Skapa RSS för dagens diarium från en av länsstyrelserna

diarium.rb

# encoding: UTF-8
require 'mechanize'
require 'builder'

@agent = Mechanize.new

def get_list(diary_id)
  page = @agent.get("http://diarium.lansstyrelsen.se/default.aspx")

  form = page.form('aspnetForm')
  form['__EVENTTARGET'] = '_ctl0$SearchPlaceHolder$CaseSearch$btnCurrentDay'
  form['__EVENTARGUMENT'] = ''
  form['_ctl0:SearchPlaceHolder:CaseSearch:ddDiaryID'] = diary_id.to_s
  form['_ctl0:SearchPlaceHolder:CaseSearch:txtHiddenDiaryID'] = diary_id.to_s

  form.submit
end

diarium = get_list 11

builder = Builder::XmlMarkup.new

builder.instruct!
builder.rss :version => "2.0", "xmlns:content" => "http://purl.org/rss/1.0/modules/content/",
            "xmlns:dc"=>"http://purl.org/dc/elements/1.1/",
            "xmlns:media"=>"http://search.yahoo.com/mrss/",
            "xmlns:atom"=>"http://www.w3.org/2005/Atom",
            "xmlns:georss"=>"http://www.georss.org/georss" do

  builder.channel do
    builder.title 'Diarium'
    builder.link 'http://diarium.lansstyrelsen.se/'
    builder.description 'Länsstyrelsernas diarier'
    builder.language 'sv-se'
    diarium.parser.xpath('//table[@id="_ctl0_SearchPlaceHolder_caseGridView"]//tr[position() > 1]').each do |row|
      builder.item do

        link = 'http://diarium.lansstyrelsen.se/Case/' + row.children.first.xpath('a/@href').text
        (diarienummer, status, in_upp_datum, arenderubrik, avsandare_mottagare, postort,
            kommun, tillkomst, beslutsdatum, organisationsenhet) = row.children.map  { |col| col.text.strip }

        builder.category organisationsenhet
        builder.title arenderubrik

        builder.link link

        builder.description do
          builder.cdata!(avsandare_mottagare)
        end
      end
    end


  end
end

puts builder.to_s