libis-format 0.9.32 → 0.9.33

data/spec/identifier_spec.rb

@@ -1,37 +1,12 @@
 # encoding: utf-8
 require 'spec_helper'
+require 'libis/format/identifier'
+require 'awesome_print'
 
-describe 'Identfier' do
-
-  dir = File.join File.absolute_path(File.dirname(__FILE__)), 'data'
-
-  before :all do
-    ::Libis::Tools::Config.logger.appenders =
-        ::Logging::Appenders.string_io('StringIO', layout: ::Libis::Tools::Config.get_log_formatter)
-    ::Libis::Tools::Config.logger.level = :all
-  end
-
-  let(:logoutput) { ::Libis::Tools::Config.logger.appenders.last.sio }
-
-  it 'should initialize correctly' do
-    # expect(Libis::Format::Identifier.fido_formats.size).to be 1
-    # expect(File.basename(Libis::Format::Identifier.fido_formats.first)).to eq 'lias_formats.xml'
-
-    expect(Libis::Format::Identifier.xml_validations.size).to be 1
-    expect(File.basename Libis::Format::Identifier.xml_validations['archive/ead']).to eq 'ead.xsd'
-  end
-
-  it 'should not attempt to identify a directory' do
-    expect(Libis::Format::Identifier.get dir).to be_nil
-  end
-
-  it 'should not attempt to identify a file that does not exist' do
-    expect(Libis::Format::Identifier.get File.join(dir, 'does_not_exist_file')).to be_nil
-  end
-
-    # noinspection RubyStringKeysInHashInspection
+# noinspection RubyStringKeysInHashInspection
+formatlist =
     {
-        'Cevennes2.ppm' => {mimetype: 'image/x‑portable‑pixmap', puid: 'fmt/408'},
+        'Cevennes2.ppm' => {mimetype: 'image/x-portable-pixmap', puid: 'fmt/408'},
         'Cevennes2.bmp' => {mimetype: 'image/bmp', puid: 'fmt/116'},
         'Cevennes2.jp2' => {mimetype: 'image/jp2', puid: 'x-fmt/392'},
         'test.odt' => {mimetype: 'application/vnd.oasis.opendocument.text', puid: 'fmt/291'},
@@ -52,14 +27,130 @@ describe 'Identfier' do
         'test.png' => {mimetype: 'image/png', puid: 'fmt/11'},
         'test.jpg' => {mimetype: 'image/jpeg', puid: 'fmt/43'},
         'test.gif' => {mimetype: 'image/gif', puid: 'fmt/4'},
-        'test.xml' => {mimetype: 'application/xml', puid: 'fmt/101'},
+        'test.xml' => {mimetype: 'text/xml', puid: 'fmt/101'},
         'test-ead.xml' => {mimetype: 'archive/ead', puid: 'fmt/101'},
         'NikonRaw-CaptureOne.tif' => {mimetype: 'image/tiff', puid: 'x-fmt/387'},
-        'NikonRaw-CameraRaw.TIF' => {mimetype: 'image/tiff', puid: 'fmt/202'},
-    }.each do |file, result|
+        'NikonRaw-CameraRaw.TIF' => {mimetype: 'image/tiff', puid: 'fmt/353'},
+    }
+
+# noinspection RubyStringKeysInHashInspection
+fidolist =
+    {
+        'Cevennes2.ppm' => {mimetype: 'image/x-portable-pixmap', puid: 'fmt/408'},
+        'Cevennes2.bmp' => {mimetype: 'image/bmp', puid: 'fmt/116'},
+        'Cevennes2.jp2' => {mimetype: 'image/jp2', puid: 'x-fmt/392'},
+        'test.odt' => {mimetype: 'application/vnd.oasis.opendocument.text', puid: 'fmt/290'},
+        'test.doc' => {mimetype: nil, puid: 'fmt/111'},
+        'test.docx' => {mimetype: 'application/vnd.openxmlformats-officedocument.wordprocessingml.document', puid: 'fmt/412'},
+        'test.pdf' => {mimetype: 'application/pdf', puid: 'fmt/18'},
+        'test.rtf' => {mimetype: 'application/rtf', puid: 'fmt/45'},
+        'test.txt' => {mimetype: 'text/plain', puid: 'x-fmt/111'},
+        'test.ods' => {mimetype: nil, puid: 'fmt/294'},
+        'test.xls' => {mimetype: 'application/vnd.ms-excel', puid: 'fmt/61'},
+        'test.xlsx' => {mimetype: 'application/vnd.openxmlformats-officedocument.spreadsheetml.sheet', puid: 'fmt/214'},
+        'test.psd' => {mimetype: 'image/vnd.adobe.photoshop', puid: 'x-fmt/92'},
+        'test.bmp' => {mimetype: 'image/bmp', puid: 'fmt/119'},
+        'test.tif' => {mimetype: 'image/tiff', puid: 'fmt/353'},
+        'test-jpg.tif' => {mimetype: 'image/tiff', puid: 'fmt/353'},
+        'test-lzw.tif' => {mimetype: 'image/tiff', puid: 'fmt/353'},
