initial

.gitignore

@@ -0,0 +1,2 @@
+src/
+run.sh

README.md

@@ -0,0 +1,6 @@
+# docs_on_kindle
+
+This project aims to put web documention for popular software tools on the
+Kindle.
+
+

css/kindle.css

@@ -0,0 +1,17 @@
+p { text-indent: 0; }
+
+p, H1, H2, H3, H4, H5, H6, H7, H8, table, pre { margin-top: 1em;}
+
+/* doesn't work apparently: */
+dt {
+  display:block;
+  margin-top: 1em;
+}
+
+.pagebreak { page-break-before: always; }
+#toc H3 {
+  text-indent: 1em;
+}
+#toc .document {
+  text-indent: 2em;
+}

lib/docs_on_kindle.rb

@@ -0,0 +1,82 @@
+=begin
+
+Require this file and include this module into each recipe.
+
+Your recipe class is responsible for getting all the source HTML necessary to
+build the ebook.
+
+=end
+
+require 'fileutils'
+require 'nokogiri'
+require 'fileutils'
+require 'yaml'
+require 'date'
+
+module DocsOnKindle
+
+  STYLESHEET = File.absolute_path "css/kindle.css"
+
+  def add_head_section(doc, title)
+    head = Nokogiri::XML::Node.new "head", doc
+    title_node = Nokogiri::XML::Node.new "title", doc
+    title_node.content = title
+    title_node.parent = head
+    css = Nokogiri::XML::Node.new "link", doc
+    css['rel'] = 'stylesheet'
+    css['type'] = 'text/css'
+    css['href'] = STYLESHEET
+    css.parent = head
+    doc.at("body").before head
+  end
+
+  def run cmd
+    puts "  #{cmd}"
+    `#{cmd}`
+  end
+
+  def download_images! doc
+    doc.search('img').each {|img|
+      src = img[:src] 
+      /(?<img_file>[^\/]+)$/ =~ src
+      FileUtils::mkdir_p 'images'
+      FileUtils::mkdir_p 'grayscale_images'
+      unless File.size?("images/#{img_file}")
+        run "curl -Ls '#{src}' > images/#{img_file}"
+      end
+      grayscale_image_path = "grayscale_images/#{img_file.gsub(/(\.\w+)$/, "-grayscale.gif")}"
+      unless File.size?(grayscale_image_path)
+        run "convert images/#{img_file}[0] -type Grayscale -depth 8 -resize '400x300>' #{grayscale_image_path}"
+      end
+      img['src'] = [Dir.pwd, grayscale_image_path].join("/")
+    }
+  end
+
+  def fixup_html! doc
+
+    # Sort of a hack to improve dt elements spacing
+    # Using a css rule margin-top doesn't work
+    doc.search('dt').each {|dt|
+      dt.children.first.before(Nokogiri::XML::Node.new("br", doc))
+    }
+
+    # We want to remove nested 'p' tags in 'li' tags, because these introduce an undesirable 
+    # blank line after the bullet. The expected CSS fix doesn't work.
+    doc.search('li').each {|li|
+      xs = li.search("p").map {|p|
+        # remove surrounding paragraph tags
+        p.children.each {|c|
+          li.add_child c
+        }
+        p.remove
+      }.flatten
+
+    }
+
+  end
+
+  def mobi!
+    File.open("_document.yml", 'w'){|f| f.puts document.to_yaml}
+    exec 'kindlerb'
+  end
+end

recipes/heroku.rb

@@ -0,0 +1,99 @@
+#!/usr/bin/env ruby
+
+require 'docs_on_kindle'
+
+class HerokuDocs
+  include ::DocsOnKindle 
+
+  OUTPUT_DIR = "src/heroku"
+  `mkdir -p #{OUTPUT_DIR}`
+
+  def get_source_files
+    start_url = "http://devcenter.heroku.com/categories/add-on-documentation" 
+    @start_doc = Nokogiri::HTML `curl -s #{start_url}`
+    File.open("#{OUTPUT_DIR}/sections.yml", 'w') {|f|f.puts extract_sections.to_yaml}
+  end
+
+  def document 
+    {
+      'doc_uuid' => "heroku-docs-#{Date.today.to_s}",
+      'title' => "Heroku Documentation",
+      'publisher' => "Heroku",
+      'author' => "Heroku",
+      'subject' => 'Reference',
+      'date' => Date.today.to_s,
+      'cover' => nil,
+      'masthead' => nil,
+      'mobi_outfile' => "heroku-guide.#{Date.today.to_s}.mobi"
+    }
+  end
+
+  def build_kindlerb_tree
+    sections = YAML::load_file "#{OUTPUT_DIR}/sections.yml"
+    sections.select! {|s| !s[:articles].empty?}
+    Dir.chdir OUTPUT_DIR do
+      sections.each_with_index {|s, section_idx|
+        title = s[:title]
+        FileUtils::mkdir_p("sections/%03d" % section_idx)
+        File.open("sections/%03d/_section.txt" % section_idx, 'w') {|f| f.puts title}
+        puts "sections/%03d -> #{title}" % section_idx
+        # save articles
+        s[:articles].each_with_index {|a, item_idx|
+          article_title = a[:title]
+          /(?<path>articles\/[\w-]+)(#\w+|)$/ =~ a[:url]
+          puts a[:url], path
+          item = Nokogiri::HTML(File.read path)
+
+          download_images! item
+          fixup_html! item
+
+          item_path = "sections/%03d/%03d.html" % [section_idx, item_idx] 
+          add_head_section item, article_title
+          # fix all image links
+          # item.search("img").each { |img|
+            #img['src'] = "#{Dir.pwd}/#{img['src']}"
+          #}
+          File.open(item_path, 'w'){|f| f.puts item.to_html}
+          puts "  #{item_path} -> #{article_title}"
+        }
+      }
+      mobi!
+    end
+  end
+
+  def extract_sections
+    @start_doc.search('select[@id=quicknav] option').map {|o| 
+      title = o.inner_text
+      $stderr.puts "#{title}"
+      s = { 
+        title: title,
+        articles: articles(`curl -s http://devcenter.heroku.com#{o[:value]}`) 
+      }
+    }
+  end
+
+  def articles html
+    category_page = Nokogiri::HTML html 
+    xs = category_page.search("ul.articles a").map {|x|
+      title = x.inner_text.strip
+      href = x[:href] =~ /^http/ ? x[:href] : "http://devcenter.heroku.com#{x[:href]}" 
+      $stderr.puts "-  #{title}"
+      a = { 
+        title: title,
+        url: href
+      }
+    }
+  end
+
+  def article href
+    /(?<filename>[\w-]+)$/ =~ href
+    a = Nokogiri::HTML `curl -s #{href}`    
+    FileUtils::mkdir_p "#{OUTPUT_DIR}/articles"
+    path = "#{OUTPUT_DIR}/articles/#{filename}"
+    File.open(path, 'w') {|f| f.puts(a.at('article').inner_html)}
+  end
+end
+
+
+HerokuDocs.new.get_source_files
+HerokuDocs.new.build_kindlerb_tree