Opal: Ruby in the Browser and the Game of Life

Welcome to the the second part of the article where we look at Opal, a Ruby to Javascript compiler! In the previous installment, I introduced Opal and showed how to get it set up on your system. We created the first half of Conway’s Game of Life in Opal. In particular, implementing the grid using the HTML 5 canvas element, and adding some interactivity to it. In this article, we will complete our application by implementing the rest of the logic, and hooking it to the canvas. The completed source can be found at the end of the article. Let’s dive right in!

Game Rules

As you might recall, Conway’s Game of Life consists of 4 simple rules:

Rule 1

Any live cell with fewer than two live neighbors dies, as if caused by under-population.

Rule 2

Any live cell with two or three live neighbors lives on to the next generation.

Rule 3

Any live cell with more than three live neighbors dies, as if by overcrowding.

Rule 4

Any dead cell with exactly three live neighbors becomes a live cell, as if by reproduction.

Protip: Using pp as console.log

Along the way, you might use lots of console.log statements for debugging. Indeed, that was what I used to do. Then, I learned an awesome trick:

@bentanweihao @opalrb #protip `console.log` is mapped to #pp (https://t.co/5mrxIiOJdV), awesome article btw :) — EliaSchito(͡° ͜ʖ ͡°) (@elia) April 4, 2014

Implementing the Game Logic

Step 1: Create conway.rb

In the app directory, go ahead and create conway.rb:

require 'opal'
require 'opal-jquery'
require 'grid'

class Conway
attr_reader :grid

def initialize(grid)
@grid = grid
end
end

We pass in an instance of Grid to the Conway initializer.

Step 2: Discovering Liveness and Population Count

Before the 4 rules can be implemented, we need to find out whether a cell at a particular coordinate is alive or dead:

def is_alive?(x, y)
state[[x, y]] == 1
end

def is_dead?(x, y)
!is_alive?(x, y)
end

Also how many neighbors for a given coordinate:

def population_at(x, y)
[
state[[x-1, y-1]],
state[[x-1, y ]],
state[[x-1, y+1]],
state[[x, y-1]],
state[[x, y+1]],
state[[x+1, y-1]],
state[[x+1, y ]],
state[[x+1, y+1]]
].map(&:to_i).reduce(:+)
end

Notice how we can happily make use of Ruby methods like map and reduce, and even Symbol#to_proc works! It almost makes you forget that you are also working with Javascript.

Step 3: Implementing the Rules

Now that we can check for the liveness and population count at a coordinate, implementing the 4 rules is simple:

# Any live cell with fewer than two live neighbours dies,
# as if caused by under-population.
def is_underpopulated?(state, x, y)
is_alive?(x, y) && population_at(x, y) < 2
end

# Any live cell with two or three live neighbours lives
# on to the next generation.
def is_living_happily?(x, y)
is_alive?(x, y) && ([2, 3].include? population_at(x, y))
end

# Any live cell with more than three live neighbours dies,
# as if by overcrowding.
def is_overpopulated?(x, y)
is_alive?(x, y) && population_at(x, y) > 3
end

# Any dead cell with exactly three live neighbours becomes a live cell,
# as if by reproduction.
def can_reproduce?(x, y)
is_dead?(x, y) && population_at(x, y) == 3
end

Next, check if a cell makes it at the next interval, or tick, in Game of Life parlance:

def get_state_at(x, y)
if is_underpopulated?(x, y)
0
elsif is_living_happily?(x, y)
1
elsif is_overpopulated?(x, y)
0
elsif can_reproduce?(x, y)
1
end
end

Step 4: Accessing the Grid State

Recall how Grid was implemented:

class Grid
attr_reader :height, :width, :canvas, :context, :max_x, :max_y
attr_accessor :state

CELL_HEIGHT = 15;
CELL_WIDTH = 15;

def initialize
@height = `$(window).height()`
@width = `$(window).width()`
@canvas = `document.getElementById(#{canvas_id})`
@context = `#{canvas}.getContext('2d')`
@max_x = (height / CELL_HEIGHT).floor
@max_y = (width / CELL_WIDTH).floor
@state = blank_state
draw_canvas
end

def blank_state
h = Hash.new
(0..max_x).each do |x|
(0..max_y).each do |y|
h[[x,y]] = 0
end
end
h
end
# ...
end

Note that the state of the grid is stored in the @state attribute. In order to access the grid state, a Conway object needs to do something like grid.state. Turns out, there’s a more idiomatic way to do this in Ruby. Enter the Forwardable module. What does this do exactly?

