Updated from tika-app-1.4.jar to tika-app-1.9.jar

mrcsparker · Jun 30, 2015 · 170290a · 170290a
1 parent cd0d5ba
commit 170290a
Show file tree

Hide file tree

Showing 4 changed files with 20 additions and 20 deletions.
diff --git a/ext/tika-app-1.4.jar → ext/tika-app-1.9.jar b/ext/tika-app-1.4.jar → ext/tika-app-1.9.jar
diff --git a/lib/ruby_tika_app.rb b/lib/ruby_tika_app.rb
@@ -24,7 +24,7 @@ def initialize(document)
 
     java_cmd = 'java'
     java_args = '-server -Djava.awt.headless=true'
-    tika_path = "#{File.join(File.dirname(__FILE__))}/../ext/tika-app-1.4.jar"
+    tika_path = "#{File.join(File.dirname(__FILE__))}/../ext/tika-app-1.9.jar"
 
     @tika_cmd = "#{java_cmd} #{java_args} -jar '#{tika_path}'"
   end

diff --git a/ruby_tika_app.gemspec b/ruby_tika_app.gemspec
@@ -3,7 +3,7 @@ $:.push File.expand_path('../lib', __FILE__)
 
 Gem::Specification.new do |s|
   s.name        = 'ruby_tika_app'
-  s.version     = '1.4.0'
+  s.version     = '1.5.0'
   s.platform    = Gem::Platform::RUBY
   s.authors     = ['Chris Parker']
   s.email       = %w(mrcsparker@gmail.com)
@@ -23,7 +23,7 @@ Gem::Specification.new do |s|
   s.add_runtime_dependency('open4')
 
   s.add_development_dependency('rake')
-  s.add_development_dependency('rspec', '~> 2.13.0')
+  s.add_development_dependency('rspec', '~> 3.3.0')
   s.add_development_dependency('bundler', '>= 1.0.15')
   s.add_development_dependency('simplecov')
   s.add_development_dependency('json')

diff --git a/spec/ruby_tika_app_spec.rb b/spec/ruby_tika_app_spec.rb
@@ -16,14 +16,14 @@
       expect {
         rta = RubyTikaApp.new('No file')
         rta.to_xml
-      }.to raise_error
+      }.to raise_error(RuntimeError)
     end
   end
 
   describe '#to_xml' do
     it 'header' do
       rta = RubyTikaApp.new(@test_file)
-      rta.to_xml[0..37].should == "<?xml version=\"1.0\" encoding=\"UTF-8\"?>"
+      expect(rta.to_xml[0..37]).to eq("<?xml version=\"1.0\" encoding=\"UTF-8\"?>")
     end
 
     it 'middle' do
@@ -32,81 +32,81 @@
 
       xml_size = xml.size / 2
 
-      xml[xml_size..(xml_size + 100)].should == "ction IV). Besides,\nMHRW performs better in well connected graphs than in\nloosely connected graphs, a"
+      expect(xml[xml_size..(xml_size + 100)]).to eq("plicated nodes make the node distribution converge\nto uniform distribution. We do not need to conside")
     end
   end
 
   describe '#to_html' do
     it 'header' do
       rta = RubyTikaApp.new(@test_file)
-      rta.to_html[0..42].should == "<html xmlns=\"http://www.w3.org/1999/xhtml\">"
+      expect(rta.to_html[0..42]).to eq("<html xmlns=\"http://www.w3.org/1999/xhtml\">")
     end
 
     it 'middle' do
       rta = RubyTikaApp.new(@test_file)
-      rta.to_html[1000 ... 1100].should == "rceName\" content=\"graph sampling simplex - 11.pdf\"/>\n<meta name=\"Last-Save-Date\" content=\"2011-03-29"
+      expect(rta.to_html[1000 ... 1100]).to eq("on/pdf\"/>\n<meta name=\"X-Parsed-By\" content=\"org.apache.tika.parser.DefaultParser\"/>\n<meta name=\"X-Pa")
     end
   end
 
