Gem Version Dependency Status CircleCI Code Climate Test Coverage PullReview stats

PageMagic

PageMagic is an API for testing web applications.

It has a simple but powerful DSL which makes modelling and interacting with your pages easy.

Wouldn't it be great if there was a framework that could:

Well PageMagic might just be the answer!

Give it a try and let us know what you think! There will undoubtedly be things that can be improved and issues that we are not aware of so your feedback/pull requests are greatly appreciated!

Contents

Installation

gem install page_magic

Quick Start

Getting started with PageMagic is easy, try running this:

require 'page_magic'

class Github
  include PageMagic

  url 'https://www.github.com'

  text_field :search_field, name: 'q' do
    watch(:url)

    after_events do
      wait_until { changed?(:url) }
    end
  end

  def search(project_name)
    search_field.set "#{project_name}\n"
  end
end
github = Github.visit(browser: :chrome)
github.search('page_magic')

This example defines a page to represent Github's home page, visits it and performs a search.

This code models a single page and will let you interact with the elements defined on it as well as use the helper method we defined.

You can do lots with PageMagic including mapping pages to a session so that they are fluidly switched in for you. You can even define hooks to run when ever a element is interacted with. So what are you wating for? there's no place better to start than the beginning. Have fun! :)

Defining Pages

To define something that PageMagic can work with, simply include PageMagic in to a class.

class Github
  include PageMagic
end

Elements

Defining elements is easy. Just give the:

  • element_type
  • id to refer to it by
  • selector to find it in the page

The following example defines a text field called 'search_field' that can be found using its name which is 'q'

class Github
  include PageMagic

  text_field :search_field, name: 'q'
end

Element types supported by PageMagic are:

  • text_field
  • button
  • link
  • checkbox
  • select_list
  • radio
  • textarea

PageMagic is very powerful and provides a number of different ways to define a page element so for more details please look at the API

Singular Results

The element types written above are the method names you need to use in order to define elements of those types on your pages.

Using them as they are written above will tell PageMagic to expect to find only one element using the selector you supply. Finding more than one result will cause PageMagic to raise an AmbiguousQueryException

Multiple Results

Applying an 's' to the end of the element definition method name will tell PageMagic that more than one result can be returned using the given selector.

results are returned in an Array.

class ResultsPage
  include PageMagic
  elements :results, css: '.repo-list-item'
end

page.results #=> Array<Element>

Interacting with elements

class Github
  include PageMagic

  text_field :search_field, name: 'q'
end

Elements are defined with an id which is the name of the method you will use to reference it. In the above example, the text field was defined with the id :search_field.

After visiting a page you are will get a Session object. Elements can be accessed through the session itself.

page.search_field.set 'page_magic'

Sub Elements

If your pages are complex you can use PageMagic to compose pages, their elements and subelements to as many levels as you need to.

class ResultsPage
  include PageMagic

  element :results, css: '.repo-list-item' do
    element :stats, css: '.repo-list-stats'
    element :meta_data, css: '.repo-list-meta'
    link :repo_link, css: 'h3 a'
  end
end

Sub elements can be accessed through their parent elements e.g:

page.results.first.repo_link.click

Custom elements

PageMagic allows you to define your own custom elements.

class SearchField < PageMagic::Element
  selector name: 'q'
  # custom stuff
end

class Github 
  include PageMagic
  element SearchField
end

If an id is not specified then the name of the element class will be used. In the above example the name given to the element of type SearchField would be search_field. The selector for the element can bespecified on the class itself or overiden when defining the element. The custom element can also be extended as with other elements.

class MyPage
  include PageMagic
  element SearchField, :search, selector: '.custom' do
    link(:extr_link, id: 'extra-link')
  do
end

Hooks

PageMagic provides hooks to allow you to define actions that are executed when you pages and elements are interacted with.

Note:

  • You may well find PageMagic's watchers useful.
  • The following examples wait for actions to happen. You can of course write you own wait code or feel free try out our wait_until helper:)

On load hook

PageMagic lets you define an on_load hook for your pages. This will be executed when the browser thinks the page has been loaded.

class Github
  # ... code defining elements as shown above

  on_load do
    # code that needs to run when the page has loaded
  end
end

Element event hooks

Frequently, you are going to have to work with pages that make heavy use of ajax. For these occasions PageMagic provides before_events and after_events hooks that you use to perform custom action.

In the following example we have added watchers and event hooks to the SearchField custom element we defined in the previous section. Encapsulating the business logic here means that we can really add value to the pages that reuse this custom element

class SearchField < PageMagic::Element
  selector name: 'q'
  watch(:url)
  after_events do
    wait_until { changed?(:url) }
  end
end

Helper methods

Helper methods can be defined to avoid writing repetive page/element specific code outside of your pages and elements.

class Github
  # ... code defining elements as shown above

  def search(project_name)
    search_field.set "#{project_name}\n"
  end
end

We can interact with helper in the same way as we did page elements.

page.search('page_magic')

Dynamic Selectors

In some cases you wont be able to specify the selector for an element until runtime. PageMagic allows you to handle such situations with support for dynamic selectors. In the case of our Github example it would be nice to select a particular result by supplying the owners organisation name. 

class ResultsPage
  include PageMagic

  element :results do |organisation:|

    selector xpath: "//h3/a[contains(text(), '#{organisation}')]/../.."

    # code for sub elements
  end
end

In the above example the selector looks for an element that has a link containing text that includes that organisation. The example uses a named parameter and is invoked as follows.

page.results(organisation: 'lvlup')

Starting a session

To start a PageMagic session simply decide what browser you want to use and pass it to PageMagic's .session method

session = PageMagic.session(browser: :chrome, url: 'https://www.github.com')

Your session won't do much besides navigating to the given url until you have mapped pages to it, so take a look at this next!

