Merge pull request #2 from andrusha/master

Fixing errors on reading files with compound names

lib/ruby_tika_app.rb

@@ -16,14 +16,13 @@ def initialize status
   end
 
   def initialize(document)
-
-    @document = document
+    @document = "file://#{document}"
 
     java_cmd = 'java'
     java_args = '-server -Djava.awt.headless=true'
     tika_path = "#{File.join(File.dirname(__FILE__))}/../ext/tika-app-1.2.jar"
 
-    @tika_cmd = "#{java_cmd} #{java_args} -jar #{tika_path}"
+    @tika_cmd = "#{java_cmd} #{java_args} -jar '#{tika_path}'"
   end
 
   def to_xml
@@ -53,8 +52,7 @@ def to_metadata
   private
 
   def run_tika(option)
-
-    final_cmd = "#{@tika_cmd} #{option} #{@document}"
+    final_cmd = "#{@tika_cmd} #{option} '#{@document}'"
     result = []
 
 
@@ -63,7 +61,7 @@ def run_tika(option)
     stdout_result = stdout.read.strip
     stderr_result = stderr.read.strip
 
-    unless stderr_result.strip == "" then
+    unless strip_stderr(stderr_result).empty?
       raise(CommandFailedError.new(stderr_result),
             "execution failed with status #{stderr_result}: #{final_cmd}")
     end
@@ -75,4 +73,8 @@ def run_tika(option)
     stderr.close
   end
 
+  def strip_stderr(s)
+    s.gsub(/^(info|warn) - .*$/i, '').strip
+  end
+
 end

spec/ruby_tika_app_spec.rb

@@ -3,7 +3,7 @@
 describe RubyTikaApp do
 
   before(:each) do
-    @test_file = "#{File.join(File.dirname(__FILE__))}/docs/graph_sampling_simplex11.pdf"
+    @test_file = "#{File.join(File.dirname(__FILE__))}/docs/graph sampling simplex - 11.pdf"
   end
 
   describe "#to_xml" do
@@ -18,7 +18,7 @@
 
       xml_size = xml.size / 2
 
-      xml[xml_size..(xml_size + 100)].should == "(Section IV). Besides,\nMHRW performs better in well connected graphs than in\nloosely connected graphs"
+      xml[xml_size..(xml_size + 100)].should == "sides,\nMHRW performs better in well connected graphs than in\nloosely connected graphs, as it was orig"
     end
   end
 
@@ -30,7 +30,7 @@
 
     it "middle" do
       rta = RubyTikaApp.new(@test_file)
-      rta.to_html[1000 ... 1100].should == "rceName\" content=\"graph_sampling_simplex11.pdf\"/>\n<meta name=\"Last-Save-Date\" content=\"2011-03-29T13"
+      rta.to_html[1000 ... 1100].should == "rceName\" content=\"graph sampling simplex - 11.pdf\"/>\n<meta name=\"Last-Save-Date\" content=\"2011-03-29"
     end
   end
 
@@ -80,7 +80,6 @@
       rta = RubyTikaApp.new(@test_file)
       rta.to_metadata[100 ... 150].should == "Type: application/pdf\nCreation-Date: 2011-03-29T12"
     end
-
   end
 
 end