undertexter 0.1.4 → 0.1.5

data/Gemfile.lock

@@ -1,10 +1,10 @@
 PATH
   remote: .
   specs:
-    undertexter (0.1.3)
-      hpricot (= 0.8.2)
+    undertexter (0.1.4)
       levenshteinish
       mimer_plus
+      nokogiri
       rest-client
 
 GEM
@@ -12,11 +12,11 @@ GEM
   specs:
     diff-lcs (1.1.2)
     hintable_levenshtein (0.0.3)
-    hpricot (0.8.2)
     levenshteinish (0.0.1)
       hintable_levenshtein
     mime-types (1.16)
     mimer_plus (0.0.4)
+    nokogiri (1.4.4)
     rest-client (1.6.1)
       mime-types (>= 1.16)
     rspec (2.4.0)

data/README.markdown

@@ -3,7 +3,7 @@
 
 Undertexter provides a basic search client that makes it possible to search for swedish and english subtitles on [Undertexter.se](http://undertexter.se)
 
-*Support for other subtitles sites will be added in the future. Follow this project to know when*
+*Support for other subtitles sites will be added in the future. Follow this project to know when.*
 
 ## How to use
 
@@ -16,7 +16,7 @@ This is how to use it in `irb`.
     # => true
     
     $ subtite = Undertexter.find("tt0840361").first
-    => #<Subtitle:0x1020fff98 @downloads=8328, @movie_title="The Town", @title="The.Town.2010....BRRip", @url="http://www.undertexter.se/?p=undertext&id=23711", @cds=1>
+    => #<Container::Subtitle:0x1020fff98 @downloads=8328, @movie_title="The Town", @title="The.Town.2010....BRRip", @url="http://www.undertexter.se/?p=undertext&id=23711", @cds=1>
     $ subtitle.downloads
     => 8328
     $ subtitle.movie_title
@@ -58,7 +58,7 @@ This is how to use it in `irb`.
 ### Find the right subtitle based on the **release name** of the movie
     
     $ Undertexter.find("tt0840361").based_on("The Town EXTENDED 2010 480p BRRip XviD AC3 FLAWL3SS")
-    => #<Subtitle:0x00000101b739d0 @cds=1, @downloads=1644, @title="The.Town.EXTENDED.2010.480p.BRRip.XviD.AC3-FLAWL3SS", @details="http://www.undertexter.se/?p=undertext&id=23752", @movie_title="The Town", @language=:swedish>
+    => #<Container::Subtitle:0x00000101b739d0 @cds=1, @downloads=1644, @title="The.Town.EXTENDED.2010.480p.BRRip.XviD.AC3-FLAWL3SS", @details="http://www.undertexter.se/?p=undertext&id=23752", @movie_title="The Town", @language=:swedish>
     
 ### Specify how sensitive the `based_on` method should be, from `0.0` to `1.0`
     
@@ -66,14 +66,14 @@ This is how to use it in `irb`.
     => nil
     
     $ Undertexter.find("tt0840361").based_on("The Town EXTENDED 2010 480p BRRip XviD AC3 FLAWL3SS", limit: 0.4)
-    => #<Subtitle:0x00000101b8d808 @cds=1, @downloads=1644, @title="The.Town.EXTENDED.2010.480p.BRRip.XviD.AC3-FLAWL3SS", @details="http://www.undertexter.se/?p=undertext&id=23752", @movie_title="The Town", @language=:swedish>
+    => #<Container::Subtitle:0x00000101b8d808 @cds=1, @downloads=1644, @title="The.Town.EXTENDED.2010.480p.BRRip.XviD.AC3-FLAWL3SS", @details="http://www.undertexter.se/?p=undertext&id=23752", @movie_title="The Town", @language=:swedish>
     
