undertexter 0.1.3 → 0.1.4

data/Gemfile.lock

@@ -1,10 +1,10 @@
 PATH
   remote: .
   specs:
-    undertexter (0.1.2)
+    undertexter (0.1.3)
+      hpricot (= 0.8.2)
       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)
@@ -32,9 +32,5 @@ PLATFORMS
   ruby
 
 DEPENDENCIES
-  levenshteinish
-  mimer_plus
-  nokogiri
-  rest-client
   rspec
   undertexter!

data/lib/undertexter.rb

@@ -2,10 +2,22 @@
 
 require 'rest-client'
 require 'subtitle'
-require 'nokogiri'
+require 'hpricot'
 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
   
@@ -42,37 +54,39 @@ class Undertexter
   end
   
   def parse!
-    noko = Nokogiri::HTML(@raw_data)
-    
     # Example output
-    # [["(1 cd)", "Nedladdningar: 11891", "Avatar (2009) PROPER DVDSCR XviD-MAXSPEED", "http://www.undertexter.se/?p=undertext&id=19751"]]
+    # [["(1 cd)", "Nedladdningar: 11891", "Avatar (2009) PROPER DVDSCR XviD-MAXSPEED", "http://www.undertexter.se/?p=undertext&id=19751", "Avatar"]]
+    
+    doc = Hpricot(@raw_data)
+    @block = []
     
-    [12,15].each do |id|
-      @block = noko.css("table:nth-child(#{id}) td").to_a.reject do |inner| 
-        inner.content.empty? or ! inner.content.match(/Nedladdningar/i) 
-      end.map do |inner| 
-        inner.content.split(/\n/).map do |i| 
-          i.gsub(/"/, "").strip
-        end
-      end
-      
-      next if @block.nil?
-      
-      noko.css("table:nth-child(#{id}) a").to_a.reject do |inner| 
-        details = inner.attr('href')
-        inner.content.empty? or details.nil? or ! details.match(/(p=undertext&id=\d+)|(p=subtitle&id=\d+)/i)
-      end.map do |y| 
-        [y.attr('href'), y.content.strip]
-      end.reject do |list|
-        list.last.empty?
-      end.each_with_index do |value, index|
-        @block[index] << value.first
-        @block[index] << value.last
-      end
+    # 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)
+    end.sort_by do |inner| 
+      inner.search('tbody').count 
+    end.first
     
-      @block.map!{|value| value.reject(&:empty?)}
-      
-      break if @block.any?
+    # Nothing found, okey!
+    return if tbody.nil?
+    
+    tbody = tbody.search('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]
+      @block[length] << line[4]
+      @block[length] << value.last.attributes['href']
+      @block[length] << value.last.attributes['title']
+
+      @block[length].map! {|i| i.gsub(/<\/?[^>]*>/, "").strip}
     end
   end
   

data/spec/undertexter_spec.rb

@@ -58,8 +58,8 @@ describe Undertexter, "trying to search for a movie using a title" do
     Undertexter.should have_at_least(41).find("die hard")
   end
   
-  it "should return the right title, again" do
-    @use.each{|subtitle| subtitle.title.should match(/die.*hard/i)}
+  it "should have 6 die hard movies that does not contain any title" do
+    @use.reject{|subtitle| ! subtitle.title.empty?}.count.should be(6)
   end
   
   it "should contain the right details, again" do
@@ -83,6 +83,14 @@ describe Undertexter, "trying to search for a movie using a title" do
       subtitle.url.match(/id=(\d+)/)[1].should eq(@use[index].details.match(/id=(\d+)/)[1])
     end
   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|
+        subtitle.send(method).to_s.should_not match(/<\/?[^>]*>/)
+      end
+    end
+  end
 end
 
 describe Undertexter, "should work when trying to fetch some english subtitles" do  

data/undertexter.gemspec

@@ -3,7 +3,7 @@ $:.push File.expand_path("../lib", __FILE__)
 
 Gem::Specification.new do |s|
   s.name        = "undertexter"
-  s.version     = "0.1.3"
+  s.version     = "0.1.4"
   s.platform    = Gem::Platform::RUBY
   s.authors     = ["Linus Oleander"]
   s.email       = ["linus@oleander.nu"]
@@ -19,7 +19,11 @@ Gem::Specification.new do |s|
   s.require_paths = ["lib"]
   
   s.add_dependency('rest-client')
-  s.add_dependency('nokogiri')
+  
+  # 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('mimer_plus')
   s.add_dependency('levenshteinish')
   s.add_development_dependency('rspec')

metadata

@@ -1,12 +1,13 @@
 --- !ruby/object:Gem::Specification 
 name: undertexter
 version: !ruby/object:Gem::Version 
-  prerelease: false
+  hash: 19
+  prerelease: 
   segments: 
   - 0
   - 1
-  - 3
-  version: 0.1.3
+  - 4
+  version: 0.1.4
 platform: ruby
 authors: 
 - Linus Oleander
@@ -14,7 +15,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-02-04 00:00:00 +01:00
+date: 2011-02-06 00:00:00 +01:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -25,22 +26,26 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
+        hash: 3
         segments: 
         - 0
         version: "0"
   type: :runtime
   version_requirements: *id001
 - !ruby/object:Gem::Dependency 
-  name: nokogiri
+  name: hpricot
   prerelease: false
   requirement: &id002 !ruby/object:Gem::Requirement 
     none: false
     requirements: 
-    - - ">="
+    - - "="
       - !ruby/object:Gem::Version 
+        hash: 59
         segments: 
         - 0
-        version: "0"
+        - 8
+        - 2
+        version: 0.8.2
   type: :runtime
   version_requirements: *id002
 - !ruby/object:Gem::Dependency 
@@ -51,6 +56,7 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
+        hash: 3
         segments: 
         - 0
         version: "0"
@@ -64,6 +70,7 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
+        hash: 3
         segments: 
         - 0
         version: "0"
@@ -77,6 +84,7 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
+        hash: 3
         segments: 
         - 0
         version: "0"
@@ -98,7 +106,6 @@ files:
 - Gemfile.lock
 - README.markdown
 - Rakefile
-- lib/.undertexter.rb.swp
 - lib/subtitle.rb
 - lib/undertexter.rb
 - lib/undertexter/array.rb
@@ -122,6 +129,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements: 
   - - ">="
     - !ruby/object:Gem::Version 
+      hash: 3
       segments: 
       - 0
       version: "0"
@@ -130,13 +138,14 @@ required_rubygems_version: !ruby/object:Gem::Requirement
   requirements: 
   - - ">="
     - !ruby/object:Gem::Version 
+      hash: 3
       segments: 
       - 0
       version: "0"
 requirements: []
 
 rubyforge_project: undertexter
-rubygems_version: 1.3.7
+rubygems_version: 1.5.0
 signing_key: 
 specification_version: 3
 summary: A subtitle search client for undertexter.se