hydra-file_characterization 0.3.2 → 1.1.2

data/lib/hydra/file_characterization/characterizers/ffprobe.rb

@@ -1,13 +1,13 @@
+# frozen_string_literal: true
 require 'hydra/file_characterization/exceptions'
 require 'hydra/file_characterization/characterizer'
 
 module Hydra::FileCharacterization::Characterizers
   class Ffprobe < Hydra::FileCharacterization::Characterizer
-
     protected
+
     def command
       "#{tool_path} -i \"#{filename}\" -print_format xml -show_streams -v quiet"
     end
-
   end
 end

data/lib/hydra/file_characterization/characterizers/fits.rb

@@ -1,18 +1,23 @@
+# frozen_string_literal: true
 require 'hydra/file_characterization/exceptions'
 require 'hydra/file_characterization/characterizer'
+require 'logger'
 module Hydra::FileCharacterization::Characterizers
   class Fits < Hydra::FileCharacterization::Characterizer
-
     protected
 
-      def command
-        "#{tool_path} -i \"#{filename}\""
-      end
+    def command
+      "#{tool_path} -i \"#{filename}\""
+    end
 
-      # Remove any residual non-XML from JHOVE
-      # See: https://github.com/harvard-lts/fits/issues/20
-      def post_process(raw_output)
-        raw_output.sub(/^READBOX seen=true\n/, '')
-      end
+    # Remove any non-XML output that precedes the <?xml> tag
+    # See: https://github.com/harvard-lts/fits/issues/20
+    #      https://github.com/harvard-lts/fits/issues/40
+    #      https://github.com/harvard-lts/fits/issues/46
+    def post_process(raw_output)
+      md = /\A(.*)(<\?xml.*)\Z/m.match(raw_output)
+      logger.warn "FITS produced non-xml output: \"#{md[1].chomp}\"" unless md[1].empty?
+      md[2]
+    end
   end
 end

data/lib/hydra/file_characterization/characterizers/fits_servlet.rb

@@ -0,0 +1,23 @@
+# frozen_string_literal: true
+require 'hydra/file_characterization/exceptions'
+require 'hydra/file_characterization/characterizer'
+require 'logger'
+module Hydra::FileCharacterization::Characterizers
+  class FitsServlet < Hydra::FileCharacterization::Characterizer
+    protected
+
+    def command
+      "curl -k -F datafile=@#{filename} #{ENV['FITS_SERVLET_URL']}/examine"
+    end
+
+    # Remove any non-XML output that precedes the <?xml> tag
+    # See: https://github.com/harvard-lts/fits/issues/20
+    #      https://github.com/harvard-lts/fits/issues/40
+    #      https://github.com/harvard-lts/fits/issues/46
+    def post_process(raw_output)
+      md = /\A(.*)(<\?xml.*)\Z/m.match(raw_output)
+      logger.warn "FITS produced non-xml output: \"#{md[1].chomp}\"" unless md[1].empty?
+      md[2]
+    end
+  end
+end

data/lib/hydra/file_characterization/exceptions.rb

@@ -1,5 +1,5 @@
+# frozen_string_literal: true
 module Hydra::FileCharacterization
-
   class FileNotFoundError < RuntimeError
   end
 
@@ -8,5 +8,4 @@ module Hydra::FileCharacterization
       super("Unable to find Hydra::FileCharacterization tool with name :#{tool_name}")
     end
   end
-
 end

data/lib/hydra/file_characterization/to_temp_file.rb

@@ -1,3 +1,4 @@
+# frozen_string_literal: true
 require 'open3'
 require 'tempfile'
 
@@ -16,7 +17,7 @@ module Hydra::FileCharacterization
     end
 
     def call(data)
-      f = Tempfile.new([File.basename(filename),File.extname(filename)])
+      f = Tempfile.new([File.basename(filename), File.extname(filename)])
       begin
         f.binmode
         if data.respond_to? :read
@@ -32,6 +33,5 @@ module Hydra::FileCharacterization
         f.unlink
       end
     end
-
   end
-end
+end

data/lib/hydra/file_characterization/version.rb

@@ -1,5 +1,6 @@
+# frozen_string_literal: true
 module Hydra
   module FileCharacterization
-    VERSION = "0.3.2"
+    VERSION = "1.1.2"
   end
 end

data/spec/lib/hydra/file_characterization/characterizer_spec.rb

@@ -1,30 +1,31 @@
+# frozen_string_literal: true
 require 'spec_helper'
 
 module Hydra::FileCharacterization
   describe Characterizer do
+    subject { characterizer }
     let(:filename) { __FILE__ }
     let(:instance_tool_path) { nil }
     let(:class_tool_path) { nil }
 