+        'test.ps' => {mimetype: 'application/postscript', puid: 'x-fmt/408'},
+        'test.png' => {mimetype: 'image/png', puid: 'fmt/11'},
+        'test.jpg' => {mimetype: 'image/jpeg', puid: 'fmt/43'},
+        'test.gif' => {mimetype: 'image/gif', puid: 'fmt/4'},
+        'test.xml' => {mimetype: 'text/xml', puid: 'fmt/101'},
+        'test-ead.xml' => {mimetype: 'text/xml', puid: 'fmt/101'},
+        'NikonRaw-CaptureOne.tif' => {mimetype: 'image/tiff', puid: 'fmt/353'},
+        'NikonRaw-CameraRaw.TIF' => {mimetype: 'image/tiff', puid: 'fmt/353'},
+    }
+
+describe 'Identfier' do
+
+  dir = File.join File.absolute_path(File.dirname(__FILE__)), 'data'
+
+  let (:identifier) { ::Libis::Format::Identifier }
+
+  before :all 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] = '/usr/local/bin/fido'
+  end
+
+  let(:logoutput) {::Libis::Tools::Config.logger.appenders.last.sio}
+
+  it 'should initialize correctly' do
+    expect(identifier.xml_validations.size).to be 1
+    expect(File.basename identifier.xml_validations['archive/ead']).to eq 'ead.xsd'
+  end
+
+  # noinspection RubyResolve
+  it 'should not attempt to identify a non-existing file or directory' do
+    result = identifier.get (File.join(dir, 'abcdef'))
+    expect(result[:output]).to be_empty
+    expect(result[:formats]).to be_empty
+    expect(result[:messages]).not_to be_empty
+    expect(result[:messages].all? do |x|
+      x[0] == :error &&
+          x[1] =~ /^Error running (Droid|Fido|File): IdentificationTool: file argument should be a path to an existing file or directory or a list of those/
+    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
+
+  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
+
+  context 'individual files' do
+
+    formatlist.each do |file, format|
       it "should identify test document '#{file}'" do
-        expect(::Libis::Format::Identifier.get(File.join(dir,file))).to include result
+        result = identifier.get (File.join(dir, file))
+        expect(result[:formats].size).to be 1
+        expect(result[:formats][File.join(dir, file)]).to include format
+      end
+    end
+
+  end
+
+  context 'Fido' do
+
+    # expect(identifier.fido_formats.size).to be 1
+    # expect(File.basename(identifier.fido_formats.first)).to eq 'lias_formats.xml'
+    it 'should identify list of test documents' do
+      filelist = formatlist.keys.map {|file| File.join(dir, file)}
+      fido_result = ::Libis::Format::Fido.instance.run_list(filelist)
+      filelist.each do |filename|
+        result = fido_result[filename]
+        result = result[0] if result
+        # noinspection RubyResolve
+        expect(result).to include fidolist[File.basename(filename)]
       end
     end
 
+    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)
+      filelist.each do |filename|
+        result = fido_result[filename]
+        result = result[0] if result
+        # noinspection RubyResolve
+        expect(result).to include fidolist[File.basename(filename)]
+      end
+    end
+
+  end
+
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: libis-format
 version: !ruby/object:Gem::Version
-  version: 0.9.32
+  version: 0.9.33
 platform: ruby
 authors:
 - Kris Dekeyser
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2017-07-31 00:00:00.000000000 Z
+date: 2017-09-17 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -66,20 +66,34 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '0.9'
+- !ruby/object:Gem::Dependency
+  name: awesome_print
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
 - !ruby/object:Gem::Dependency
   name: libis-tools
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.9'
+        version: 0.9.52
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.9'
+        version: 0.9.52
 - !ruby/object:Gem::Dependency
   name: os
   requirement: !ruby/object:Gem::Requirement
@@ -162,7 +176,10 @@ files:
