kindler 0.0.1 → 0.0.2

data/bin/kindler

@@ -0,0 +1,29 @@
+#!/usr/bin/env ruby
+require 'rubygems'
+require 'optparse'
+require 'open-uri'
+require File.dirname(__FILE__) + "/../lib/kindler"
+# get options
+options = {:title=>'kindler_ebook',:urls=>ARGV}
+option_parser = OptionParser.new do |opts|
+	opts.banner = "Usage kindler url1 url2 url3 [-t my_book] [-d]"
+
+	opts.on("-d",'--debug','show debug infos') do |d|
+		options[:debug] = d
+	end
+
+	opts.on("-t",'--title','title of book') do |t|
+		options[:title] = t
+	end
+
+	opts.on_tail('-h','--help','show this message') do
+		puts opts
+		exit
+	end
+	
+end
+
+exit unless ARGV.length > 0
+book = Kindler::Book.new(options)
+book.generate
+

data/lib/kindler/version.rb

@@ -1,3 +1,3 @@
 module Kindler
-  VERSION = "0.0.1"
+  VERSION = "0.0.2"
 end

data/lib/kindler.rb

@@ -1,4 +1,5 @@
-require "kindler/version"
+require 'rubygems'
+require File.dirname(__FILE__)+"/kindler/version"
 require "readability"
 require "open-uri"
 # require 'mini_magick'
@@ -15,12 +16,17 @@ module Kindler
 		# @option urls [Array] urls to generate
 		# @option title [String] book title
 		# @option output_dir [String] directory want to generate
+		# @option debug [Boolean] whether puts debug infos
+		# @option keep_image [Boolean] whether keep images, default to true
 		def initialize(options={})
 			@urls = options[:urls] || {}
 			@title = options[:title] || ''
 			@output_dir = options[:output_dir] || './'
+			@keep_image = options[:keep_image] || true
+			@debug = options[:debug]
 			raise KindlerError.new("urls option could not be empty") if @urls.empty?
 			@author = options[:author] || ''
+			@images = []
 			@doc_infos = {}
 			# init doc infos by url
 			@urls.each {|url| @doc_infos[url]= {} }
@@ -41,9 +47,11 @@ module Kindler
 			make_generated_dirs
 			# generate
 			generate_html
+			localize_images
 			generate_toc
 			generate_opf
 			generate_ncx
+			write_to_disk
 			kindlegen
 			# clear
 		end
@@ -57,7 +65,7 @@ module Kindler
 		# make sure kindlegen is installed
 		# you can use "sudo brew install " to install it
 		def kindlegen
-			puts 'begin generate mobi'
+			debug 'begin generate mobi'
 			system("kindlegen #{tmp_dir}/#{@title}.opf ")
 		end
 
@@ -80,9 +88,8 @@ module Kindler
 			end
 			# append footer
 			contents << "</ul></body></html>"
-			File.open(file_path('contents'),'w') do |f|
-				f.puts contents
-			end
+
+			@toc = contents
 		end
 
 		# generate ncx , which is navigation
@@ -145,7 +152,7 @@ module Kindler
 				files_count += 1
 			end
 			contents << "</navMap></ncx>"
-			File.open("#{tmp_dir}/nav-contents.ncx",'w') { |f| f.puts contents }
+			@ncx = contents
 		end
 
 		# generate the opf, manifest of book,including all articles and images and css
@@ -189,7 +196,7 @@ module Kindler
 				files_count += 1
 			end
 			contents << "</spine><guide><reference href='contents.html' type='toc' title='Table of Contents'/></guide></package>"
-			File.open("#{tmp_dir}/#{@title}.opf",'w') {|f| f.puts contents}
+			@opf = contents
 		end
 
 		# generate every url to article in readable format
@@ -200,13 +207,32 @@ module Kindler
 				infos[:content] = html_wrap(article.title,article.content)
 				infos[:title] = article.title
 			end
-			# make html files
-			files_count = 1
+		end
+
+		def localize_images
+			images_count = 1
 			@doc_infos.each do |url,infos|
-				File.open(file_path(files_count.to_s.rjust(3,'0')),'w') do |f|
-					f.puts infos[:content]
+				article = Nokogiri::HTML(infos[:content])
+				article.css('img').each do |img|
+					image_remote_address = img.attr('src')
+					unless image_remote_address.start_with?('http')
+						image_remote_address = "http://#{URI(url).host}#{image_remote_address}"
+					end
+					image_local_address = File.join(tmp_dir,"#{images_count}#{File.extname(image_remote_address)}")
+					# download images
+					debug "begin fetch image #{image_remote_address}"
+					debug "save to #{image_local_address}"
+					File.open(image_local_address,'w') do |f|
+						f.puts open(image_remote_address).read
+					end
+					debug 'Image saved'
+					# replace local url address
+					img.attributes['src'].value = "#{images_count}#{File.extname(image_remote_address)}"
+					infos[:content] = article.inner_html
+					# add to manifest
+					@images << "#{images_count}#{File.extname(image_remote_address)}"
+					images_count += 1 
 				end
-				files_count += 1
 			end
 		end
 