   describe '#to_json' do
     it 'header' do
       rta = RubyTikaApp.new(@test_file)
-      rta.to_json[0..42].should == "{ \"Application\":\"\\u0027Certified by IEEE PD"
+      expect(rta.to_json[0..42]).to eq("{\"Application\":\"\\u0027Certified by IEEE PDF")
     end
 
     it 'middle' do
       rta = RubyTikaApp.new(@test_file)
-      rta.to_json[100 ... 150].should == "h\":171510, \n\"Content-Type\":\"application/pdf\", \n\"Cr"
+      expect(rta.to_json[100 ... 150]).to eq("\"171510\",\"Content-Type\":\"application/pdf\",\"Creatio")
     end
   end
 
   describe '#to_text' do
     it 'header' do
       rta = RubyTikaApp.new(@test_file)
-      rta.to_text[0..42].should == "Understanding Graph Sampling Algorithms\nfor"
+      expect(rta.to_text[0..42]).to eq("Understanding Graph Sampling Algorithms\nfor")
     end
 
     it 'middle' do
       rta = RubyTikaApp.new(@test_file)
-      rta.to_text[100 ... 150].should == "n Zhang3, Tianyin Xu2\n\nLong Jin1, Pan Hui4, Beixin"
+      expect(rta.to_text[100 ... 150]).to eq("n Zhang3, Tianyin Xu2\n\nLong Jin1, Pan Hui4, Beixin")
     end
   end
 
   describe '#to_text_main' do
     it 'header' do
       rta = RubyTikaApp.new(@test_file)
-      rta.to_text_main[0..42].should == 'Understanding Graph Sampling Algorithms for'
+      expect(rta.to_text_main[0..42]).to eq('Understanding Graph Sampling Algorithms for')
     end
 
     it 'middle' do
       rta = RubyTikaApp.new(@test_file)
-      rta.to_text_main[100 ... 150].should == "n Zhang3, Tianyin Xu2\nLong Jin1, Pan Hui4, Beixing"
+      expect(rta.to_text_main[100 ... 150]).to eq("n Zhang3, Tianyin Xu2\nLong Jin1, Pan Hui4, Beixing")
     end
   end
 
   describe '#to_metadata' do
     it 'header' do
       rta = RubyTikaApp.new(@test_file)
-      rta.to_metadata[0..42].should == "Application: 'Certified by IEEE PDFeXpress "
+      expect(rta.to_metadata[0..42]).to eq("Application: 'Certified by IEEE PDFeXpress ")
     end
 
     it 'middle' do
       rta = RubyTikaApp.new(@test_file)
-      rta.to_metadata[100 ... 150].should == "Type: application/pdf\nCreation-Date: 2011-03-29T12"
+      expect(rta.to_metadata[100 ... 150]).to eq("Type: application/pdf\nCreation-Date: 2011-03-29T12")
     end
   end
 
   describe 'external URLs' do
     it 'should be able to parse an http url' do
       rta = RubyTikaApp.new('http://localhost:9299/cnn.com')
-      rta.to_text.should_not be_nil
-      rta.to_text.should eq(RubyTikaApp.new(@cnn_com_file).to_text)
+      expect(rta.to_text).to_not be_nil
+      expect(rta.to_text).to eq(RubyTikaApp.new(@cnn_com_file).to_text)
     end
 
     it 'should be able to parse another http url' do
       rta = RubyTikaApp.new('http://localhost:9299/news.ycombinator.com')
-      rta.to_text.should_not be_nil
-      rta.to_text.should eq(RubyTikaApp.new(@news_ycombinator_com_file).to_text)
+      expect(rta.to_text).to_not be_nil
+      expect(rta.to_text).to eq(RubyTikaApp.new(@news_ycombinator_com_file).to_text)
     end
   end