When it comes to run time introspection and dynamic code generation I don't think ruby has any rivals except possibly for some lisp dialects. The other day I was doing some code exercise to explore ruby's dynamic facilities and I started to wonder about ways of adding methods to existing objects. Here are 3 ways I could think of:
obj = Object.new
# add a method directly
def obj.new_method
...
end
# add a method indirectly with the singleton class
class << obj
def new_method
...
end
end
# add a method by opening up the class
obj.class.class_eval do
def new_method
...
end
end
This is just the tip of the iceberg because I still haven't explored various combinations of instance_eval, module_eval and define_method. Is there an online/offline resource where I can find out more about such dynamic tricks?
Ruby Metaprogramming seems to be a good resource. (And, linked from there, The Book of Ruby.)
If obj has a superclass, you can add methods to obj from the superclass using define_method (API) as you mentioned. If you ever look at the Rails source code, you'll notice that they do this quite a bit.
Also while this isn't exactly what you're asking for, you can easily give the impression of creating an almost infinite number of methods dynamically by using method_missing:
def method_missing(name, *args)
string_name = name.to_s
return super unless string_name =~ /^expected_\w+/
# otherwise do something as if you have a method called expected_name
end
Adding that to your class will allow it to respond to any method call which looks like
#instance.expected_something
I like the book Metaprogramming Ruby which is published by the publishers of the pickaxe book.
Related
I don't know how to correctly phrase the title, I think the best way to explain this issue is just with code samples.
My goal
I want to define a meta method like this (in Rails 5):
class Post < ApplicationRecord
override_this_attribute_writer :some_attribute
end
The override_this_attribute_writer follows a common pattern, it overrides the original writer by doing some filtering on top of it. I find this way of overriding very convenient and clear.
First approach
module MyCommonModule
extend ActiveSupport::Concern
module ClassMethods
def override_this_attribute_writer(attribute_name)
alias_method :"#{attribute_name}_old=", :"#{attribute_name}="
define_method :"#{attribute_name}=" do |a_value|
# Do my stuff
send(:"#{attribute_name}_old=", a_value)
end
end
end
When doing this, I was getting an exception at the call of alias_method, because, apparently, the method I was trying to copy didn't exist (yet).
Second approach
module MyCommonModule
extend ActiveSupport::Concern
module ClassMethods
def override_this_attribute_writer(attribute_name)
define_method :"#{attribute_name}=" do |a_value|
# Do my stuff
send(:write_attribute, attribute_name, a_value)
end
end
end
I was expecting this not to work: if, when running the meta method, ActiveRecord hasn't created the attribute writer yet, this means that it will do it later and override the method that I just defined.
But surprisingly it worked! So I put my hands inside ActiveRecord (5.1.5) to find out more.
Dig into ActiveRecord 5.1.5
I wanted to ensure that what I did was safe and it wasn't just working by accident: I looked into the definition of method writer, and put binding.pry around the method.
This is the result of the experiment:
For attributes that I did not override,
This line is called
Then the method is defined inside this module eval call
Finally, the newly created writer method is correctly called when performing object.attribute=
For attributes that I DID override,
My own method is defined before anything else (when the ActiveRecord writers aren't there yet
Then ActiveRecord calls the same line that handles writer creation, as in the previous example
The method gets (apparently) correctly created by ActiveRecord, since it passes again by this point
But now, surprisingly, when calling object.attribute= my own method is still called in place of the ActiveRecord one
So, this is what I don't understand: if ActiveRecord seems to be overriding my method but it doesn't, what prevents it from doing it?
My questions
What in the end I need to know is whether the fix I have done is actually a good practice (and robust) or it's at risk and it might break if in the future we do upgrades.
If you think that my fix is dangerous, would you be able to suggest a different way to achieve the same goal?
Calling super is even more idiomatic:
module MyCommonModule
extend ActiveSupport::Concern
module ClassMethods
def override_this_attribute_writer(attribute_name)
define_method :"#{attribute_name}=" do |value|
# do some stuff
super value
end
end
end
end
I am new to Ruby and I saw methods defined like:
def method_one
puts "method 1"
end
class MyClass
method_one
def method_two
puts "method 2"
end
end
The way method_one is used reminds me of Python decorators.The output of
c = MyClass.new
c.method_two
is
method 1
method 2
I have been trying to search for more information about this syntax/language feature in the Ruby documentation on the web but I don't know what keywords to search for.
What this is thing called?
TL;DR
This code doesn't do what you think it does. Don't do stuff like this.
Ruby's Top-Level Object
Ruby lets you define methods outside a class. These methods exist on a top-level object, which you can (generally) treat as a sort of catch-all namespace. You can see various posts like What is the Ruby Top-Level? for more details, but you shouldn't really need to care.
In your original post, method_one is just a method defined in the top-level. It is therefore available to classes and methods nested within the top-level, such as MyClass.
Methods in Classes
Despite what you think, the following doesn't actually declare a :method_one class or instance method on MyClass:
class MyClass
method_one
def method_two; end
end
Instead, Ruby calls the top-level ::method_one during the definition of the class, but it never becomes a class method (e.g. MyClass::method_one) or an instance method (e.g. MyClass.new.method_one). There might be a few use cases for doing this (e.g. printing debugging information, test injection, etc.) but it's confusing, error-prone, and generally to be avoided unless you have a really strong use case for it.
Better Options
In general, when you see something like this outside an academic lesson, the programmer probably meant to do one of the following:
Extend a class.
Add a singleton method to a class.
Include a module in a class.
Set up a closure during class definition.
The last gets into murky areas of metaprogramming, at which point you should probably be looking at updating your class initializer, or passing Proc or lambda objects around instead. Ruby lets you do all sorts of weird and wonderful things, but that doesn't mean you should.
I think you're a little mislead; the output of:
c = MyClass.new
c.method_two
is
#<MyClass:0x007feda41acf18>
"method 2"
You're not going to see method one until the class is loaded or if you're in IRB you enter the last end statement.
I would suggest looking into ruby's initialize method.
In the code below, method roar is not defined in class Lion, but still can be called using method_missing.
class Lion
def method_missing(name, *args)
puts "Lion will #{name}: #{args[0]}"
end
end
lion = Lion.new
lion.roar("ROAR!!!") # => Lion will roar: ROAR!!!
In which situations and how should I use this method_missing? And is it safe to use?
It's entirely safe to use provided you use it in expected ways and don't get carried away. Not everything you can do is worth doing, after all.
The advantage of method_missing is you can respond to all kinds of things in unique ways.
The disadvantage is you don't advertise your capabilities. Other objects that expect you to respond_to? something will not get confirmation and might treat your custom object in ways you don't intend.
For building Domain Specific Languages and providing very loose glue between components, this sort of thing is invaluable.
A great example of where this is a good fit is the Ruby OpenStruct class.
Summary: When to use? When it will make your life easier and not complicate others' lives.
Here's one example that pops to mind. It's from redis_failover gem.
# Dispatches redis operations to master/slaves.
def method_missing(method, *args, &block)
if redis_operation?(method)
dispatch(method, *args, &block)
else
super
end
end
Here we check if the method called is actually a command of redis connection. If so, we delegate it to underlying connection(s). If not, relay to super.
Another famous example of method_missing application is ActiveRecord finders.
User.find_by_email_and_age('me#example.com', 20)
There's not, of course, a method find_by_email_and_age. Instead, the method_missing breaks the name, analyzes the parts and invokes find with proper parameters.
Here's a favorite of mine
class Hash
def method_missing(sym,*args)
fetch(sym){fetch(sym.to_s){super}}
end
end
Which lets me access values of a hash as if they were attributes. This is particular handy when working with JSON data.
So for example, rather than having to write tweets.collect{|each|each['text']} I can just write tweets.collect(&:text) which is much shorter. Or also, rather than tweets.first['author'] I can just write tweets.first.author which feels much more natural. Actually, it gives you Javascript-style access to values of a hash.
NB, I'm expecting the monkey patching police at my door any minute…
First and foremost, stick to Sergio Tulentsev's summary.
Apart from that, I think looking at examples is the best way to get a feeling for right and wrong situations for method_missing; so here is another simple example:
I recently made use of method_missing in a Null Object.
The Null Object was a replacement for a Order model.
The Order stores different prices for different currencies.
Without method_missing it looks like this:
class NullOrder
def price_euro
0.0
end
def price_usd
0.0
end
# ...
# repeat for all other currencies
end
With method_missing, I can shorten it to:
class NullOrder
def method_missing(m, *args, &block)
m.to_s =~ /price_/ ? 0.0 : super
end
end
With the added benefit of not having to (remember to) update the NullOrder when I add new price_xxx attributes to Order.
I also found a blog post from (Paolo Perrotta) where it demonstrated when to use method_missing:
class InformationDesk
def emergency
# Call emergency...
"emergency() called"
end
def flights
# Provide flight information...
"flights() called"
end
# ...even more methods
end
Check if a service has been asked during lunch time.
class DoNotDisturb
def initialize
#desk = InformationDesk.new
end
def method_missing(name, *args)
unless name.to_s == "emergency"
hour = Time.now.hour
raise "Out for lunch" if hour >= 12 && hour < 14
end
#desk.send(name, *args)
end
end
# At 12:30...
DoNotDisturb.new.emergency # => "emergency() called"
DoNotDisturb.new.flights # ~> -:37:in `method_missing': Out for lunch (RuntimeError)
Let's say I have a bunch of related functions that have no persistent state, say various operations in a string differencing package. I can either define them in a class or module (using self) and they can be accessed the exact same way:
class Diff
def self.diff ...
def self.patch ...
end
or
module Diff
def self.diff ...
def self.patch ...
end
I can then do Diff.patch(...). Which is 'better' (or 'correct')?
The main reason I need to group them up is namespace issues, common function names are all used elsewhere.
Edit: Changed example from matrix to diff. Matrix is a terrible example as it does have state and everyone started explaining why it's better to write them as methods rather than answer the actual question. :(
In your two examples, you are not actually defining methods in a Class or a Module; you are defining singleton methods on an object which happens to be a Class or a Module, but could be just about any object. Here's an example with a String:
Diff = "Use me to access really cool methods"
def Diff.patch
# ...
end
You can do any of these and that will work, but the best way to group related methods is in a Module as normal instance methods (i.e. without self.):
module Diff
extend self # This makes the instance methods available to the Diff module itself
def diff ... # no self.
def patch ...
end
Now you can:
use this functionality from within any Class (with include Diff) or from any object (with extend Diff)
an example of this use is the extend self line which makes it possible to call Diff.patch.
even use these methods in the global namespace
For example, in irb:
class Foo
include Diff
end
Foo.new.patch # => calls the patch method
Diff.patch # => also calls Diff.patch
include Diff # => now you can call methods directly:
patch # => also calls the patch method
Note: the extend self will "modify" the Diff module object itself but it won't have any effect on inclusions of the module. Same thing happens for a def self.foo, the foo won't be available to any class including it. In short, only instance methods of Diff are imported with an include (or an extend), not the singleton methods. Only subclassing a class will provide inheritance of both instance and singleton methods.
When you actually want the inclusion of a module to provide both instance methods and singleton methods, it's not completely easy. You have to use the self.included hook:
module Foo
def some_instance_method; end
module ClassMethods
def some_singleton_method; end
end
def self.included(base)
base.send :extend, ClassMethods
end
def self.will_not_be_included_in_any_way; end
end
class Bar
include Foo
end
# Bar has now instance methods:
Bar.new.some_instance_method # => nil
# and singleton methods:
Bar.some_singleton_method # => nil
The main difference between modules and classes is that you can not instantiate a module; you can't do obj = MyModule.new. The assumption of your question is that you don't want to instantiate anything, so I recommend just using a module.