-    let(:characterizer) { Hydra::FileCharacterization::Characterizer.new(filename, instance_tool_path) }
-    subject { characterizer }
-    around(:each) do |example|
-      Hydra::FileCharacterization::Characterizer.tool_path = class_tool_path
+    let(:characterizer) { described_class.new(filename, instance_tool_path) }
+    around do |example|
+      described_class.tool_path = class_tool_path
       example.run
-      Hydra::FileCharacterization::Characterizer.tool_path = nil
+      described_class.tool_path = nil
     end
 
     context 'call' do
       context 'with missing file' do
         let(:filename) { '/dev/path/to/bogus/file' }
-        it 'should raise FileNotFoundError' do
+        it 'raises FileNotFoundError' do
           expect { subject.call }.to raise_error(FileNotFoundError)
         end
       end
 
       context 'with a callable tool path' do
-        let(:class_tool_path) { lambda { |filename| [filename, :output] }}
-        it 'should raise FileNotFoundError' do
+        let(:class_tool_path) { ->(filename) { [filename, :output] } }
+        it 'raises FileNotFoundError' do
           expect(subject.call).to eq [filename, :output]
         end
       end

data/spec/lib/hydra/file_characterization/characterizers/ffprobe_spec.rb

@@ -1,19 +1,16 @@
+# frozen_string_literal: true
 require 'spec_helper'
 require 'hydra/file_characterization/characterizers/ffprobe'
 
 module Hydra::FileCharacterization::Characterizers
-
   describe Ffprobe do
-
-    subject { Ffprobe.new(filename) }
+    subject { described_class.new(filename) }
 
     describe 'invalidFile' do
       let(:filename) { fixture_file('nofile.pdf') }
-      it "should raise an error if the path does not contain the file" do
-        expect {subject.call}.to raise_error(Hydra::FileCharacterization::FileNotFoundError)
+      it "raises an error if the path does not contain the file" do
+        expect { subject.call }.to raise_error(Hydra::FileCharacterization::FileNotFoundError)
       end
     end
-
   end
-
 end

data/spec/lib/hydra/file_characterization/characterizers/fit_servlet_spec.rb

@@ -0,0 +1,71 @@
+# frozen_string_literal: true
+require 'spec_helper'
+require 'hydra/file_characterization/characterizers/fits_servlet'
+
+module Hydra::FileCharacterization::Characterizers
+  describe FitsServlet do
+    let(:fits) { Fits.new(filename) }
+
+    describe "#call", unless: ENV['TRAVIS'] do
+      subject { fits.call }
+
+      context 'validfile' do
+        let(:filename) { fixture_file('brendan_behan.jpeg') }
+        it { is_expected.to include(%(<identity format="JPEG File Interchange Format" mimetype="image/jpeg")) }
+      end
+
+      context 'invalidFile' do
+        let(:filename) { fixture_file('nofile.pdf') }
+        it "raises an error" do
+          expect { subject }.to raise_error(Hydra::FileCharacterization::FileNotFoundError)
+        end
+      end
+
+      context 'corruptFile' do
+        let(:filename) { fixture_file('brendan_broken.dxxd') }
+        it { is_expected.to include(%(<identity format="Unknown Binary" mimetype="application/octet-stream")) }
+      end
+
+      context 'zip file should be characterized not its contents' do
+        let(:filename) { fixture_file('archive.zip') }
+        it { is_expected.to include(%(<identity format="ZIP Format" mimetype="application/zip")) }
+      end
+    end
+
+    context 'when JHOVE adds non-xml' do
+      # https://github.com/harvard-lts/fits/issues/20
+      subject { fits.call }
+
+      before do
+        expect(fits.logger).to receive(:warn)
+        allow(fits).to receive(:internal_call).and_return(
+          'READBOX seen=true
+<?xml version="1.0" encoding="UTF-8"?>
+<fits xmlns="http://hul.harvard.edu/ois/xml/ns/fits/fits_output" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://hul.harvard.edu/ois/xml/ns/fits/fits_output http://hul.harvard.edu/ois/xml/xsd/fits/fits_output.xsd" version="0.8.2" timestamp="15/09/14 10:00 AM">
+<identification/></fits>'
+        )
+      end
+
+      let(:filename) { fixture_file('brendan_behan.jpeg') }
+      it { is_expected.not_to include('READBOX') }
+    end
+
+    context "when FITS itself adds non-xml" do
+      # https://github.com/harvard-lts/fits/issues/46
+      subject { fits.call }
+
+      before do
+        expect(fits.logger).to receive(:warn)
+        allow(fits).to receive(:internal_call).and_return(
+          '2015-10-15 17:14:25,761 ERROR [main] ToolBelt:79 - Thread 1 error initializing edu.harvard.hul.ois.fits.tools.droid.Droid: edu.harvard.hul.ois.fits.exceptions.FitsToolException  Message: DROID cannot run under Java 8
+<?xml version="1.0" encoding="UTF-8"?>
+<fits xmlns="http://hul.harvard.edu/ois/xml/ns/fits/fits_output" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://hul.harvard.edu/ois/xml/ns/fits/fits_output http://hul.harvard.edu/ois/xml/xsd/fits/fits_output.xsd" version="0.8.2" timestamp="15/09/14 10:00 AM">
+<identification/></fits>'
+        )
+      end
+
+      let(:filename) { fixture_file('brendan_behan.jpeg') }
+      it { is_expected.not_to include('FitsToolException') }
+    end
+  end
+end

