1
0
Fork 0
mirror of https://github.com/rails/rails.git synced 2022-11-09 12:12:34 -05:00
rails--rails/activerecord/lib/active_record/inheritance.rb

254 lines
9.8 KiB
Ruby
Raw Normal View History

# frozen_string_literal: true
require "active_support/core_ext/hash/indifferent_access"
2013-01-06 04:40:12 -05:00
module ActiveRecord
# == Single table inheritance
#
# Active Record allows inheritance by storing the name of the class in a column that by
# default is named "type" (can be changed by overwriting <tt>Base.inheritance_column</tt>).
# This means that an inheritance looking like this:
#
# class Company < ActiveRecord::Base; end
# class Firm < Company; end
# class Client < Company; end
# class PriorityClient < Client; end
#
# When you do <tt>Firm.create(name: "37signals")</tt>, this record will be saved in
# the companies table with type = "Firm". You can then fetch this row again using
# <tt>Company.where(name: '37signals').first</tt> and it will return a Firm object.
#
# Be aware that because the type column is an attribute on the record every new
# subclass will instantly be marked as dirty and the type column will be included
# in the list of changed attributes on the record. This is different from non
# Single Table Inheritance(STI) classes:
#
# Company.new.changed? # => false
# Firm.new.changed? # => true
# Firm.new.changes # => {"type"=>["","Firm"]}
#
# If you don't have a type column defined in your table, single-table inheritance won't
# be triggered. In that case, it'll work just like normal subclasses with no special magic
# for differentiating between them or reloading the right type with find.
#
# Note, all the attributes for all the cases are kept in the same table. Read more:
# https://www.martinfowler.com/eaaCatalog/singleTableInheritance.html
#
module Inheritance
extend ActiveSupport::Concern
included do
2013-05-27 14:51:06 -04:00
# Determines whether to store the full constant name including namespace when using STI.
# This is true, by default.
class_attribute :store_full_sti_class, instance_writer: false, default: true
end
module ClassMethods
# Determines if one of the attributes passed in is the inheritance column,
# and if the inheritance column is attr accessible, it initializes an
2013-05-27 14:51:06 -04:00
# instance of the given subclass instead of the base class.
def new(attributes = nil, &block)
if abstract_class? || self == Base
raise NotImplementedError, "#{self} is an abstract class and cannot be instantiated."
end
if has_attribute?(inheritance_column)
subclass = subclass_from_attributes(attributes)
if subclass.nil? && base_class == self
subclass = subclass_from_attributes(column_defaults)
end
end
if subclass && subclass != self
subclass.new(attributes, &block)
else
super
end
end
2013-05-27 14:51:06 -04:00
# Returns +true+ if this does not need STI type condition. Returns
# +false+ if STI type condition needs to be applied.
def descends_from_active_record?
if self == Base
false
elsif superclass.abstract_class?
superclass.descends_from_active_record?
else
superclass == Base || !columns_hash.include?(inheritance_column)
end
end
def finder_needs_type_condition? #:nodoc:
# This is like this because benchmarking justifies the strange :false stuff
:true == (@finder_needs_type_condition ||= descends_from_active_record? ? :false : :true)
end
# Returns the class descending directly from ActiveRecord::Base, or
# an abstract class, if any, in the inheritance hierarchy.
#
# If A extends ActiveRecord::Base, A.base_class will return A. If B descends from A
# through some arbitrarily deep hierarchy, B.base_class will return A.
#
# If B < A and C < B and if A is an abstract_class then both B.base_class
# and C.base_class would return B as the answer since A is an abstract_class.
def base_class
unless self < Base
raise ActiveRecordError, "#{name} doesn't belong in a hierarchy descending from ActiveRecord"
end
if superclass == Base || superclass.abstract_class?
self
else
superclass.base_class
end
end
# Set this to true if this is an abstract class (see <tt>abstract_class?</tt>).
# If you are using inheritance with ActiveRecord and don't want child classes
# to utilize the implied STI table name of the parent class, this will need to be true.
# For example, given the following:
#
# class SuperClass < ActiveRecord::Base
# self.abstract_class = true
# end
# class Child < SuperClass
# self.table_name = 'the_table_i_really_want'
# end
#
#
# <tt>self.abstract_class = true</tt> is required to make <tt>Child<.find,.create, or any Arel method></tt> use <tt>the_table_i_really_want</tt> instead of a table called <tt>super_classes</tt>
#
attr_accessor :abstract_class
# Returns whether this class is an abstract class or not.
def abstract_class?
