Skip to content

Instantly share code, notes, and snippets.

@akitaonrails
Created January 24, 2010 05:28
Show Gist options
  • Save akitaonrails/285032 to your computer and use it in GitHub Desktop.
Save akitaonrails/285032 to your computer and use it in GitHub Desktop.
Download from onemanga.com
#!/usr/bin/env ruby
#
# Put this script in your PATH and download from onemanga.com like this:
# onemanga_downloader.rb Bleach [chapter number]
#
# You will find the downloaded chapters under $HOME/Documents/OneManga/Bleach
#
# If you run this script without arguments, it will check your local manga downloads
# and check if there are any new chapters
#
# Updates
# 05/24 - taking into account redirection to 1000manga.com and age verification cookie
require 'rubygems'
require 'mechanize'
require 'nokogiri'
require 'open-uri'
manga_root = "http://www.onemanga.com/"
manga_download_folder = File.join(ENV['HOME'],"/Documents/OneManga/")
agent = Mechanize.new { |agent| agent.user_agent_alias = 'Mac Safari' }
if ARGV.size == 0
# no args means just to check for new chapters
mangas = Dir.glob(File.join(manga_download_folder, "*")).map do |f|
f.gsub(manga_download_folder, '')
end
mangas.each do |manga_name|
downloaded_chapters = Dir.glob(File.join(manga_download_folder, manga_name, "*")).map do |f|
f.gsub(File.join(manga_download_folder, manga_name, "/"), "").to_i
end.sort
last_chapter = downloaded_chapters.last
# index page
agent.get(manga_root + manga_name)
# find chapter
chapters = agent.page.links.map do |l|
$1.to_i if l.href =~ /#{manga_name}\/(\d+)/
end.compact.sort
most_recent_chapter = chapters.last
puts "#{last_chapter}/#{most_recent_chapter} - #{manga_name}"
end
exit 0 # go away
end
manga_name = ARGV.first || "Bakuman"
start_from_chapter = ARGV.size > 1 ? ARGV[1] : nil
manga_folder = File.join(manga_download_folder, manga_name)
puts "Creating #{manga_folder}"
FileUtils.mkdir_p(manga_folder)
# index page
agent.get(manga_root + manga_name)
# find chapter
chapter_link = agent.page.links.select do |l|
if start_from_chapter
l.href =~ /#{manga_name}\/#{start_from_chapter}\//
else
l.href =~ /#{manga_name}\/\d+/
end
end.reverse.first
# click the chapter link in the index page
agent.click chapter_link
another_site = agent.page.link_with(:text => "Read this series at 1000manga.com")
if another_site
agent.get :url => another_site.href, :referer => agent.page, :headers => { "cookie" => "age_verified=30" }
end
# first time in a chapter starts with "Begin reading ..."
agent.get :url => agent.page.links.select { |l| l.text =~ /Begin/ }.first.href, :referer => agent.page, :headers => { "cookie" => "age_verified=30" }
chapter_number = nil
chapter_folder = ""
# go all the way. the navigation stop in the last chapter with a bookmark link
while (agent.page / "#id_bookmark_click").empty?
break if agent.page.forms.empty?
# create the chapter folder if it changes
current_chapter_number = agent.page.uri.to_s.split("/")[-2] # /[manga]/[chapter]/[page]
if chapter_number != current_chapter_number
chapter_number = current_chapter_number
chapter_folder = File.join(manga_download_folder, manga_name, chapter_number)
puts "Creating #{chapter_folder}"
FileUtils.mkdir_p(chapter_folder)
end
# download image file
img_uri = agent.page.search("//img[@class='manga-page']").first["src"]
image_file = File.join(chapter_folder, img_uri.split("/").last)
open(image_file, 'wb') do |file|
puts "Downloading #{img_uri} to #{image_file}"
file.write(open(img_uri).read)
end
# next page
agent.get :url => agent.page.links.select { |link| link.text.strip.empty? }.first.href, :referer => agent.page, :headers => { "cookie" => "age_verified=30" }
end
#!/usr/bin/env ruby
# Generate PDFs for the OneManga downloaded images
# Use like this:
# ruby onemanga_pdf.rb Bakuman 5 10 50
#
# Meaning:
# Manga: Bakuman (name of the folder)
# Number of volumes: 5 (number of generated PDF files)
# Starting chapter: 10 (default is 0)
# Last chapter: 50 (default is -1, which is everything)
#
require 'rubygems'
require 'enumerator'
require 'prawn'
if ARGV.size < 1
puts "Pass a manga name.\nonemanga_pdf.rb manga_name [# volumes] [# start chapter] [# last chapter]"
exit(1)
end
manga_name = ARGV[0]
volumes = ARGV[1].to_i || 1 # defaults to just one volume
start_chapter = ARGV[2].to_i || 1
last_chapter = ARGV[3].to_i > 0 ? ARGV[3].to_i : -1 # defaults to all chapters
puts "Preparing to pring #{manga_name} in #{volumes} volumes."
manga_download_folder = File.join(ENV['HOME'],"/Documents/OneManga/")
manga_current_folder = File.join(manga_download_folder, manga_name)
folders = Dir.glob(File.join(manga_current_folder, "*")) # fetch all chapters
# correctly sort directories names based o the chapter number at the end
folders.sort! do |a, b|
a =~ /\/(\d+)$/
x = $1.to_i
b =~ /\/(\d+)$/
y = $1.to_i
x <=> y
end
folders = folders[start_chapter..last_chapter] # limit chapters to process
current_volume = 1
folders.each_slice(folders.size / volumes) do |chapter_folders|
manga_file = File.join(manga_download_folder, manga_name + "_#{current_volume}.pdf")
File.delete(manga_file) if File.exists?(manga_file)
print "\nPrinting volume #{manga_file} ..."
Prawn::Document.generate(manga_file, :margin => 0, :page_size => "A4") do
chapter_folders.each do |chapter|
Dir.glob(File.join(chapter, "*")).each do |page|
next if page =~ /credit/
image page, :fit => [595.28,841.89]
start_new_page
print "."
end
end
end
current_volume += 1
end
puts "Finished printing out #{volumes} volumes for #{manga_name}."
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment