libis-format 0.9.38 → 0.9.39

data/spec/identifier_spec.rb

@@ -69,8 +69,6 @@ describe 'Identfier' do
     ::Libis::Tools::Config.logger.appenders =
         ::Logging::Appenders.string_io('StringIO', layout: ::Libis::Tools::Config.get_log_formatter)
     ::Libis::Tools::Config.logger.level = :all
-    ::Libis::Format::Config[:droid_path] = '/opt/droid/droid.sh'
-    ::Libis::Format::Config[:fido_path] = File.join(ENV['HOME'], 'bin', 'fido')
   end
 
   let (:identifier) {::Libis::Format::Identifier}
@@ -106,53 +104,46 @@ describe 'Identfier' do
     end).to be_truthy
   end
 
-  it 'should identify all files in a folder at once' do
-    result = identifier.get (dir)
-    expect(result[:formats].size).to be >= formatlist.size
-    formatlist.each do |file, format|
-      expect(result[:formats][File.join(dir, file)]).to include format
-    end
-  end
+  context 'Fido and Droid', if: (File.exists?(Libis::Format::Config[:droid_path]) &&
+      File.exists?(Libis::Format::Config[:fido_path])) do
 
-  it 'should identify all files in a folder with base_dir option' do
-    result = identifier.get(dir, base_dir: dir)
-    expect(result[:formats].size).to be >= formatlist.size
-    formatlist.each do |file, format|
-      expect(result[:formats][file]).to include format
+    it 'should identify all files in a folder at once' do
+      result = identifier.get (dir)
+      expect(result[:formats].size).to be >= formatlist.size
+      formatlist.each do |file, format|
+        expect(result[:formats][File.join(dir, file)]).to include format
+      end
     end
-  end
 
-  it 'should identify all files in a list at once' do
-    filelist = fidolist.keys.map {|file| File.join(dir, file)}
-    result = identifier.get (filelist)
-    expect(result[:formats].size).to be >= formatlist.size
-    formatlist.each do |file, format|
-      expect(result[:formats][File.join(dir, file)]).to include format
+    it 'should identify all files in a folder with base_dir option' do
+      result = identifier.get(dir, base_dir: dir)
+      expect(result[:formats].size).to be >= formatlist.size
+      formatlist.each do |file, format|
+        expect(result[:formats][file]).to include format
+      end
     end
-  end
 
-  it 'should identify all files in a list with base_dir option' do
-    filelist = fidolist.keys.map {|file| File.join(dir, file)}
-    result = identifier.get(filelist, base_dir: dir)
-    expect(result[:formats].size).to be >= formatlist.size
-    formatlist.each do |file, format|
-      expect(result[:formats][file]).to include format
+    it 'should identify all files in a list at once' do
+      filelist = fidolist.keys.map {|file| File.join(dir, file)}
+      result = identifier.get (filelist)
+      expect(result[:formats].size).to be >= formatlist.size
+      formatlist.each do |file, format|
+        expect(result[:formats][File.join(dir, file)]).to include format
+      end
     end
-  end
-
-  context 'individual files' do
 
-    formatlist.each do |file, format|
-      it "should identify test document '#{file}'" do
-        result = identifier.get (File.join(dir, file))
-        expect(result[:formats].size).to be 1
-        expect(result[:formats][File.join(dir, file)]).to include format
+    it 'should identify all files in a list with base_dir option' do
+      filelist = fidolist.keys.map {|file| File.join(dir, file)}
+      result = identifier.get(filelist, base_dir: dir)
+      expect(result[:formats].size).to be >= formatlist.size
+      formatlist.each do |file, format|
+        expect(result[:formats][file]).to include format
       end
     end
 
   end
 
-  context 'Fido' do
+  context 'Fido', if: File.exists?(Libis::Format::Config[:fido_path]) do
 