data/spec/lib/hydra/file_characterization/characterizers/fits_spec.rb

@@ -1,11 +1,12 @@
+# frozen_string_literal: true
 require 'spec_helper'
 require 'hydra/file_characterization/characterizers/fits'
 
 module Hydra::FileCharacterization::Characterizers
   describe Fits do
-    let(:fits) { Fits.new(filename) }
+    let(:fits) { described_class.new(filename) }
 
-    describe "#call" do
+    describe "#call", unless: ENV['TRAVIS'] do
       subject { fits.call }
 
       context 'validfile' do
@@ -27,22 +28,44 @@ module Hydra::FileCharacterization::Characterizers
 
       context 'zip file should be characterized not its contents' do
         let(:filename) { fixture_file('archive.zip') }
-        it { is_expected.to include(%(<identity format="ZIP Format" mimetype="application/zip"))}
+        it { is_expected.to include(%(<identity format="ZIP Format" mimetype="application/zip")) }
       end
+    end
+
+    context 'when JHOVE adds non-xml' do
+      # https://github.com/harvard-lts/fits/issues/20
+      subject { fits.call }
 
-      context 'when JHOVE adds non-xml' do
-        # https://github.com/harvard-lts/fits/issues/20
-        before do
-          allow(fits).to receive(:internal_call).and_return(
-            'READBOX seen=true
+      before do
+        expect(fits.logger).to receive(:warn)
+        allow(fits).to receive(:internal_call).and_return(
+          'READBOX seen=true
 <?xml version="1.0" encoding="UTF-8"?>
 <fits xmlns="http://hul.harvard.edu/ois/xml/ns/fits/fits_output" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://hul.harvard.edu/ois/xml/ns/fits/fits_output http://hul.harvard.edu/ois/xml/xsd/fits/fits_output.xsd" version="0.8.2" timestamp="15/09/14 10:00 AM">
-  <identification/></fits>')
-        end
+<identification/></fits>'
+        )
+      end
 
-        let(:filename) { fixture_file('brendan_behan.jpeg') }
-        it { is_expected.not_to include('READBOX') }
+      let(:filename) { fixture_file('brendan_behan.jpeg') }
+      it { is_expected.not_to include('READBOX') }
+    end
+
+    context "when FITS itself adds non-xml" do
+      # https://github.com/harvard-lts/fits/issues/46
+      subject { fits.call }
+
+      before do
+        expect(fits.logger).to receive(:warn)
+        allow(fits).to receive(:internal_call).and_return(
+          '2015-10-15 17:14:25,761 ERROR [main] ToolBelt:79 - Thread 1 error initializing edu.harvard.hul.ois.fits.tools.droid.Droid: edu.harvard.hul.ois.fits.exceptions.FitsToolException  Message: DROID cannot run under Java 8
+<?xml version="1.0" encoding="UTF-8"?>
+<fits xmlns="http://hul.harvard.edu/ois/xml/ns/fits/fits_output" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://hul.harvard.edu/ois/xml/ns/fits/fits_output http://hul.harvard.edu/ois/xml/xsd/fits/fits_output.xsd" version="0.8.2" timestamp="15/09/14 10:00 AM">
+<identification/></fits>'
+        )
       end
+
+      let(:filename) { fixture_file('brendan_behan.jpeg') }
+      it { is_expected.not_to include('FitsToolException') }
     end
   end
 end

data/spec/lib/hydra/file_characterization/characterizers_spec.rb

@@ -1,3 +1,4 @@
+# frozen_string_literal: true
 require 'spec_helper'
 require 'hydra/file_characterization/characterizers'
 
