suby 0.0.3

data/bin/suby

@@ -0,0 +1,101 @@
+#!/usr/bin/env ruby
+
+if ARGV.include? '-h'
+  puts "#{File.basename $0} video"
+  puts "  Find and download subtitles for the given video file"
+  exit
+end
+
+require 'net/http'
+require 'nokogiri'
+
+SUB_EXTENSIONS = %w[srt sub]
+archive = '__archive__'
+site = 'www.tvsubtitles.net'
+search_url = '/search.php'
+lang = 'en'
+
+def extract_subs_from_archive(archive)
+  case `file #{archive}`
+  when /Zip archive data/
+    subs = `unzip -qql #{archive}`.scan(/\d{2}:\d{2}   (.+?(?:#{SUB_EXTENSIONS.join '|'}))$/).map(&:first)
+    raise "no subtitles in #{archive}" if subs.empty?
+    subs_for_unzip = subs.map { |sub| sub.gsub(/(\[|\])/) { "\\#{$1}" } }
+    system 'unzip', archive, *subs_for_unzip, 1 => :close
+    puts "found subtitles: #{subs.join(', ')}"
+  else
+    raise "unknown archive type (#{archive})"
+  end
+
+  # Cleaning
+  File.unlink archive
+  subs
+end
+
+# cache
+show_urls = {}
+show_pages = {}
+
+ARGV.each { |file|
+  next if SUB_EXTENSIONS.include? File.extname(file)
+  next(puts "Skipping: #{file}") if SUB_EXTENSIONS.any? { |ext|
+    File.exist? File.basename(file, File.extname(file)) + ".#{ext}" }
+
+  unless /^(?<tvshow>.+) (?<season>\d{1,2})x(?<episode>\d{1,2}) - .+\.[a-z]+?$/ =~ file
+    next puts "wrong file format (#{file}). Must be:\n<tvshow> <season>x<episode> - <title>.<ext>"
+  end
+  season, episode = [season, episode].map(&:to_i)
+  puts "Searching subtitles for #{file}:"
+  puts "Show: #{tvshow}, Season: #{season}, Episode: #{episode}"
+
+  Net::HTTP.start(site) { |tvsubtitles|
+    # search tvshow
+    tvshow_url = (show_urls[tvshow] ||= begin
+      post = Net::HTTP::Post.new(search_url)
+      post.form_data = { 'q' => tvshow }
+      results = Nokogiri tvsubtitles.request(post).body
+      url = results.css('ul li div a').first[:href]
+
+      raise 'could not find the show' unless /^\/tvshow-(\d+)\.html$/ =~ url
+      "/tvshow-#{$1}-#{season}.html"
+    end)
+    puts "tvshow url: #{tvshow_url}"
+
+    # search episode
+    tvshow = (show_pages[tvshow] ||= Nokogiri tvsubtitles.get(tvshow_url).body)
+    tr = tvshow.css('div.left_articles table tr').find { |tr|
+      tr.children.find { |td| td.name == 'td' && td.text =~ /\A#{season}x0?#{episode}\z/ }
+    }
+    episode_url = nil
+    tr.children.find { |td|
+      td.children.find { |a|
+        a.name == 'a' && a[:href].start_with?('episode') && episode_url = a[:href]
+      }
+    }
+
+    raise "invalid episode url: #{episode_url}" unless episode_url =~ /^episode-(\d+)\.html$/
+    episode_url = "/episode-#{$1}-#{lang}.html"
+    puts "episode url: #{episode_url}"
+
+    # subtitles
+    subtitles = Nokogiri tvsubtitles.get(episode_url).body
+
+    # TODO: choose 720p or most downloaded instead of first found
+    subtitle_url = subtitles.css('div.left_articles a').find { |a| a.name == 'a' && a[:href].start_with?('/subtitle') }[:href]
+    raise 'invalid subtitle url' unless subtitle_url =~ /^\/subtitle-(\d+)\.html/
+    puts "subtitle url: #{subtitle_url}"
+
+    # download
+    download_url = tvsubtitles.get("/download-#{$1}.html")['Location']
+    download_url = URI.escape('/'+download_url)
+    puts "download url: #{download_url}"
+
+    # extract
+    zip = tvsubtitles.get(download_url).body
+    open(archive, 'wb') { |f| f.write zip }
+    subs = extract_subs_from_archive(archive)
+    new_name = File.basename(file, File.extname(file))+File.extname(subs.first)
+    File.rename(subs.first, new_name)
+    puts "Renaming to #{new_name}"
+  }
+}

metadata

@@ -0,0 +1,46 @@
+--- !ruby/object:Gem::Specification
+name: suby
+version: !ruby/object:Gem::Version
+  version: 0.0.3
+  prerelease: 
+platform: ruby
+authors:
+- eregon
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2011-06-15 00:00:00.000000000 Z
+dependencies: []
+description: Find and download subtitles
+email: 
+executables:
+- suby
+extensions: []
+extra_rdoc_files: []
+files:
+- bin/suby
+homepage: 
+licenses: []
+post_install_message: 
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project: 
+rubygems_version: 1.8.5
+signing_key: 
+specification_version: 3
+summary: Subtitles' downloader
+test_files: []