Still you should reconsider your approach: rather than using arrays of arrays or whatever you are doing to represent a Matrix, it would be more elegant to make your own class to represent a matrix, or find a good class that someone else has already written.
Ruby Modules are used to specify behaviour, pieces of related functionality.
Ruby Classes are used to specify both state and behaviour, a singular entity.
There is a maxim in software design that says that code is a liability, so use the less code possible. In the case of Ruby, the difference in code lines is cero. So you can use either way (if you don't need to save state)
If you want to be a purist, then use a Module, since you won't be using the State functionality. But I wouldn't say that using a class is wrong.
As a trivia info: In Ruby a Class is a kind of Module.
http://www.ruby-doc.org/core-1.9.3/Class.html
The following also works
Matrix = Object.new
def Matrix.add ...
def Matrix.equals ...
That's because so-called "class methods" are just methods added to a single object, and it doesn't really matter what that object class is.
As a matter of form, the Module is more correct. You can still create instances of the class, even if it has only class methods. You can think of a module here as a static class of C# or Java. Classes also always have the instance related methods (new, allocate, etc.). Use the Module. Class methods usually have something to do with objects (creating them, manipulating them).
What I'm trying to find out is whether there is some sort of equivalence to what I see in Groovy as ExpandoMetaClasses. I've been reading about Open Classes but I can't quite see what level of scoping Ruby allows of the class modifications.
Borrowing an example from the blog above, in Groovy, I could modify Java's String class and add a method to it like so:
String.metaClass.shout = {->
return delegate.toUpperCase()
}
println "Hello MetaProgramming".shout()
// output
// HELLO METAPROGRAMMING
And I think that Ruby would have you redefine the class and possibly alias it (please help clarify my misunderstandings at this point):
class String
def foo
"foo"
end
end
puts "".foo # prints "foo"
In Groovy, there are ways to scope the redefinition of core Java library methods to single instances or to a group of instances using Categories, which feel similar to what I would define as mixins in Ruby.
What are the ways to scope open classes to specific instances or to subsets of modules?
If I were to install a gem that had redefined some core class, would only that module be affected, or would any .rb file I include that gem with be affected with it?
Apologies in advance for making some possible assumptions on both Ruby and Groovy, I'm new to both but have been trying to find equivalence between the two.
Ruby's classes are never "closed". So when you say:
class String
def omg!
self.replace "OMG"
end
end
You are defining the omg! method on the String class. Unlike in Groovy, which requires the usage of a special metaclass concept, Ruby classes are always open, period.
If you wanted to modify a particular set of Strings, you could do this:
module Magic
def presto
puts "OMG A HAT!"
end
end
class Array
include Magic
end
x = "Hello".extend(Magic)
puts x #=> Hello
x.presto #=> OMG A HAT!
[].presto #=> OMG A HAT!
def x.really?
true
end
x.really? #=> true
Effectively, a module is a collection of methods that can be added to a class or specific instances.
So you can either open a class directly or add new methods to a class using a module. You can also open an instance directly or add new methods to an instance using a module. That's because a class is just an instance of Class ;) Pretty nifty!
In addition to what Yehuda said, instances in Ruby also have metaclasses (technically called "singleton classes"), accessed with class <<whatever. For example, to redo Yehuda's Magic example with a singleton class:
x = "Hello"
class <<x
include Magic
def magical?
true
end
end
x.presto #=> OMG A HAT!
x.magical? #=> true
"Something else".magical? #=> NoMethodError
There's no scoping on modifications to classes. As soon as a class is modified, the modified class is accessible to all later requires and following code.