From 4f62f365efce1c18f8e035f491fc66f0c4cebc0e Mon Sep 17 00:00:00 2001
From: Andrew Korzhuev <wolfon@gmail.com>
Date: Tue, 25 Dec 2012 15:37:04 +0400
Subject: [PATCH 1/3] Fixing errors on reading files with compound names

---
 lib/ruby_tika_app.rb                                |   7 +++----
 ...implex11.pdf => graph sampling simplex - 11.pdf} | Bin
 spec/ruby_tika_app_spec.rb                          |   7 +++----
 3 files changed, 6 insertions(+), 8 deletions(-)
 rename spec/docs/{graph_sampling_simplex11.pdf => graph sampling simplex - 11.pdf} (100%)

diff --git a/lib/ruby_tika_app.rb b/lib/ruby_tika_app.rb
index 0ecea02..01aa5c5 100644
--- a/lib/ruby_tika_app.rb
+++ b/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
@@ -54,7 +53,7 @@ def to_metadata
 
   def run_tika(option)
 
-    final_cmd = "#{@tika_cmd} #{option} #{@document}"
+    final_cmd = "#{@tika_cmd} #{option} '#{@document}'"
     result = []
 
 
diff --git a/spec/docs/graph_sampling_simplex11.pdf b/spec/docs/graph sampling simplex - 11.pdf
similarity index 100%
rename from spec/docs/graph_sampling_simplex11.pdf
rename to spec/docs/graph sampling simplex - 11.pdf
diff --git a/spec/ruby_tika_app_spec.rb b/spec/ruby_tika_app_spec.rb
index f05e965..f252c7c 100644
--- a/spec/ruby_tika_app_spec.rb
+++ b/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

From 7b3f9bebc8de4e1ec18df84512bcc792f084061a Mon Sep 17 00:00:00 2001
From: Andrew Korzhuev <wolfon@gmail.com>
Date: Wed, 26 Dec 2012 20:52:36 +0400
Subject: [PATCH 2/3] Do not treat 'INFO - ' in stderr output as error

---
 lib/ruby_tika_app.rb | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/lib/ruby_tika_app.rb b/lib/ruby_tika_app.rb
index 01aa5c5..7fb8331 100644
--- a/lib/ruby_tika_app.rb
+++ b/lib/ruby_tika_app.rb
@@ -52,7 +52,6 @@ def to_metadata
   private
 
   def run_tika(option)
-
     final_cmd = "#{@tika_cmd} #{option} '#{@document}'"
     result = []
 
@@ -62,7 +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
@@ -74,4 +73,8 @@ def run_tika(option)
     stderr.close
   end
 
+  def strip_stderr(s)
+    s.gsub(/^info - .*$/i, '').strip
+  end
+
 end

From fad8c86cfdeabd2673b790d8eb0db93f0cf81a7f Mon Sep 17 00:00:00 2001
From: Andrew Korzhuev <wolfon@gmail.com>
Date: Fri, 4 Jan 2013 16:59:27 +0400
Subject: [PATCH 3/3] Ignore warnings along with infos

---
 lib/ruby_tika_app.rb | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lib/ruby_tika_app.rb b/lib/ruby_tika_app.rb
index 7fb8331..fae69f7 100644
--- a/lib/ruby_tika_app.rb
+++ b/lib/ruby_tika_app.rb
@@ -74,7 +74,7 @@ def run_tika(option)
   end
 
   def strip_stderr(s)
-    s.gsub(/^info - .*$/i, '').strip
+    s.gsub(/^(info|warn) - .*$/i, '').strip
   end
 
 end