middleman conversion.rb

require 'pathname'
require "open-uri"
require 'uri'
require 'reverse_markdown'

# Custom gem (not published to rubygems, so reference the git repo)
# https://github.com/bensheldon/middleman_writer 
require 'middleman_writer'

# Helper method to save an array of articles
def save(articles)
  articles.each do |article|
    begin
      content_string = article.article.serialize
      article.path.open('w') { |f| f.write(content_string) }
    rescue => e
      puts "#{e}: #{article.path}"
    end
  end
end

# 1. Load up all the articles
articles = MiddlemanWriter::ArticleFile.load('source/articles')

# 2. Regex all the file urls and add them to their article's frontmatter
articles.each do |article|
  # Will need to reference the domain name of the blog
  matches = article.content.scan(/(http:[^"\n]*island94[^"\n]*(?:wp-content|\/files\/)[^"]*)/).flatten.uniq
  if matches.size > 0
    article.frontmatter['files'] = matches.map { |m| URI.escape(m) }
  else
    article.frontmatter.delete('files')
  end
end;
 
# articles.select{ |a| a.content.strip =~ /\A(?!<p|<ul|<blockquote|<div|<h).*/ }.each{ |a| a.content.gsub!(/^(.*)$/, '<p>\1</p>') }

# 3. Save the articles with their file urls in the frontmatter
save(articles)
 
articles.select{ |a| a.frontmatter.fetch('files', false) }.each do |article|
  directory = Pathname.new "source/articles/#{article.path.basename.to_s.match /[^\.]*/ }"
  directory.mkdir unless directory.directory?
 
  # clear out old failed file urls
  article.frontmatter.delete("failed_files")
 
  article.frontmatter['files'].each do |image_url|
    image_path = directory + Pathname.new(image_url).basename
 
    # check if the file has already been downloaded
    unless image_path.file?
      begin
        filename = directory + Pathname.new(image_path).basename
        File.open(filename, 'wb') do |fo|
          fo.write open(image_url).read 
        end
      rescue
        begin
          # whelp, delete the file
          filename.delete
        rescue
        end
        failed_files = article.frontmatter.fetch("failed_files", []) << image_url
        article.frontmatter["failed_files"] = failed_files.uniq
        puts "Failed: #{image_url}"
      end
    end
 
    # Now let's try to download the full-sizec big image for posterity
    bimage_url = image_url.sub(/-\d*x\d*\./, ".")
    bimage_path = directory + Pathname.new(bimage_url).basename
 
    unless bimage_path.file?
      filename = directory + Pathname.new(bimage_path).basename
 
      begin
        File.open(filename, 'wb') do |fo|
          fo.write open(bimage_url).read 
        end
      rescue
        begin
          filename.delete
        rescue
        end
        failed_files = article.frontmatter.fetch("failed_files", []) << bimage_url
        article.frontmatter["failed_files"] = failed_files.uniq
        puts "Failed (big): #{bimage_url}" 
      end
    end
  end
end;

# Save the articles
save(articles)

# Rewrite the content to reference the new relative paths
articles.each do |article|
  article.frontmatter.fetch('files', []).each do |image_url|
    rel_path = article.path.basename( '.html.markdown' ) + Pathname.new(image_url).basename
    article.content.gsub!(image_url.to_s, rel_path.to_s)
    puts "Rewrite content: #{image_url} -> #{rel_path} "
  end
end;

# Save the articles
save(articles)

redirects.rb

require 'csv'
text = File.read('redirects_p.csv').gsub(/\\"/,'""')

redirects = []
CSV.parse(text, headers: true, header_converters: :symbol) do |row|
  redirects << row
end

redirects.each do |redirect|
  article = articles.find { |a| a.frontmatter['wp:post_id'] == redirect[:id] }
  # articles.find{|f| p f.frontmatter['wp_link']; false}
  if article
    article.frontmatter['redirects'] = article.frontmatter.fetch('redirects', []) << redirect[:source]
  end
end;

articles.each do |article|
  redirects = article.frontmatter.fetch('redirects', []) << "/?p=#{article.frontmatter['wp:post_id']}"
  article.frontmatter['redirects'] = redirects.uniq
end

save(articles)