     # expect(identifier.fido_formats.size).to be 1
     # expect(File.basename(identifier.fido_formats.first)).to eq 'lias_formats.xml'
@@ -169,7 +160,7 @@ describe 'Identfier' do
 
     it 'should identify dir of test documents' do
       filelist = fidolist.keys.map {|file| File.join(dir, file)}
-      fido_result = ::Libis::Format::Fido.instance.run_dir(dir)
+      fido_result = ::Libis::Format::Fido.instance.run_dir(dir, false)
       filelist.each do |filename|
         result = fido_result[filename]
         result = result[0] if result
@@ -179,5 +170,4 @@ describe 'Identfier' do
     end
 
   end
-
 end

data/spec/spec_helper.rb

@@ -6,4 +6,53 @@ Bundler.setup
 
 require 'rspec'
 require 'libis-format'
-require 'libis-tools'
+require 'libis-tools'
+
+require 'chromaprint'
+
+RSpec::Matchers.define(:be_same_file_as) do |exected_file_path|
+  match do |actual_file_path|
+    expect(md5_hash(actual_file_path)).to eq md5_hash(exected_file_path)
+  end
+
+  def md5_hash(file_path)
+    Digest::MD5.hexdigest(File.read(file_path))
+  end
+end
+
+RSpec::Matchers.define :sound_like do |exp_file, threshold, rate, channels|
+  match do |tgt_file|
+    rate ||= 96000
+    channels ||= 1
+    threshold ||= 0.9
+
+    if File.exists?(exp_file) && File.exists?(tgt_file)
+      # Convert input files into raw 16-bit signed audio (WAV) to process with Chramaprint
+      exp_raw = File.join('','tmp', File.basename(exp_file) + '.wav')
+      tgt_raw = File.join('','tmp', File.basename(tgt_file) + '.wav')
+      FileUtils.rm(exp_raw, force: true)
+      FileUtils.rm(tgt_raw, force: true)
+      cvt_cmd = "sox %s -e signed -b 16 -t wav %s rate #{rate} channels #{channels}"
+      %x"#{cvt_cmd % [exp_file,exp_raw]}"
+      %x"#{cvt_cmd % [tgt_file,tgt_raw]}"
+      exp_audio = File.binread(exp_raw)
+      tgt_audio = File.binread(tgt_raw)
+
+      # Get audio fingerprints
+      chromaprint = Chromaprint::Context.new(rate, channels, Chromaprint::ALGORITHM_TEST3)
+      exp_fp = chromaprint.get_fingerprint(exp_audio)
+      tgt_fp = chromaprint.get_fingerprint(tgt_audio)
+
+      # Cleanup files
+      FileUtils.rm(exp_raw, force: true)
+      FileUtils.rm(tgt_raw, force: true)
+
+      # Compare fingerprints and compare result against threshold
+      cmp = exp_fp.compare(tgt_fp)
+      # puts "Threshold[#{File.basename(exp_file)},#{File.basename(tgt_file)}: #{cmp}"
+      cmp > threshold
+    else
+      false
+    end
+  end
+end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: libis-format
 version: !ruby/object:Gem::Version
-  version: 0.9.38
+  version: 0.9.39
 platform: ruby
 authors:
 - Kris Dekeyser
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2017-10-03 00:00:00.000000000 Z
+date: 2017-11-15 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -136,6 +136,20 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '0.3'
+- !ruby/object:Gem::Dependency
+  name: chromaprint
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 0.0.2
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 0.0.2
 description: Collection of tools and classes that help to identify formats of binary
   files and create derivative copies (e.g. PDF from Word).
 email:
@@ -168,6 +182,7 @@ files:
 - lib/libis/format.rb
 - lib/libis/format/config.rb
 - lib/libis/format/converter.rb
+- lib/libis/format/converter/audio_converter.rb