@@ -228,9 +254,13 @@ module Kindler
 
 		# get readable document by url, using ruby-readability here
 		def readable_article(url)
-			puts "begin fetch url : #{url}"
+			debug "begin fetch url : #{url}"
 			source = open(url).read
-			Readability::Document.new(source)
+			if @keep_image
+				Readability::Document.new(source,:tags=>%w(div p img a),:attributes => %w[src href],:remove_empty_nodes => false)
+			else
+				Readability::Document.new(source)
+			end
 		end
 
 		# the dir path to generated files
@@ -244,11 +274,32 @@ module Kindler
 			FileUtils.mkdir_p tmp_dir unless File.exist?(tmp_dir)
 		end
 
+		def write_to_disk
+			File.open("#{tmp_dir}/nav-contents.ncx",'w') { |f| f.puts @ncx }
+			File.open(file_path('contents'),'w') {|f| f.puts @toc }
+			File.open("#{tmp_dir}/#{@title}.opf",'w') {|f| f.puts @opf}
+			# make html files
+			files_count = 1
+			@doc_infos.each do |url,infos|
+				File.open(file_path(files_count.to_s.rjust(3,'0')),'w') do |f|
+					f.puts infos[:content]
+				end
+				files_count += 1
+			end
+
+		end
+
 		# exist to clear tmp files such as ncx,opf or html other than mobi file
 		# keep them right now
 		def clear_tmp_dirs
 			
 		end
 
+		def debug(str)
+			return unless @debug
+			Rails.logger.info(str) if defined?(Rails)
+			puts str
+		end
+
 	end
 end

data/spec/cases/generator_spec.rb

@@ -5,7 +5,7 @@ describe "Mobi html file generator" do
 		title = "Test_book"
 		book = Kindler::Book.new ({:urls=>["http://blog.farmostwood.net/643.html",
 															"http://www.ifanr.com/69878","http://www.oneplus.info/archives/455"],
-															:title=>title,:author=>'mike'})
+															:title=>title,:author=>'mike',:debug=>true})
 		book.generate
 		File.exist?(mobi_book_path(title)).should == true
 	end
@@ -20,7 +20,7 @@ describe "Mobi html file generator" do
 		urls << "http://tumblr.intranation.com/post/766290565/how-set-up-your-own-private-git-server-linux"
 		urls << "http://antirez.com/post/what-is-wrong-with-2006-programming.html"
 		urls << "http://fak3r.com/2009/09/14/howto-build-your-own-open-source-dropbox-clone/"
-		book = Kindler::Book.new :urls=>urls,:title=>title,:author=>'mike'
+		book = Kindler::Book.new :urls=>urls,:title=>title,:author=>'mike',:debug=>true
 		book.generate 
 		File.exist?(mobi_book_path(title)).should == true
 	end
@@ -31,7 +31,7 @@ describe "Mobi html file generator" do
 		urls << "http://www.wowsai.com/home/space.php?uid=1&do=blog&id=4362&classid=2"
 		urls << "http://www.honeykennedy.com/2012/01/miss-moss-love-letters/"
 		urls << "http://www.mysenz.com/?p=3692"
-		book = Kindler::Book.new :urls=>urls,:title=>title,:author=>'mike',:output_dir=>'/Users/lidongbin/projects'
+		book = Kindler::Book.new :urls=>urls,:title=>title,:author=>'mike',:output_dir=>'/Users/lidongbin/projects',:debug=>true
 		book.generate 
 		File.exist?(mobi_book_path(title,'/Users/lidongbin/projects')).should == true
 	end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: kindler
 version: !ruby/object:Gem::Version
-  version: 0.0.1
+  version: 0.0.2
   prerelease: 
 platform: ruby
 authors:
@@ -13,7 +13,7 @@ date: 2012-01-30 00:00:00.000000000Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: mini_magick
-  requirement: &2162549440 !ruby/object:Gem::Requirement
+  requirement: &2157488800 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -21,10 +21,10 @@ dependencies:
         version: '0'
   type: :runtime
   prerelease: false
-  version_requirements: *2162549440
+  version_requirements: *2157488800
 - !ruby/object:Gem::Dependency
   name: ruby-readability
-  requirement: &2162549020 !ruby/object:Gem::Requirement
+  requirement: &2157488380 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -32,11 +32,12 @@ dependencies:
         version: '0'
   type: :runtime
   prerelease: false
-  version_requirements: *2162549020
+  version_requirements: *2157488380
 description: a simple gem to generate kindle mobi book
 email:
 - mike.d.1984@gmail.com
-executables: []
+executables:
+- kindler
 extensions: []
 extra_rdoc_files: []
 files:
@@ -45,11 +46,10 @@ files:
 - Guardfile
 - Rakefile
 - Readme.md
+- bin/kindler
 - kindler.gemspec
 - lib/kindler.rb
-- lib/kindler/html_generator.rb
 - lib/kindler/railtie.rb
-- lib/kindler/toc_generator.rb
 - lib/kindler/version.rb
 - spec/cases/generator_spec.rb
 - spec/spec_helper.rb

data/lib/kindler/html_generator.rb

@@ -1,7 +0,0 @@
-module Kindler
-	module HtmlGenerator
-		def g_html
-			
-		end
-	end
-end

data/lib/kindler/toc_generator.rb

@@ -1,7 +0,0 @@
-module Kindler
-	module TocGenerator
-		def g_toc
-			
-		end
-	end
-end