.new
end
def extractor(object, options)
# dispatch to any class AutoExtractor can, plus more
if detect_obscurity(object)
@my_field_extractor
else
super
end
end
end
Blueprinter.configure do |config|
config.extractor_default = MyAutoExtractor
end
Sorting Fields
By default the response sorts the keys by name. If you want the fields to be sorted in the order of definition, use the below configuration option.
Usage:
Blueprinter.configure do |config|
config.sort_fields_by = :definition
end
class UserBlueprint < Blueprinter::Base
identifier :name
field :email
field :birthday, datetime_format: "%m/%d/%Y"
end
Output:
{
"name": "John Doe",
"email": "[email protected]",
"birthday": "03/04/1994"
}
Deprecations
When functionality in Blueprinter is invoked, that has been deprecated, the default behavior is to write a deprecation notice to stderror.
However, deprecations can be configured to report at three different levels:
Key | Result |
---|---|
:stderr (Default) |
Deprecations will be written to stderror |
:raise |
Deprecations will be raised as Blueprinter::BlueprinterError s |
:silence |
Deprecations will be silenced and will not be raised or logged |
Blueprinter.configure do |config|
config.deprecations = :raise
end
render_as_hash
Same as render
, returns a Ruby Hash.
Usage:
puts UserBlueprint.render_as_hash(user, company: company)
Output:
{
uuid: "733f0758-8f21-4719-875f-262c3ec743af",
company_name: "My Company LLC"
}
render_as_json
Same as render
, returns a Ruby Hash JSONified. This will call JSONify all keys and values.
Usage:
puts UserBlueprint.render_as_json(user, company: company)
Output:
{
"uuid" => "733f0758-8f21-4719-875f-262c3ec743af",
"company_name" => "My Company LLC"
}
Add this line to your application's Gemfile:
gem 'blueprinter'
And then execute:
bundle
Or install it yourself as:
gem install blueprinter
You should also have require 'json'
already in your project if you are not using Rails or if you are not using Oj.
By default, Blueprinter will be calling JSON.generate(object)
internally and it expects that you have require 'json'
already in your project's code. You may use Oj
to generate in place of JSON
like so:
require 'oj' # you can skip this if OJ has already been required.
Blueprinter.configure do |config|
config.generator = Oj # default is JSON
end
Ensure that you have the Oj
gem installed in your Gemfile if you haven't already:
# Gemfile
gem 'oj'
yajl-ruby is a fast and powerful JSON generator/parser. To use yajl-ruby
in place of JSON / OJ
, use:
require 'yajl' # you can skip this if yajl has already been required.
Blueprinter.configure do |config|
config.generator = Yajl::Encoder # default is JSON
config.method = :encode # default is generate
end
NOTE: You should be doing this only if you aren't using yajl-ruby
through the JSON API by requiring yajl/json_gem
. More details here. In this case, JSON.generate
is patched to use Yajl::Encoder.encode
internally.
Please read our Contributing file
You can run tests with bundle exec rake
.
We use Yard for documentation. Here are the following documentation rules:
- Document all public methods we expect to be utilized by the end developers.
- Methods that are not set to private due to ruby visibility rule limitations should be marked with
@api private
.
We use Yard for documentation. Here are the following documentation rules:
- Document all public methods we expect to be utilized by the end developers.
- Methods that are not set to private due to ruby visibility rule limitations should be marked with
@api private
.
To release a new version, change the version number in version.rb
, and update the CHANGELOG.md
. Finally, maintainers need to run bundle exec rake release
, which will automatically create a git tag for the version, push git commits and tags to Github, and push the .gem
file to rubygems.org.
The gem is available as open source under the terms of the MIT License.