Script to export posts from a Wordpress XML file to Jekyll (Textile) files. Collects comments in a YAML file too.

wordpressxml2jekyll.rb

#!/usr/bin/env ruby

# Input: WordPress XML export file.
# Outputs: a series of Textile files ready to be included in a Jekyll site,
# and comments.yml which contains all approved comments with metadata which
# can be used for a Disqus import.

require 'rubygems'
require 'hpricot'
require 'clothred'
require 'time'
require 'yaml'

WORDPRESS_XML_FILE_PATH = "/home/marko/Documents/wordpress.2010-01-01.xml"
OUTPUT_PATH = "/tmp/export"
ORIGINAL_DOMAIN = "http://example.com"

class Post
  attr_accessor :title, :post_date, :created_at, :slug, :url, :content, :textile_content
  attr_accessor :hpricot_element

  def initialize(item)
    @hpricot_element = item

    @title = item.search("title").first.inner_text

    @post_date = item.search("wp:post_date").first.inner_text
    @created_at = Date.parse(post_date)

    @slug = item.search("wp:post_name").first.inner_text

    @url = ORIGINAL_DOMAIN + "/" + item.search("wp:post_date_gmt").first.inner_text[0, 10].gsub(/-/, "/") + "/" + @slug

    @content = item.search("content:encoded").first.inner_text
    text = ClothRed.new(content)
    @textile_content = text.to_textile
  end

  def to_jekyll
    buf = ""
    buf << "---\n"
    buf << "layout: post\n"
    buf << "title: #{title}\n"
    buf << "---\n\n"
    buf << textile_content
  end

  def save(root_path)
    File.open("#{root_path}/#{created_at}-#{slug}.textile", "w") { |file| file.write self.to_jekyll }
    self
  end

  def save_comments(path)
    comment_elements = @hpricot_element.search("wp:comment").reject do |c|
      c.search("wp:comment_approved").inner_text != "1"
    end

    File.open("#{path}/comments.yml", "a") do |yaml_file|
      comment_elements.collect { |el| Comment.new(self, el) }.each { |comment| comment.write_to yaml_file }
    end
  end

  class << self
    def parse(element, path)
      return nil unless element.is_a?(Hpricot::Elem)
      post = Post.new(element)
      post.save(path)
    end
  end
end

class Comment
  attr_accessor :author_name, :author_email, :author_url, :content, :post

  def initialize(post, element)
    @post_url = post.url + "/"
    @author_name = element.search("wp:comment_author").first.inner_text
    @author_email = element.search("wp:comment_author_email").first.inner_text
    @author_url = element.search("wp:comment_author_url").first.inner_text
    @content = element.search("wp:comment_content").first.inner_text || ""

    comment_date = element.search("wp:comment_date_gmt").first.inner_text
    @created_at = Time.parse("#{comment_date} GMT")
  end

  def write_to(file)
    file.write self.to_yaml + "\n" unless @content.size == 0
  end
end

# main

doc = Hpricot(File.open(WORDPRESS_XML_FILE_PATH))

File.open("#{OUTPUT_PATH}/comments.yml", "w") { |f| }

(doc / "item").each do |item|
  post = Post.parse(item, OUTPUT_PATH)
  post.save_comments(OUTPUT_PATH)
end

yamlcomments2disqus.rb

#!/usr/bin/env ruby

# Takes comments.yml generated by wordpressxml2jekyll.rb and posts them to your Disqus forum.
# sudo gem install disqus

require 'disqus'
require 'disqus/api'

COMMENTS_YAML_FILE = '/tmp/export/comments.yml'

Disqus::defaults[:api_key] = "N4wWciM45UAfBJe6QbylR0mfQ340WH7kdEKlBi7q5Tb0QeAKHOxP7wC6W5WyJWWz"

forum_id = Disqus::Api.get_forum_list["message"].first["id"]
fak = Disqus::Api.get_forum_api_key(:forum_id => forum_id)["message"]

File.open(COMMENTS_YAML_FILE) do |yf|
  YAML.each_document( yf ) do |c|
    thread = Disqus::Api.get_thread_by_url(:forum_api_key => fak, :url => c.ivars["post_url"])
    Disqus::Api.create_post(:forum_api_key => fak,
                            :thread_id => thread["message"]["id"],
                            :message => c.ivars["content"],
                            :author_name => c.ivars["author_name"],
                            :author_email => c.ivars["author_email"],
                            :author_url => c.ivars["author_url"],
                            :created_at => Time.parse(c.ivars["created_at"].to_s).strftime("%Y-%m-%dT%H:%M"))
  end
end

Amazingly useful - thank you!

This looks awesome!
Would this support non-Latin characters?

@neozheng I don't know, my blog posts don't have them. But I think that it would. If you try it please let me know.

@markoa just tried. perfectly transformed the file contents but not the names. so much better than the built-in converter which breaks everything.

@neozhang since slug and content are read pretty much the same way I suspect something special needs to be done in the File.open call, but I haven't really worked with such issues before. Feel free to fork the gist and send more comments.

The title in the YAML Front Matter is not properly escaped. If the title of the post is for example "command not found: clear" the YAML Front Matter becomes
---
layout: post
title: WordPress migration: ” (quotes) and ‘ (apostrophe) being replaced with “ and ’

---

but and it should be

---
layout: post
title: "Wordpress migration:  \" (quotes) and ' (apostrophe) being replaced with \xC3\xA2\xE2\x82\xAC\xC5\x93 and \xC3\xA2\xE2\x82\xAC\xE2\x84\xA2"


---

I fixed that in my gist http://gist.github.com/500506