 - lib/libis/format/converter/base.rb
 - lib/libis/format/converter/chain.rb
 - lib/libis/format/converter/image_converter.rb
@@ -175,8 +190,10 @@ files:
 - lib/libis/format/converter/office_converter.rb
 - lib/libis/format/converter/pdf_converter.rb
 - lib/libis/format/converter/repository.rb
+- lib/libis/format/converter/video_converter.rb
 - lib/libis/format/droid.rb
 - lib/libis/format/extension_identification.rb
+- lib/libis/format/ffmpeg.rb
 - lib/libis/format/fido.rb
 - lib/libis/format/file_tool.rb
 - lib/libis/format/identification_tool.rb
@@ -197,6 +214,58 @@ files:
 - spec/data/Cevennes2.ppm
 - spec/data/NikonRaw-CameraRaw.TIF
 - spec/data/NikonRaw-CaptureOne.tif
+- spec/data/audio/carlin_disappointed.aac
+- spec/data/audio/carlin_disappointed.ac3
+- spec/data/audio/carlin_disappointed.aiff
+- spec/data/audio/carlin_disappointed.au
+- spec/data/audio/carlin_disappointed.flac
+- spec/data/audio/carlin_disappointed.m4a
+- spec/data/audio/carlin_disappointed.mka
+- spec/data/audio/carlin_disappointed.mp3
+- spec/data/audio/carlin_disappointed.ogg
+- spec/data/audio/carlin_disappointed.ra
+- spec/data/audio/carlin_disappointed.voc
+- spec/data/audio/carlin_disappointed.wav
+- spec/data/audio/carlin_disappointed.wma
+- spec/data/audio/greensleeves.aac
+- spec/data/audio/greensleeves.ac3
+- spec/data/audio/greensleeves.aiff
+- spec/data/audio/greensleeves.au
+- spec/data/audio/greensleeves.flac
+- spec/data/audio/greensleeves.m4a
+- spec/data/audio/greensleeves.mka
+- spec/data/audio/greensleeves.mp3
+- spec/data/audio/greensleeves.ogg
+- spec/data/audio/greensleeves.ra
+- spec/data/audio/greensleeves.voc
+- spec/data/audio/greensleeves.wav
+- spec/data/audio/greensleeves.wma
+- spec/data/audio/king_nonviolence.aac
+- spec/data/audio/king_nonviolence.ac3
+- spec/data/audio/king_nonviolence.aiff
+- spec/data/audio/king_nonviolence.au
+- spec/data/audio/king_nonviolence.flac
+- spec/data/audio/king_nonviolence.m4a
+- spec/data/audio/king_nonviolence.mka
+- spec/data/audio/king_nonviolence.mp3
+- spec/data/audio/king_nonviolence.ogg
+- spec/data/audio/king_nonviolence.ra
+- spec/data/audio/king_nonviolence.voc
+- spec/data/audio/king_nonviolence.wav
+- spec/data/audio/king_nonviolence.wma
+- spec/data/audio/top_gun_secret.aac
+- spec/data/audio/top_gun_secret.ac3
+- spec/data/audio/top_gun_secret.aiff
+- spec/data/audio/top_gun_secret.au
+- spec/data/audio/top_gun_secret.flac
+- spec/data/audio/top_gun_secret.m4a
+- spec/data/audio/top_gun_secret.mka
+- spec/data/audio/top_gun_secret.mp3
+- spec/data/audio/top_gun_secret.ogg
+- spec/data/audio/top_gun_secret.ra
+- spec/data/audio/top_gun_secret.voc
+- spec/data/audio/top_gun_secret.wav
+- spec/data/audio/top_gun_secret.wma
 - spec/data/multipage.tif
 - spec/data/multipage.tif.jp2
 - spec/data/test-ead.xml
@@ -224,6 +293,23 @@ files:
 - spec/data/test.xlsx
 - spec/data/test.xml
 - spec/data/test_pdfa.pdf
+- spec/data/video/SampleVideo_1080x720_2mb.flv
+- spec/data/video/SampleVideo_1080x720_2mb.mkv
+- spec/data/video/SampleVideo_1080x720_2mb.mp4
+- spec/data/video/SampleVideo_176x144_2mb.3gp
+- spec/data/video/SampleVideo_320x240_2mb.3gp
+- spec/data/video/SampleVideo_360x240_2mb.flv
+- spec/data/video/SampleVideo_360x240_2mb.mkv
+- spec/data/video/SampleVideo_360x240_2mb.mp4
+- spec/data/video/sample.avi
+- spec/data/video/sample.flv