     $ Undertexter.find("tt0840361").based_on("The.Town.EXTENDED.2010.480p.BRRip.XviD.AC3-FLAWL3SS", limit: 0.0)
-    => #<Subtitle:0x00000101b8d718 @cds=1, @downloads=1644, @title="The.Town.EXTENDED.2010.480p.BRRip.XviD.AC3-FLAWL3SS", @details="http://www.undertexter.se/?p=undertext&id=23752", @movie_title="The Town", @language=:swedish>
+    => #<Container::Subtitle:0x00000101b8d718 @cds=1, @downloads=1644, @title="The.Town.EXTENDED.2010.480p.BRRip.XviD.AC3-FLAWL3SS", @details="http://www.undertexter.se/?p=undertext&id=23752", @movie_title="The Town", @language=:swedish>
 
 ## What is being returned?
 
-The find method returns an `Array` with zero or more subtitles. Every subtitle provides some basic accessors.
+The find method returns an `Array` with zero or more `Container::Subtitle` instances. Every object provides some basic accessors.
 
 - `movie_title` (String) The official name of the movie.
 - `cds` (Integer) The amount of cds that the release should contain.

data/lib/subtitle.rb

@@ -1,65 +1,67 @@
 require 'fileutils'
 require 'mimer_plus'
 
-class Subtitle
-  attr_accessor :details, :downloads, :cds, :title, :movie_title, :url
+module Container
+  class Subtitle
+    attr_accessor :details, :downloads, :cds, :title, :movie_title, :url
   
-  def initialize(args)
-    args.keys.each { |name| instance_variable_set "@" + name.to_s, args[name] }
-  end
-  
-  def url
-    @language == :english ? "http://eng.undertexter.se/subtitle.php?id=#{id}" : "http://undertexter.se/laddatext.php?id=#{id}"
-  end
+    def initialize(args)
+      args.keys.each { |name| instance_variable_set "@" + name.to_s, args[name] }
+    end
   
-  # Downloading the file and saves it disk
-  def download!(args = {})
-    if args[:to].nil?
-      dir = "/tmp"
-      file_name = "#{dir}/#{generate_file_name}"
-    else
-      dir = generate_custom_file_path(args)
-      file_name = "#{dir}/#{generate_file_name}"
+    def url
+      @language == :english ? "http://eng.undertexter.se/subtitle.php?id=#{id}" : "http://undertexter.se/laddatext.php?id=#{id}"
     end
+  
+    # Downloading the file and saves it disk
+    def download!(args = {})
+      if args[:to].nil?
+        dir = "/tmp"
+        file_name = "#{dir}/#{generate_file_name}"
+      else
+        dir = generate_custom_file_path(args)
+        file_name = "#{dir}/#{generate_file_name}"
+      end
     
-    data = RestClient.get(self.url, :timeout => 10) rescue nil
-    file = File.new(file_name, 'w')
-    file.write(data)
-    file.close
+      data = RestClient.get(self.url, :timeout => 10) rescue nil
+      file = File.new(file_name, 'w')
+      file.write(data)
+      file.close
     
-    type = Mimer.identify(file_name)
+      type = Mimer.identify(file_name)
     
-    if type.zip?
-      file_ending = ".zip"
-    elsif type.rar?
-      file_ending = ".rar"
-    else
-      file_ending = ""
-    end
+      if type.zip?
+        file_ending = ".zip"
+      elsif type.rar?
+        file_ending = ".rar"
+      else
+        file_ending = ""
+      end
     
-    new_file_name = "#{dir}/#{title.gsub(/\s+/, '.')}#{file_ending}"
+      new_file_name = "#{dir}/#{title.gsub(/\s+/, '.')}#{file_ending}"
     
-    # Changing the name on the file
-    FileUtils.mv(file_name, new_file_name)
+      # Changing the name on the file
+      FileUtils.mv(file_name, new_file_name)
     