defined?(@abstract_class) && @abstract_class == true
end
def sti_name
store_full_sti_class ? name : name.demodulize
end
Cache results of computing model type We faced a significant performance decrease when we started using STI without storing full namespaced class name in type column (because of PostgreSQL length limit for ENUM types). We realized that the cause of it is the slow STI model instantiation. Problematic method appears to be `ActiveRecord::Base.compute_type`, which is used to find the right class for STI model on every instantiation. It builds an array of candidate types and then iterates through it calling `safe_constantize` on every type until it finds appropriate constant. So if desired type isn't the first element in this array there will be at least one unsuccessful call to `safe_constantize`, which is very expensive, since it's defined in terms of `begin; rescue; end`. This commit is an attempt to speed up `compute_type` method simply by caching results of previous calls. ```ruby class MyCompany::MyApp::Business::Accounts::Base < ApplicationRecord self.table_name = 'accounts' self.store_full_sti_class = false end class MyCompany::MyApp::Business::Accounts::Free < Base end class MyCompany::MyApp::Business::Accounts::Standard < Base # patch .compute_type there end puts '======================= .compute_type =======================' Benchmark.ips do |x| x.report("original method") do MyCompany::MyApp::Business::Accounts::Free.send :compute_type, 'Free' end x.report("with types cached") do MyCompany::MyApp::Business::Accounts::Standard.send :compute_type, 'Standard' end x.compare! end ``` ``` ======================= .compute_type ======================= with types cached: 1529019.4 i/s original method: 2850.2 i/s - 536.46x slower ``` ```ruby 5_000.times do |i| MyCompany::MyApp::Business::Accounts::Standard.create!(name: "standard_#{i}") end 5_000.times do |i| MyCompany::MyApp::Business::Accounts::Free.create!(name: "free_#{i}") end puts '====================== .limit(100).to_a =======================' Benchmark.ips do |x| x.report("without .compute_type patch") do MyCompany::MyApp::Business::Accounts::Free.limit(100).to_a end x.report("with .compute_type patch") do MyCompany::MyApp::Business::Accounts::Standard.limit(100).to_a end x.compare! end ``` ``` ====================== .limit(100).to_a ======================= with .compute_type patch: 360.5 i/s without .compute_type patch: 24.7 i/s - 14.59x slower ```
2016-12-31 13:42:53 -05:00
def inherited(subclass)
subclass.instance_variable_set(:@_type_candidates_cache, Concurrent::Map.new)
super
end
protected
# Returns the class type of the record using the current module as a prefix. So descendants of
# MyApp::Business::Account would appear as MyApp::Business::AccountSubclass.
def compute_type(type_name)
Cache results of computing model type We faced a significant performance decrease when we started using STI without storing full namespaced class name in type column (because of PostgreSQL length limit for ENUM types). We realized that the cause of it is the slow STI model instantiation. Problematic method appears to be `ActiveRecord::Base.compute_type`, which is used to find the right class for STI model on every instantiation. It builds an array of candidate types and then iterates through it calling `safe_constantize` on every type until it finds appropriate constant. So if desired type isn't the first element in this array there will be at least one unsuccessful call to `safe_constantize`, which is very expensive, since it's defined in terms of `begin; rescue; end`. This commit is an attempt to speed up `compute_type` method simply by caching results of previous calls. ```ruby class MyCompany::MyApp::Business::Accounts::Base < ApplicationRecord self.table_name = 'accounts' self.store_full_sti_class = false end class MyCompany::MyApp::Business::Accounts::Free < Base end class MyCompany::MyApp::Business::Accounts::Standard < Base # patch .compute_type there end puts '======================= .compute_type =======================' Benchmark.ips do |x| x.report("original method") do MyCompany::MyApp::Business::Accounts::Free.send :compute_type, 'Free' end x.report("with types cached") do MyCompany::MyApp::Business::Accounts::Standard.send :compute_type, 'Standard' end x.compare! end ``` ``` ======================= .compute_type ======================= with types cached: 1529019.4 i/s original method: 2850.2 i/s - 536.46x slower ``` ```ruby 5_000.times do |i| MyCompany::MyApp::Business::Accounts::Standard.create!(name: "standard_#{i}") end 5_000.times do |i| MyCompany::MyApp::Business::Accounts::Free.create!(name: "free_#{i}") end puts '====================== .limit(100).to_a =======================' Benchmark.ips do |x| x.report("without .compute_type patch") do MyCompany::MyApp::Business::Accounts::Free.limit(100).to_a end x.report("with .compute_type patch") do MyCompany::MyApp::Business::Accounts::Standard.limit(100).to_a end x.compare! end ``` ``` ====================== .limit(100).to_a ======================= with .compute_type patch: 360.5 i/s without .compute_type patch: 24.7 i/s - 14.59x slower ```
2016-12-31 13:42:53 -05:00
if type_name.start_with?("::".freeze)
# If the type is prefixed with a scope operator then we assume that
# the type_name is an absolute reference.
ActiveSupport::Dependencies.constantize(type_name)
else
Cache results of computing model type We faced a significant performance decrease when we started using STI without storing full namespaced class name in type column (because of PostgreSQL length limit for ENUM types). We realized that the cause of it is the slow STI model instantiation. Problematic method appears to be `ActiveRecord::Base.compute_type`, which is used to find the right class for STI model on every instantiation. It builds an array of candidate types and then iterates through it calling `safe_constantize` on every type until it finds appropriate constant. So if desired type isn't the first element in this array there will be at least one unsuccessful call to `safe_constantize`, which is very expensive, since it's defined in terms of `begin; rescue; end`. This commit is an attempt to speed up `compute_type` method simply by caching results of previous calls. ```ruby class MyCompany::MyApp::Business::Accounts::Base < ApplicationRecord self.table_name = 'accounts' self.store_full_sti_class = false end class MyCompany::MyApp::Business::Accounts::Free < Base end class MyCompany::MyApp::Business::Accounts::Standard < Base # patch .compute_type there end puts '======================= .compute_type =======================' Benchmark.ips do |x| x.report("original method") do MyCompany::MyApp::Business::Accounts::Free.send :compute_type, 'Free' end x.report("with types cached") do MyCompany::MyApp::Business::Accounts::Standard.send :compute_type, 'Standard' end x.compare! end ``` ``` ======================= .compute_type ======================= with types cached: 1529019.4 i/s original method: 2850.2 i/s - 536.46x slower ``` ```ruby 5_000.times do |i| MyCompany::MyApp::Business::Accounts::Standard.create!(name: "standard_#{i}") end 5_000.times do |i| MyCompany::MyApp::Business::Accounts::Free.create!(name: "free_#{i}") end puts '====================== .limit(100).to_a =======================' Benchmark.ips do |x| x.report("without .compute_type patch") do MyCompany::MyApp::Business::Accounts::Free.limit(100).to_a end x.report("with .compute_type patch") do MyCompany::MyApp::Business::Accounts::Standard.limit(100).to_a end x.compare! end ``` ``` ====================== .limit(100).to_a ======================= with .compute_type patch: 360.5 i/s without .compute_type patch: 24.7 i/s - 14.59x slower ```
2016-12-31 13:42:53 -05:00
type_candidate = @_type_candidates_cache[type_name]
if type_candidate && type_constant = ActiveSupport::Dependencies.safe_constantize(type_candidate)
return type_constant
end
# Build a list of candidates to search for
candidates = []
name.scan(/::|$/) { candidates.unshift "#{$`}::#{type_name}" }
candidates << type_name
candidates.each do |candidate|
constant = ActiveSupport::Dependencies.safe_constantize(candidate)
Cache results of computing model type We faced a significant performance decrease when we started using STI without storing full namespaced class name in type column (because of PostgreSQL length limit for ENUM types). We realized that the cause of it is the slow STI model instantiation. Problematic method appears to be `ActiveRecord::Base.compute_type`, which is used to find the right class for STI model on every instantiation. It builds an array of candidate types and then iterates through it calling `safe_constantize` on every type until it finds appropriate constant. So if desired type isn't the first element in this array there will be at least one unsuccessful call to `safe_constantize`, which is very expensive, since it's defined in terms of `begin; rescue; end`. This commit is an attempt to speed up `compute_type` method simply by caching results of previous calls. ```ruby class MyCompany::MyApp::Business::Accounts::Base < ApplicationRecord self.table_name = 'accounts' self.store_full_sti_class = false end class MyCompany::MyApp::Business::Accounts::Free < Base end class MyCompany::MyApp::Business::Accounts::Standard < Base # patch .compute_type there end puts '======================= .compute_type =======================' Benchmark.ips do |x| x.report("original method") do MyCompany::MyApp::Business::Accounts::Free.send :compute_type, 'Free' end x.report("with types cached") do MyCompany::MyApp::Business::Accounts::Standard.send :compute_type, 'Standard' end x.compare! end ``` ``` ======================= .compute_type ======================= with types cached: 1529019.4 i/s original method: 2850.2 i/s - 536.46x slower ``` ```ruby 5_000.times do |i| MyCompany::MyApp::Business::Accounts::Standard.create!(name: "standard_#{i}") end 5_000.times do |i| MyCompany::MyApp::Business::Accounts::Free.create!(name: "free_#{i}") end puts '====================== .limit(100).to_a =======================' Benchmark.ips do |x| x.report("without .compute_type patch") do MyCompany::MyApp::Business::Accounts::Free.limit(100).to_a end x.report("with .compute_type patch") do MyCompany::MyApp::Business::Accounts::Standard.limit(100).to_a end x.compare! end ``` ``` ====================== .limit(100).to_a ======================= with .compute_type patch: 360.5 i/s without .compute_type patch: 24.7 i/s - 14.59x slower ```
2016-12-31 13:42:53 -05:00
if candidate == constant.to_s
@_type_candidates_cache[type_name] = candidate
return constant
end
end
raise NameError.new("uninitialized constant #{candidates.first}", candidates.first)
end
end
private
# Called by +instantiate+ to decide which class to use for a new
# record instance. For single-table inheritance, we check the record
# for a +type+ column and return the corresponding class.
def discriminate_class_for_record(record)
if using_single_table_inheritance?(record)
find_sti_class(record[inheritance_column])
else
super
end
end
def using_single_table_inheritance?(record)
record[inheritance_column].present? && has_attribute?(inheritance_column)
end
def find_sti_class(type_name)
type_name = base_class.type_for_attribute(inheritance_column).cast(type_name)
subclass = begin
if store_full_sti_class
ActiveSupport::Dependencies.constantize(type_name)
else
compute_type(type_name)
end
rescue NameError
raise SubclassNotFound,
"The single-table inheritance mechanism failed to locate the subclass: '#{type_name}'. " \
"This error is raised because the column '#{inheritance_column}' is reserved for storing the class in case of inheritance. " \
"Please rename this column if you didn't intend it to be used for storing the inheritance class " \
"or overwrite #{name}.inheritance_column to use another column for that information."
end
unless subclass == self || descendants.include?(subclass)
raise SubclassNotFound, "Invalid single-table inheritance type: #{subclass.name} is not a subclass of #{name}"
end
subclass
end
def type_condition(table = arel_table)
sti_column = arel_attribute(inheritance_column, table)
sti_names = ([self] + descendants).map(&:sti_name)
sti_column.in(sti_names)
end
# Detect the subclass from the inheritance column of attrs. If the inheritance column value
# is not self or a valid subclass, raises ActiveRecord::SubclassNotFound
def subclass_from_attributes(attrs)
attrs = attrs.to_h if attrs.respond_to?(:permitted?)
if attrs.is_a?(Hash)
subclass_name = attrs[inheritance_column] || attrs[inheritance_column.to_sym]
if subclass_name.present?
find_sti_class(subclass_name)
end
end
end
end
def initialize_dup(other)
super
ensure_proper_type
end
private
def initialize_internals_callback
super
ensure_proper_type
end
# Sets the attribute used for single table inheritance to this class name if this is not the
# ActiveRecord::Base descendant.
# Considering the hierarchy Reply < Message < ActiveRecord::Base, this makes it possible to
# do Reply.new without having to set <tt>Reply[Reply.inheritance_column] = "Reply"</tt> yourself.
# No such attribute would be set for objects of the Message class in that example.
def ensure_proper_type
klass = self.class
if klass.finder_needs_type_condition?
2017-06-18 10:37:06 -04:00
_write_attribute(klass.inheritance_column, klass.sti_name)
end
end
end
end