+- spec/data/video/sample.mkv
+- spec/data/video/sample.mov
+- spec/data/video/sample.mp4
+- spec/data/video/sample.mpg
+- spec/data/video/sample.swf
+- spec/data/video/sample.webm
+- spec/data/video/sample.wmv
 - spec/identifier_spec.rb
 - spec/spec_helper.rb
 - spec/test_types.yml
@@ -264,6 +350,58 @@ test_files:
 - spec/data/Cevennes2.ppm
 - spec/data/NikonRaw-CameraRaw.TIF
 - spec/data/NikonRaw-CaptureOne.tif
+- spec/data/audio/carlin_disappointed.aac
+- spec/data/audio/carlin_disappointed.ac3
+- spec/data/audio/carlin_disappointed.aiff
+- spec/data/audio/carlin_disappointed.au
+- spec/data/audio/carlin_disappointed.flac
+- spec/data/audio/carlin_disappointed.m4a
+- spec/data/audio/carlin_disappointed.mka
+- spec/data/audio/carlin_disappointed.mp3
+- spec/data/audio/carlin_disappointed.ogg
+- spec/data/audio/carlin_disappointed.ra
+- spec/data/audio/carlin_disappointed.voc
+- spec/data/audio/carlin_disappointed.wav
+- spec/data/audio/carlin_disappointed.wma
+- spec/data/audio/greensleeves.aac
+- spec/data/audio/greensleeves.ac3
+- spec/data/audio/greensleeves.aiff
+- spec/data/audio/greensleeves.au
+- spec/data/audio/greensleeves.flac
+- spec/data/audio/greensleeves.m4a
+- spec/data/audio/greensleeves.mka
+- spec/data/audio/greensleeves.mp3
+- spec/data/audio/greensleeves.ogg
+- spec/data/audio/greensleeves.ra
+- spec/data/audio/greensleeves.voc
+- spec/data/audio/greensleeves.wav
+- spec/data/audio/greensleeves.wma
+- spec/data/audio/king_nonviolence.aac
+- spec/data/audio/king_nonviolence.ac3
+- spec/data/audio/king_nonviolence.aiff
+- spec/data/audio/king_nonviolence.au
+- spec/data/audio/king_nonviolence.flac
+- spec/data/audio/king_nonviolence.m4a
+- spec/data/audio/king_nonviolence.mka
+- spec/data/audio/king_nonviolence.mp3
+- spec/data/audio/king_nonviolence.ogg
+- spec/data/audio/king_nonviolence.ra
+- spec/data/audio/king_nonviolence.voc
+- spec/data/audio/king_nonviolence.wav
+- spec/data/audio/king_nonviolence.wma
+- spec/data/audio/top_gun_secret.aac
+- spec/data/audio/top_gun_secret.ac3
+- spec/data/audio/top_gun_secret.aiff
+- spec/data/audio/top_gun_secret.au
+- spec/data/audio/top_gun_secret.flac
+- spec/data/audio/top_gun_secret.m4a
+- spec/data/audio/top_gun_secret.mka
+- spec/data/audio/top_gun_secret.mp3
+- spec/data/audio/top_gun_secret.ogg
+- spec/data/audio/top_gun_secret.ra
+- spec/data/audio/top_gun_secret.voc
+- spec/data/audio/top_gun_secret.wav
+- spec/data/audio/top_gun_secret.wma
 - spec/data/multipage.tif
 - spec/data/multipage.tif.jp2
 - spec/data/test-ead.xml
@@ -291,6 +429,23 @@ test_files:
 - spec/data/test.xlsx
 - spec/data/test.xml
 - spec/data/test_pdfa.pdf
+- spec/data/video/SampleVideo_1080x720_2mb.flv
+- spec/data/video/SampleVideo_1080x720_2mb.mkv
+- spec/data/video/SampleVideo_1080x720_2mb.mp4
+- spec/data/video/SampleVideo_176x144_2mb.3gp
+- spec/data/video/SampleVideo_320x240_2mb.3gp
+- spec/data/video/SampleVideo_360x240_2mb.flv
+- spec/data/video/SampleVideo_360x240_2mb.mkv
+- spec/data/video/SampleVideo_360x240_2mb.mp4
+- spec/data/video/sample.avi
+- spec/data/video/sample.flv
+- spec/data/video/sample.mkv
+- spec/data/video/sample.mov
+- spec/data/video/sample.mp4
+- spec/data/video/sample.mpg
+- spec/data/video/sample.swf
+- spec/data/video/sample.webm
+- spec/data/video/sample.wmv
 - spec/identifier_spec.rb
 - spec/spec_helper.rb
 - spec/test_types.yml