NOTE: This is not real software, maybe it will be one day but it is not today. It is an exercise on Readme Driven Development, or, put in an alternative way, insane fantasy.
Fragments is a Ruby library to scrape HTML using sample templates. I use it for testing, but for sure it has more uses (basically, anything that implies parsing HTML and treating it as structured data, and that you could do using Scrapi, Scrubyt, or plain Nokogiri). Think of it like a template system, the other way around:
+------+ +-----------+ +------+
| data | ----> | template | ----> | HTML |
+------+ +-----------+ +------+
+------+ +-----------+ +------+
| HTML | ----> | fragments | ----> | data |
+------+ +-----------+ +------+
Actually, syntax is very similar to a subset of Liquid's.
Simple example:
template = Fragments::Template.new(%q{
<h1>{{ title }}</h1>
<div class="content">
{{ content }}
</div>
})
data = template.parse(%q{
<html>
<head>
<title>Hello, world</title>
</head>
<body>
<h1>Hello, world</h1>
<div class="content">
Lorem ipsum dolor sit amet...
</div>
</body>
</html>
})
data.title # "Hello, world"
data.content # "Lorem ipsum dolor sit amet..."
If you expect more than one instance of your fragment, you can use find instead of parse:
template = Fragments::Template.new(%q{
<h1>{{ title }}</h1>
<div class="content">
{{ content }}
</div>
})
data = template.find(%q{
<html>
<head>
<title>My blog</title>
</head>
<body>
<h1>Hello world</h1>
<div class="content">
Lorem ipsum dolor sit amet...
</div>
<h1>A second article</h1>
<div class="content">
Wadus wadus...
</div>
</body>
</html>
})
data.size # 2
data[0].title # "Hello, world"
data[0].content # "Lorem ipsum dolor sit amet..."
data[1].title # "A second article"
data[1].content # "Wadus wadus..."
But in that case you can also simply use Liquid's iterators:
template = Fragments::Template.new(%q{
<h1>Products from {{ company.name }} </h1>
<ul id="products">
{% for product in products %}
<li>
<h2>{{ product.name }}</h2>
Only {{ product.price }}
{{ product.description }}
</li>
{% endfor %}
</ul>
})
data = template.parse(%q{
<html>
<head>
<title>Products from The Foo Bar</title>
</head>
<body>
<h1>Products from The Foo Bar</h1>
<ul id="products">
<li>
<h2>Club Mate</h2>
Only 1.00€
Mate based drink
</li>
<li>
<h2>Berliner</h2>
Only 2.00€
The classical beer from Berlin
</li>
</ul>
</body>
</html>
})
data.company.name # "The Foo Bar"
data.products.size # 2
data.products[0].name # "Club Mate"
data.products[0].price # "1.00€"
data.products[0].description # "Mate based drink"
data.products[1].name # "Berliner"
data.products[1].price # "2.00€"
data.products[1].description # "The classical beer from Berlin"
(quick and dirty, not very sure about this API; maybe it's simpler to have a helper that returns the scraped response, and do plain method calls and assertions)
Provided you store your fragment templates under spec/support/fragments/<name>.html.fragment, you can match against them in your Capybara specs:
# for using with the second example template
page.should have_fragment(:product, :count => 2)
# for using with the third example template
page.should have_fragment(:company, :company => { :name => "The Foo Bar" },
:products => [ {:name => 'Club Mate', :price => '1.00€'},
{:name => 'Berliner', :price => '2.00€'}])
An example of the aforementioned simpler approach:
company = page.as(:company)
company.name.should == "The Foo Bar"
company.should have(2).products
company.products[0].name.should == 'Club Mate'
...