require 'jbuilder/jbuilder' require 'jbuilder/blank' require 'jbuilder/key_formatter' require 'jbuilder/errors' require 'multi_json' require 'ostruct' class Jbuilder @@key_formatter = nil @@ignore_nil = false def initialize(options = {}) @attributes = {} @key_formatter = options.fetch(:key_formatter){ @@key_formatter ? @@key_formatter.clone : nil} @ignore_nil = options.fetch(:ignore_nil, @@ignore_nil) yield self if ::Kernel.block_given? end # Yields a builder and automatically turns the result into a JSON string def self.encode(*args, &block) new(*args, &block).target! end BLANK = Blank.new NON_ENUMERABLES = [ ::Struct, ::OpenStruct ].to_set def set!(key, value = BLANK, *args) result = if ::Kernel.block_given? if !_blank?(value) # json.comments @post.comments { |comment| ... } # { "comments": [ { ... }, { ... } ] } _scope{ array! value, &::Proc.new } else # json.comments { ... } # { "comments": ... } _merge_block(key){ yield self } end elsif args.empty? if ::Jbuilder === value # json.age 32 # json.person another_jbuilder # { "age": 32, "person": { ... } value.attributes! else # json.age 32 # { "age": 32 } value end elsif _is_collection?(value) # json.comments @post.comments, :content, :created_at # { "comments": [ { "content": "hello", "created_at": "..." }, { "content": "world", "created_at": "..." } ] } _scope{ array! value, *args } else # json.author @post.creator, :name, :email_address # { "author": { "name": "David", "email_address": "david@loudthinking.com" } } _merge_block(key){ extract! value, *args } end _set_value key, result end def method_missing(*args) if ::Kernel.block_given? set!(*args, &::Proc.new) else set!(*args) end end # Specifies formatting to be applied to the key. Passing in a name of a function # will cause that function to be called on the key. So :upcase will upper case # the key. You can also pass in lambdas for more complex transformations. # # Example: # # json.key_format! :upcase # json.author do # json.name "David" # json.age 32 # end # # { "AUTHOR": { "NAME": "David", "AGE": 32 } } # # You can pass parameters to the method using a hash pair. # # json.key_format! camelize: :lower # json.first_name "David" # # { "firstName": "David" } # # Lambdas can also be used. # # json.key_format! ->(key){ "_" + key } # json.first_name "David" # # { "_first_name": "David" } # def key_format!(*args) @key_formatter = KeyFormatter.new(*args) end # Same as the instance method key_format! except sets the default. def self.key_format(*args) @@key_formatter = KeyFormatter.new(*args) end # If you want to skip adding nil values to your JSON hash. This is useful # for JSON clients that don't deal well with nil values, and would prefer # not to receive keys which have null values. # # Example: # json.ignore_nil! false # json.id User.new.id # # { "id": null } # # json.ignore_nil! # json.id User.new.id # # {} # def ignore_nil!(value = true) @ignore_nil = value end # Same as instance method ignore_nil! except sets the default. def self.ignore_nil(value = true) @@ignore_nil = value end # Turns the current element into an array and yields a builder to add a hash. # # Example: # # json.comments do # json.child! { json.content "hello" } # json.child! { json.content "world" } # end # # { "comments": [ { "content": "hello" }, { "content": "world" } ]} # # More commonly, you'd use the combined iterator, though: # # json.comments(@post.comments) do |comment| # json.content comment.formatted_content # end def child! @attributes = [] unless ::Array === @attributes @attributes << _scope{ yield self } end # Turns the current element into an array and iterates over the passed collection, adding each iteration as # an element of the resulting array. # # Example: # # json.array!(@people) do |person| # json.name person.name # json.age calculate_age(person.birthday) # end # # [ { "name": David", "age": 32 }, { "name": Jamie", "age": 31 } ] # # If you are using Ruby 1.9+, you can use the call syntax instead of an explicit extract! call: # # json.(@people) { |person| ... } # # It's generally only needed to use this method for top-level arrays. If you have named arrays, you can do: # # json.people(@people) do |person| # json.name person.name # json.age calculate_age(person.birthday) # end # # { "people": [ { "name": David", "age": 32 }, { "name": Jamie", "age": 31 } ] } # # If you omit the block then you can set the top level array directly: # # json.array! [1, 2, 3] # # [1,2,3] def array!(collection = [], *attributes) array = if collection.nil? [] elsif ::Kernel.block_given? _map_collection(collection, &::Proc.new) elsif attributes.any? _map_collection(collection) { |element| extract! element, *attributes } else collection.to_a end merge! array end # Extracts the mentioned attributes or hash elements from the passed object and turns them into attributes of the JSON. # # Example: # # @person = Struct.new(:name, :age).new('David', 32) # # or you can utilize a Hash # # @person = { name: 'David', age: 32 } # # json.extract! @person, :name, :age # # { "name": David", "age": 32 }, { "name": Jamie", "age": 31 } # # You can also use the call syntax instead of an explicit extract! call: # # json.(@person, :name, :age) def extract!(object, *attributes) if ::Hash === object _extract_hash_values(object, attributes) else _extract_method_values(object, attributes) end end def call(object, *attributes) if ::Kernel.block_given? array! object, &::Proc.new else extract! object, *attributes end end # Returns the nil JSON. def nil! @attributes = nil end alias_method :null!, :nil! # Returns the attributes of the current builder. def attributes! @attributes end # Merges hash or array into current builder. def merge!(hash_or_array) @attributes = _merge_values(@attributes, hash_or_array) end # Encodes the current builder as JSON. def target! ::MultiJson.dump(@attributes) end private def _extract_hash_values(object, attributes) attributes.each{ |key| _set_value key, object.fetch(key) } end def _extract_method_values(object, attributes) attributes.each{ |key| _set_value key, object.public_send(key) } end def _merge_block(key) current_value = _blank? ? BLANK : @attributes.fetch(_key(key), BLANK) raise NullError.build(key) if current_value.nil? new_value = _scope{ yield self } _merge_values(current_value, new_value) end def _merge_values(current_value, updates) if _blank?(updates) current_value elsif _blank?(current_value) || updates.nil? updates elsif ::Array === updates ::Array === current_value ? current_value + updates : updates elsif ::Hash === current_value current_value.merge(updates) else raise "Can't merge #{updates.inspect} with #{current_value.inspect}" end end def _key(key) @key_formatter ? @key_formatter.format(key) : key.to_s end def _set_value(key, value) raise NullError.build(key) if @attributes.nil? raise ArrayError.build(key) if ::Array === @attributes return if @ignore_nil && value.nil? or _blank?(value) @attributes = {} if _blank? @attributes[_key(key)] = value end def _map_collection(collection) collection.map do |element| _scope{ yield element } end - [BLANK] end def _scope parent_attributes, parent_formatter = @attributes, @key_formatter @attributes = BLANK yield @attributes ensure @attributes, @key_formatter = parent_attributes, parent_formatter end def _is_collection?(object) _object_respond_to?(object, :map, :count) && NON_ENUMERABLES.none?{ |klass| klass === object } end def _blank?(value=@attributes) BLANK == value end def _object_respond_to?(object, *methods) methods.all?{ |m| object.respond_to?(m) } end end require 'jbuilder/railtie' if defined?(Rails)