-    # I like return :)
-    return new_file_name
-  end
-  
-  private
-    def id
-      @details.match(/id=(\d+)/)[1]
+      # I like return :)
+      return new_file_name
     end
+  
+    private
+      def id
+        @details.match(/id=(\d+)/)[1]
+      end
     
-    def generate_file_name
-      (0...30).map{65.+(rand(25)).chr}.join.downcase
-    end
+      def generate_file_name
+        (0...30).map{65.+(rand(25)).chr}.join.downcase
+      end
     
-    def generate_custom_file_path(args)
-      # If the path is relative
-      args[:to] = File.expand_path(args[:to]) unless args[:to].match(/^\//)
+      def generate_custom_file_path(args)
+        # If the path is relative
+        args[:to] = File.expand_path(args[:to]) unless args[:to].match(/^\//)
 
-      # Makes sure that every directory structure looks the same
-      Dir.new(args[:to]).path
-    end
+        # Makes sure that every directory structure looks the same
+        Dir.new(args[:to]).path
+      end
+  end
 end

data/lib/undertexter.rb

@@ -2,22 +2,10 @@
 
 require 'rest-client'
 require 'subtitle'
-require 'hpricot'
+require 'nokogiri'
 require 'iconv'
 require 'undertexter/array'
 
-module Hpricot
-
-  # Monkeypatch to fix an Hpricot bug that causes HTML entities to be decoded
-  # incorrectly.
-  def self.uxs(str)
-    str.to_s.
-      gsub(/&(\w+);/) { [Hpricot::NamedCharacters[$1] || ??].pack("U*") }.
-      gsub(/\&\#(\d+);/) { [$1.to_i].pack("U*") }
-  end
-
-end
-
 class Undertexter
   attr_accessor :raw_data, :base_details, :subtitles
   
@@ -57,43 +45,40 @@ class Undertexter
     # Example output
     # [["(1 cd)", "Nedladdningar: 11891", "Avatar (2009) PROPER DVDSCR XviD-MAXSPEED", "http://www.undertexter.se/?p=undertext&id=19751", "Avatar"]]
     
-    doc = Hpricot(@raw_data)
+    doc = Nokogiri::HTML(@raw_data)
     @block = []
     
     # Trying to find the {tbody} that does not contain any tbody's
-    tbody = doc.search("tbody").to_a.reject do |inner, index|
-      not inner.inner_html.match(/Nedladdningar/i)
+    tbody = doc.css("tbody").to_a.reject do |inner, index|
+      not inner.content.match(/Nedladdningar/i)
     end.sort_by do |inner| 
-      inner.search('tbody').count 
+      inner.css('tbody').count 
     end.first
     
     # Nothing found, okey!
     return if tbody.nil?
     
-    tbody = tbody.search('tr').drop(3)
+    tbody = tbody.css('tr').drop(3)
     
     tbody.each_with_index do |value, index|
       next unless index % 3 == 0
       length = @block.length
       @block[length] = [] if @block[length].nil?
-
-      line = tbody[index + 1].inner_html.split('<br />').map(&:strip)
-      value = value.search('a')
-
-      @block[length] << line[0]
-      @block[length] << line[2]
+      line = tbody[index + 1].content.split(/\n/).map(&:strip)
+      value = value.css('a')
+      @block[length] << line[1]
+      @block[length] << line[3]
       @block[length] << line[4]
-      @block[length] << value.last.attributes['href']
-      @block[length] << value.last.attributes['title']
-
-      @block[length].map! {|i| i.gsub(/<\/?[^>]*>/, "").strip}
+      @block[length] << value.last.attr('href')
+      @block[length] << value.last.attr('title')
+      @block[length].map!(&:strip)
     end
   end
   
   def build!
     @block.each do |movie|
       next unless movie.count == 5
-      @subtitles << Subtitle.new({
+      @subtitles << Container::Subtitle.new({
         :cds => movie[0].match(/\d+/)[0].to_i,
         :downloads => movie[1].match(/\d+$/)[0].to_i,
         :title => movie[2],

data/spec/subtitle_spec.rb

@@ -1,4 +1,6 @@
 require 'spec_helper'
+include Container
+
 describe Subtitle do
   before(:each) do
     @title = "127 Hours 2010 DVDSCR XViD-MC8"

data/spec/undertexter_spec.rb

@@ -84,6 +84,10 @@ describe Undertexter, "trying to search for a movie using a title" do
     end
   end
   