The Forwardable module provides delegation of specified methods to a designated object, using the methods def_delegator and def_delegators.

In the code:

require 'opal'
require 'opal-jquery'
require 'grid'
require 'forwardable'

class Conway
attr_reader :grid
extend Forwardable

def_delegators :@grid, :state, :state=

def initialize(grid)
@grid = grid
end
# ...
end

First, require and extend Forwardable. Then, declare which object (@grid) and which methods we want delegated. This means that both of these are now the same thing:

Conway.new(Grid.new).grid.state

Conway.new(Grid.new).state

Step 5: Computing the Next State

At each tick, the next state for each cell needs to be computed. In other words, find out whether that cell is alive or dead. The computed new state is then saved into grid. The entire canvas is redrawn with these new coordinates.

def tick
# def_delegators at work again!
# This call is delegate to grid.state=
self.state = new_state
redraw_canvas
end

def new_state
new_state = Hash.new
state.each do |cell, _|
new_state[cell] = get_state_at(cell[0], cell[1])
end
new_state
end

Redrawing the Canvas

We have yet to implement redraw_canvas. Let’s rectify that. Open up grid.rb and add this:

class Grid
# ...
def redraw_canvas
draw_canvas
state.each do |cell, liveness|
fill_cell(cell[0], cell[1]) if liveness == 1
end
end
# ...
end

redraw_canvas redraws a blank canvas and iterates through state. If the coordinate value is 1 (alive), it fills it, otherwise it is left unfilled. Head back to conway.rb and make an addition to the def_delegators:

def_delegators :@grid, :state, :state=, :redraw_canvas

Step 6. Looping

There is no way to transition into the next tick. In Opal, you cannot write an infinite loop in the traditional Ruby way. So, doing something like this does not work:

loop do
tick
end

Neither will this:

while true do
tick
end

Let’s think for a bit. What we really want is to run tick at certain, fixed intervals. Javascript has setInterval! How can we use this in our code? Create yet another file called interval.rb in app.

class Interval
# Note that time is in ms.
def initialize(time=0, &block)
@interval = `setInterval(function(){#{block.call}}, time)`
end

def stop
`clearInterval(#@interval)`
end
end

I hope your mind is blown after you realize what is going on here. The Interval initializer takes in a time (no surprise), and a block! Then it calls the block in the Javascript setInterval function! This means that we can now implement our infinite loop. Remember to require interval and then pass in a block containing the tick method like so:

require 'interval'
#...

class Conway
# ...
def run
Interval.new do
tick
end
end
# ...
end

run kickstarts the loop which drives the entire animation. But before we get to that, we have a bit more work left to do.

Step 7: Seeding

We have yet to tackle the issue of the initial state of the board, also known as the seed. seed should contain the coordinates of all the clicked positions of the board. When the grid is first loaded, it is blank. The user gets to click the cells to fill them up. This makes up the seed. The seed coordinates are recorded. When the user is done, he/she presses the Enter key, triggering Conway#run. The user now gets to enjoy a lovely animation. Open up grid.rb, and add the seed attribute. For convenience, we call add_mouse_event_listener when Grid is initialized.

class Grid
attr_accessor :state, :seed

def initialize
# ...
@seed = []
# ...
add_mouse_event_listener
end
# ...
end

When a click occurs, add the coordinates into seed. Similarly, for double-click, remove the coordinates from seed.

def add_mouse_event_listener
Element.find("##{canvas_id}").on :click do |event|
# ...
seed << [x, y]
end

Element.find("##{canvas_id}").on :dblclick do |event|
# ...
seed.delete([x, y])
end
end

Turn your attention now to conway.rb. First, add seed to def_delegators:

def_delegators :@grid, :state, :state=, :redraw_canvas, :seed

We need to detect when the ‘Enter’ key is pressed.

def add_enter_event_listener
Document.on :keypress do |event|
if enter_pressed?(event)
seed.each do |x, y|
state[[x, y]] = 1
end

run
end
end
end

When ‘Enter’ is detected, this method iterates through the coordinates of seed and populate the grid‘s initial state. After that, the game is run. How do we detect if an ‘Enter’ is pressed? Just like in JavaScript/jQuery, check event.which for the ‘Enter’ key code:

def enter_pressed?(event)
event.which == 13
end

Again for convenience, we run add_enter_event_listener when Conway is initialized:

class Conway
# ...
def initialize(grid)
@grid = grid
add_enter_event_listener
end
# ...
end

Step 8: Bonus! The Glider Gun Pattern

Drop this code into grid.rb:

class Grid

def initialize
# ...
add_demo_event_listener
end

# Ctrl+D displays a demo of the Glider Gun:
# https://en.wikipedia.org/wiki/File:Game_of_life_glider_gun.svg
def add_demo_event_listener
Document.on :keypress do |event|
if ctrl_d_pressed?(event)
[
[25, 1],
[23, 2], [25, 2],
[13, 3], [14, 3], [21, 3], [22, 3],
[12, 4], [16, 4], [21, 4], [22, 4], [35, 4], [36, 4],
[1, 5], [2, 5], [11, 5], [17, 5], [21, 5], [22, 5], [35, 5], [36, 5],
[1, 6], [2, 6], [11, 6], [15, 6], [17, 6], [18, 6], [23, 6], [25, 6],
[11, 7], [17, 7], [25, 7],
[12, 8], [16, 8],
[13, 9], [14, 9]
].each do |x, y|
fill_cell(x, y)
seed << [x, y]
end
end
end
end

def ctrl_d_pressed?(event)
event.ctrl_key == true && event.which == 4
end
# ...
end

Here, when you type Ctrl + D together, the Glider Gun pattern would be drawn. Press ‘Enter’ to start the animation.

Step 9: All Done!

Give yourself a big pat on the back. We have reached the end. Go ahead and try out a couple of patterns. If you need any inspiration for starting patterns, a quick search will yield many interesting results.

Concluding Remarks

I hope you had a bit of fun experimenting with Opal and, along the way, maybe even found ways to integrate Opal into your existing project. Rails developers will be pleased to know that Opal plays nice with the asset pipeline with the opal-rails gem. In my short experience playing with Opal, I didn’t find debugging that big of an issue. Usually, inspecting the error log in the browser was sufficient in figuring out what went wrong. This is especially true since during compilation, the Javascript function has the same name as the Ruby method. Most of the times the error messages were descriptive enough. You will notice that this implementation is not exactly performant, and depends on the size of the canvas. I chose to go for the most straightforward implementation, therefore, some performance had to be sacrificed. Nonetheless, I hope you had lot of fun building this application! Another thing to note is that there are a few libraries already written for Opal, such as opal-browser, opal-rails and opal-rspec. You can see even more on the Opal github page. We could make our application shorter with opal-browser, for example, but I wanted to demonstrate what could be achieved with the bare minimum. Opal is very promising, and I hope it gains more widespread adoption. It will be very interesting to see what other developers do with this. Thanks for reading!

Resources

• Opal, A new hope (for Ruby programmers) (Ruby Conf 2013) – video
• Opal mailing list is pretty active, and a good place to ask questions if you get stuck.

The Complete Source

For your convenience, the complete source is on Github

Frequently Asked Questions (FAQs) about Opal, Ruby, and Browser Game Life