Contents

#!/usr/bin/env ruby
# -*- coding: utf-8 -*-

# Author::    TAC ([e-mail])

require 'pp'
require 'time'
require 'mechanize'

require_relative 'lib/yasuri/yasuri'

agent = Mechanize.new

uri = "http://www.asahi.com/"

# Node tree constructing by DSL
root = Yasuri.links_top '//*[@id="MainInner"]/div[1]/ul/li/a' do
  text_title   '//*[@id="MainInner"]/div[1]/div/h1'
  text_article '//*[@id="MainInner"]/div/div[@class="ArticleText"]'
end

# Node tree constructing by JSON
src = <<-EOJSON
   { "node"     : "links",
     "name"     : "root",
     "path"     : "//*[@id='MainInner']/div[1]/ul/li/a",
     "children" : [
                    { "node" : "text",
                      "name" : "title",
                      "path" : "//*[@id='MainInner']/div[1]/div/h1"
                    },
                    { "node" : "text",
                      "name" : "article",
                      "path" : "//*[@id='MainInner']/div/div[@class='ArticleText']"
                    }
                  ]
   }
EOJSON
root = Yasuri.json2tree(src)

# Access to parsed resources
page = agent.get(uri)
contents = root.inject(agent, page)

contents.each do |h|
  t = h['title']
  a = h['article']

  puts t
  puts a
  puts "=" * 100
end

Version data entries

16 entries across 16 versions & 1 rubygems

Version	Path
yasuri-3.2.0	app.rb
yasuri-3.1.0	app.rb
yasuri-3.0.0	app.rb
yasuri-2.0.13	app.rb
yasuri-2.0.12	app.rb
yasuri-1.9.12	app.rb
yasuri-1.9.11	app.rb
yasuri-2.0.11	app.rb
yasuri-0.0.11	app.rb
yasuri-0.0.10	app.rb
yasuri-0.0.9	app.rb
yasuri-0.0.8	app.rb
yasuri-0.0.7	app.rb
yasuri-0.0.6	app.rb
yasuri-0.0.5	app.rb
yasuri-0.0.4	app.rb