@@ -7,36 +8,35 @@ module Hydra::FileCharacterization
 
     describe 'with :fits tool_name' do
       let(:tool_name) { :fits }
-      it { should eq(Characterizers::Fits) }
+      it { is_expected.to eq(Characterizers::Fits) }
     end
 
     describe 'with :ffprobe tool_name' do
       let(:tool_name) { :ffprobe }
-      it { should eq(Characterizers::Ffprobe) }
+      it { is_expected.to eq(Characterizers::Ffprobe) }
     end
 
     context '.characterize_with' do
+      subject { Hydra::FileCharacterization.characterize_with(tool_name, filename, tool_path) }
       let(:tool_name) { :fits }
       let(:filename) { __FILE__ }
       let(:tool_path) { nil }
-      subject { Hydra::FileCharacterization.characterize_with(tool_name, filename, tool_path) }
 
       context 'with callable tool_path and missing tool name' do
-        let(:tool_path) { lambda {|filename| [filename, :tool_path]} }
+        let(:tool_path) { ->(filename) { [filename, :tool_path] } }
         let(:tool_name) { :chunky_salsa }
-        it { should eq [filename, :tool_path] }
+        it { is_expected.to eq [filename, :tool_path] }
       end
 
       context 'with missing tool name and non-callable tool_path' do
         let(:tool_name) { :chunky_salsa }
         let(:tool_path) { '/path' }
-        it 'should raise exception' do
-          expect {
+        it 'raises exception' do
+          expect do
             subject
-          }.to raise_error(ToolNotFoundError)
+          end.to raise_error(ToolNotFoundError)
         end
       end
     end
-
   end
-end
+end

data/spec/lib/hydra/file_characterization/to_temp_file_spec.rb

@@ -1,10 +1,9 @@
+# frozen_string_literal: true
 require 'spec_helper'
 require 'hydra/file_characterization/to_temp_file'
 
 module Hydra::FileCharacterization
-
   describe 'ToTempFile' do
-
     let(:content) { "This is the content of the file." }
     let(:filename) { "hello.rb" }
 
@@ -44,6 +43,5 @@ module Hydra::FileCharacterization
         end
       end
     end
-
   end
-end
+end

data/spec/lib/hydra/file_characterization_spec.rb

@@ -6,7 +6,7 @@ module Hydra
 
   describe FileCharacterization do
 
-    describe '.characterize' do
+    describe '.characterize', unless: ENV['TRAVIS'] do
       describe "for content in memory" do
         let(:content) { "class Test; end\n" }
         let(:filename) { 'test.rb' }
@@ -18,10 +18,15 @@ module Hydra
         end
 
         describe 'with configured path' do
+          let(:tool_path) do
+            `which fits || which fits.sh`.strip
+          end
+
           it {
             response = Hydra::FileCharacterization.characterize(content, filename, :fits) do |config|
-              config[:fits] = `which fits || which fits.sh`.strip
+              config[:fits] = tool_path
             end
+
             expect(response).to match(/#{'<identity format="Plain text" mimetype="text/plain"'}/)
           }
         end
@@ -89,7 +94,7 @@ module Hydra
         Hydra::FileCharacterization::Characterizers::Fits.tool_path = old_tool_path
       end
 
-      it 'without configuration' do
+      it 'without configuration', unless: ENV['CI'] do
         Hydra::FileCharacterization.configure do |config|
           config.tool_path(:fits, nil)
         end

data/spec/spec_helper.rb

@@ -1,3 +1,4 @@
+# frozen_string_literal: true
 # This file was generated by the `rspec --init` command. Conventionally, all
 # specs live under a `spec` directory, which RSpec adds to the `$LOAD_PATH`.
 # Require this file using `require "spec_helper"` to ensure that it is only
@@ -6,7 +7,10 @@
 # See http://rubydoc.info/gems/rspec-core/RSpec/Core/Configuration
 
 GEM_ROOT = File.expand_path("../../", __FILE__)
-$:.unshift File.join(GEM_ROOT, "lib")
+$LOAD_PATH.unshift File.join(GEM_ROOT, "lib")
+
+require 'coveralls'
+Coveralls.wear!
 
 require 'hydra/file_characterization'
 
@@ -27,6 +31,7 @@ RSpec.configure do |config|
   #     --seed 1234
   config.order = 'random'
   config.before(:suite) do
-    Hydra::FileCharacterization::Characterizers::Fits.tool_path = `which fits || which fits.sh`.strip
+    tool_path = `which fits || which fits.sh`.strip
+    Hydra::FileCharacterization::Characterizers::Fits.tool_path = tool_path
   end
 end