+  it "should only contain Container::Subtitle instances" do
+    @use.each { |subtitle| subtitle.should be_instance_of(Container::Subtitle) }
+  end
+  
   it "should not contain any attributes that contain any html tags" do
     @use.each do |subtitle|
       [:details, :downloads, :cds, :title, :movie_title, :url].each do |method|

data/undertexter.gemspec

@@ -3,7 +3,7 @@ $:.push File.expand_path("../lib", __FILE__)
 
 Gem::Specification.new do |s|
   s.name        = "undertexter"
-  s.version     = "0.1.4"
+  s.version     = "0.1.5"
   s.platform    = Gem::Platform::RUBY
   s.authors     = ["Linus Oleander"]
   s.email       = ["linus@oleander.nu"]
@@ -18,12 +18,8 @@ Gem::Specification.new do |s|
   s.executables   = `git ls-files -- bin/*`.split("\n").map{ |f| File.basename(f) }
   s.require_paths = ["lib"]
   
-  s.add_dependency('rest-client')
-  
-  # Using hpricot 0.8.2 due to a bug
-  # https://github.com/hpricot/hpricot/issues/#issue/33
-  
-  s.add_dependency('hpricot', '0.8.2')
+  s.add_dependency('rest-client')  
+  s.add_dependency('nokogiri')
   s.add_dependency('mimer_plus')
   s.add_dependency('levenshteinish')
   s.add_development_dependency('rspec')

metadata

@@ -1,13 +1,8 @@
 --- !ruby/object:Gem::Specification 
 name: undertexter
 version: !ruby/object:Gem::Version 
-  hash: 19
   prerelease: 
-  segments: 
-  - 0
-  - 1
-  - 4
-  version: 0.1.4
+  version: 0.1.5
 platform: ruby
 authors: 
 - Linus Oleander
@@ -15,7 +10,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-02-06 00:00:00 +01:00
+date: 2011-02-14 00:00:00 +01:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -26,26 +21,18 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        hash: 3
-        segments: 
-        - 0
         version: "0"
   type: :runtime
   version_requirements: *id001
 - !ruby/object:Gem::Dependency 
-  name: hpricot
+  name: nokogiri
   prerelease: false
   requirement: &id002 !ruby/object:Gem::Requirement 
     none: false
     requirements: 
-    - - "="
+    - - ">="
       - !ruby/object:Gem::Version 
-        hash: 59
-        segments: 
-        - 0
-        - 8
-        - 2
-        version: 0.8.2
+        version: "0"
   type: :runtime
   version_requirements: *id002
 - !ruby/object:Gem::Dependency 
@@ -56,9 +43,6 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        hash: 3
-        segments: 
-        - 0
         version: "0"
   type: :runtime
   version_requirements: *id003
@@ -70,9 +54,6 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        hash: 3
-        segments: 
-        - 0
         version: "0"
   type: :runtime
   version_requirements: *id004
@@ -84,9 +65,6 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        hash: 3
-        segments: 
-        - 0
         version: "0"
   type: :development
   version_requirements: *id005
@@ -129,18 +107,12 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements: 
   - - ">="
     - !ruby/object:Gem::Version 
-      hash: 3
-      segments: 
-      - 0
       version: "0"
 required_rubygems_version: !ruby/object:Gem::Requirement 
   none: false
   requirements: 
   - - ">="
     - !ruby/object:Gem::Version 
-      hash: 3
-      segments: 
-      - 0
       version: "0"
 requirements: []