Note PageMagic supports having multiple sessions using different browsers at the same time :)

Rack applications and Rack::Test

To run a session against a rack application instead of a live site, simply supply the rack application when creating the session

session = PageMagic.session(application: YourRackApp, url: '/path_to_start_at')

By default PageMagic uses the Rack::Test driver for capybara however you are free to use any browser you like as long as the driver is registered for it.

session = PageMagic.session(application: YourRackApp, browser: :your_chosen_browser, url: '/path_to_start_at')

Out of the box, PageMagic supports the following as parameters to browser:

  • :chrome
  • :firefox
  • :poltergeist
  • :rack_test

Under the hood, PageMagic is using Capybara so you can register any Capybara compliant driver you want. See below for how to do this.

Note: We don't want to impose particular driver versions so PageMagic does not list any as dependencies. Therefore you will need add the requiste gems to your Gemfile.

Page mapping

With PageMagic you can map which pages should be used to handle which resources. Meaning that when a the page in the browser changes, PageMagic loads the correct PageObject class to handle it. This feature removes a lot of the juggling and brings back fluency to your code!

# define what pages map to what
session.define_page_mappings '/' => GitHub, '/search' => ResultsPage

You can use even use regular expressions and provide more than one mapping to the same page object class.

Note: By default mappings are matched against a URL's path. In addition, PageMagic supports mapping against both query string parameters and the fragement identifer (see below). Any combination of these can be used to define a page mapping.

Mapping against query string parameters

browser.define_page_mappings PageMagic.mapping(parameters: {parameter_name: string_or_regex}) => ResultsPage

Mapping against fragment identifiers

JavaScript MVC frameworks allow different resources to be mapped the fragment portion of URLs. That is the part of the URL that follows the Fragement identififer (#). PageMagic supports mapping page_objects against URL fragments.

browser.define_page_mappings PageMagic.mapping(fragment: string_or_regex) => ResultsPage

Loading pages from source

PageMagic supports loading page objects using html source. This technique can be useful for getting quick feedback that your templates correctly render based on your view objects. I.e you can test your templates in isolation.

class MyPage
  include PageMagic
  #element definitions
end

page_instance = Page.load(html_string)

Watchers

PageMagic lets you set a watcher on any element. Use watchers to decide when things have changed. The watch method can be called from anywhere within an element definition. For PageObjects it can only be called from within hooks and helper methods.

Note: Watchers are not inherited

Method watchers

Method watchers watch the output of the given method name.

button :javascript_button, css: '.fancy_button' do
  watch(:url)

  after_events do
    wait_until{changed?(:url)}
  end
end

Simple watchers

Use watch method passing two parameters, the first is the name of the element you want to keep an eye on and the second is the method that needs to be called to get the value that should be observed.

element :product_row, css '.cta' do
  watch(:total, :text)

  after_events do
    wait_until{changed?(:total)}
  end
end

Custom watchers

Custom watchers are defined by passing a name and block parameter to the watch method. The block returns the value that needs to be observed. Use watch in this way if you need to do something non standard to obtain a value or to access an element not located within the current element but elsewhere within the page.

element :product_row, css '.cta' do
  watch(:total) do
    session.nav.total.text
  end

  after_events do
    wait_until{changed?(:total)}
  end
end

Waiting

It's inevitable that if there is JavaScript on the page that you are going to have to wait for things to happen. PageMagic supplies the wait_until method that can be used anywhere you might need it. The wait_until method takes a block that it will execute until either that block returns true or the timeout occurs. See the method docs for details on configuring timeouts and retry intervals.

Drivers

Registering a custom driver

You can register any Capybara compliant driver as follows

#1. Define driver, constructor to PageMagic::Driver takes a list of browser aliases.
# Selenium Webdriver for example supports driving more than one.
Webkit = PageMagic::Driver.new(:webkit) do |app, options, browser_alias_chosen|
  # Write the code necessary to initialise the driver you have chosen
  require 'capybara/webkit'
  Capybara::Webkit::Driver.new(app, options)
end

#2. Register driver
PageMagic.drivers.register Webkit

#3. Use registered driver
session = PageMagic.session(browser: webkit, url: 'https://www.github.com')

Cucumber quick start

You can obviously use PageMagic anywhere you fancy but one of the places you might decide to use it is within a Cucumber test suite. If that's the case something like the following could prove useful.

Helper methods

Put the following in to features/support/page_magic.rb to make these helpers available to all of your steps.

require 'page_magic'
require 'active_support/inflector'
require 'your_pages'

World(Module.new do
        def page_class(string)
          "#{string}Page".delete(' ').constantize
        end

        def snake_case(string)
          string.delete(' ').underscore
        end

        def session
          $session ||= begin
            PageMagic.session(browser: :chrome, url: the_base_url).tap do |session|

              session.define_page_mappings '/login' => LoginPage,
                                           '/' => HomePage

            end
          end
        end
      end)

Example steps

Use the above helpers to navigate to pages with steps like the following.

Given /^I am on the '(.*)' page$/ do |page_name|
  session.visit(page_class(page_name))
end

And /^I set '(.*)' to be '(.*)'$/ do |field, value|
  session.send(snake_case(field)).set value
end

When /^I click '(.*)'$/ do |element|
  session.send(snake_case(element)).click
end

Then /^I should be on the '(.*)' page$/ do |page_name|
  current_page = session.current_page.class
  expected_page = page_class(page_name)
  fail "On #{current_page}, expected #{expected_page}" unless current_page == expected_page
end