Gemfile

source "https://rubygems.org"

gem 'mechanize' 

import.rb

require 'bundler'
Bundler.setup

LOGIN_TUMBLR = "[email protected]"
PASS_TUMBLR = "password"
TUMBLR_BLOG = 'xxx.tumblr.com'
CANALBLOG_NAME = "xxx"

require 'mechanize'
agent = Mechanize.new
page = agent.get("http://#{CANALBLOG_NAME}.canalblog.com/archives/index.html")

article_link = []

page.links.each do |l|
  if l.href =~ /\d{4}\/\d{2}/
    p_archive = l.click
    p_archive.links.each do |la|
      if la.href =~ /\d{4}\/\d{2}\/\d{2}\/\d+\.html$/
        article_link << la
      end
    end
  end
end

require 'net/http'


article_link.each do |article|
  year = article.href[/(\d{4})\/(\d{2})\/(\d{2})\/\d+\.html$/, 1]
  month = article.href[/(\d{4})\/(\d{2})\/(\d{2})\/\d+\.html$/, 2]
  day = article.href[/(\d{4})\/(\d{2})\/(\d{2})\/\d+\.html$/, 3]
  page = agent.get(article.href)
  title = page.search(".blogbody h2").text.gsub('Rétroliens', '')
  text = page.search(".blogbody").to_s[/<\/h2>((\W|\w)*)<div class="itemfooter"/, 1]
  while text.include?('itemfooter')
    text = text[/((\W|\w)*)<div class="itemfooter"/, 1]
  end

  text = text.gsub(/class=".+"/, '')
  text = text.gsub(/id=".+"/, '')
  text = text.gsub(/style=".+"/, '')

  p title
  p text
  p "date : #{year}-#{month}-#{day}"
  options = {
    :email => LOGIN_TUMBLR,
    :password => PASS_TUMBLR,
    :type => 'regular',
    :date => "#{year}-#{month}-#{day}",
    :format => 'html',
    :group => TUMBLR_BLOG,
    :title => title,
    :body => text
  }
  p Net::HTTP.post_form(URI.parse("http://www.tumblr.com/api/write"), options)
end