 - lib/libis/format/converter/pdf_converter.rb
 - lib/libis/format/converter/repository.rb
 - lib/libis/format/droid.rb
+- lib/libis/format/extension_identification.rb
 - lib/libis/format/fido.rb
+- lib/libis/format/file_tool.rb
+- lib/libis/format/identification_tool.rb
 - lib/libis/format/identifier.rb
 - lib/libis/format/office_to_pdf.rb
 - lib/libis/format/pdf_copy.rb
@@ -214,128 +231,6 @@ files:
 - tools/PdfTool.jar
 - tools/bcpkix-jdk15on-1.49.jar
 - tools/bcprov-jdk15on-1.49.jar
-- tools/droid/DROID_SignatureFile_V90.xml
-- tools/droid/container-signature-20170330.xml
-- tools/droid/droid-command-line-6.3.jar
-- tools/droid/droid.bat
-- tools/droid/droid.sh
-- tools/droid/lib/XmlSchema-1.4.7.jar
-- tools/droid/lib/activation-1.1.jar
-- tools/droid/lib/aopalliance-1.0.jar
-- tools/droid/lib/asm-2.2.3.jar
-- tools/droid/lib/aspectjrt-1.8.7.jar
-- tools/droid/lib/aspectjweaver-1.8.7.jar
-- tools/droid/lib/bcmail-jdk14-138.jar
-- tools/droid/lib/bcprov-jdk14-138.jar
-- tools/droid/lib/beansbinding-1.2.1.jar
-- tools/droid/lib/byteseek-2.0.3.jar
-- tools/droid/lib/cglib-nodep-2.2.2.jar
-- tools/droid/lib/classmate-1.0.0.jar
-- tools/droid/lib/commons-cli-1.2.jar
-- tools/droid/lib/commons-codec-1.10.jar
-- tools/droid/lib/commons-collections-3.2.2.jar
-- tools/droid/lib/commons-compress-1.4.1.jar
-- tools/droid/lib/commons-configuration-1.8.jar
-- tools/droid/lib/commons-dbcp-1.4.jar
-- tools/droid/lib/commons-httpclient-3.1.jar
-- tools/droid/lib/commons-io-2.4.jar
-- tools/droid/lib/commons-lang-2.6.jar
-- tools/droid/lib/commons-logging-1.1.1.jar
-- tools/droid/lib/commons-pool-1.5.4.jar
-- tools/droid/lib/cxf-api-2.2.12.jar
-- tools/droid/lib/cxf-common-schemas-2.2.12.jar
-- tools/droid/lib/cxf-common-utilities-2.2.12.jar
-- tools/droid/lib/cxf-rt-bindings-http-2.2.12.jar
-- tools/droid/lib/cxf-rt-bindings-soap-2.2.12.jar
-- tools/droid/lib/cxf-rt-bindings-xml-2.2.12.jar
-- tools/droid/lib/cxf-rt-core-2.2.12.jar
-- tools/droid/lib/cxf-rt-databinding-jaxb-2.2.12.jar
-- tools/droid/lib/cxf-rt-frontend-jaxws-2.2.12.jar
-- tools/droid/lib/cxf-rt-frontend-simple-2.2.12.jar
-- tools/droid/lib/cxf-rt-transports-http-2.2.12.jar
-- tools/droid/lib/cxf-rt-ws-addr-2.2.12.jar
-- tools/droid/lib/cxf-tools-common-2.2.12.jar
-- tools/droid/lib/de.huxhorn.lilith.3rdparty.flyingsaucer.core-renderer-8RC1.jar
-- tools/droid/lib/derby-10.10.2.0.jar
-- tools/droid/lib/droid-container-6.3.jar
-- tools/droid/lib/droid-core-6.3.jar
-- tools/droid/lib/droid-core-interfaces-6.3.jar
-- tools/droid/lib/droid-export-6.3.jar
-- tools/droid/lib/droid-export-interfaces-6.3.jar
-- tools/droid/lib/droid-help-6.3.jar
-- tools/droid/lib/droid-report-6.3.jar
-- tools/droid/lib/droid-report-interfaces-6.3.jar
-- tools/droid/lib/droid-results-6.3.jar
-- tools/droid/lib/geronimo-activation_1.1_spec-1.0.2.jar
-- tools/droid/lib/geronimo-annotation_1.0_spec-1.1.1.jar
-- tools/droid/lib/geronimo-javamail_1.4_spec-1.6.jar
-- tools/droid/lib/geronimo-jaxws_2.1_spec-1.0.jar
-- tools/droid/lib/geronimo-stax-api_1.0_spec-1.0.1.jar
-- tools/droid/lib/geronimo-ws-metadata_2.0_spec-1.1.2.jar
-- tools/droid/lib/hibernate-validator-5.1.0.Final.jar
-- tools/droid/lib/itext-2.0.8.jar
-- tools/droid/lib/javahelp-2.0.05.jar
-- tools/droid/lib/jaxb-api-2.1.jar
-- tools/droid/lib/jaxb-impl-2.1.13.jar
-- tools/droid/lib/jboss-logging-3.1.3.GA.jar
-- tools/droid/lib/joda-time-1.6.2.jar
-- tools/droid/lib/jra-1.0-alpha-4.jar
-- tools/droid/lib/jta-1.1.jar
-- tools/droid/lib/jwat-arc-1.0.3.jar
-- tools/droid/lib/jwat-archive-common-1.0.3.jar
-- tools/droid/lib/jwat-common-1.0.3.jar
-- tools/droid/lib/jwat-gzip-1.0.3.jar
-- tools/droid/lib/jwat-warc-1.0.2.jar
-- tools/droid/lib/log4j-1.2.13.jar
-- tools/droid/lib/neethi-2.0.4.jar
-- tools/droid/lib/opencsv-2.3.jar
-- tools/droid/lib/org-netbeans-swing-outline-7.2.jar
-- tools/droid/lib/org-openide-util-7.2.jar
-- tools/droid/lib/org-openide-util-lookup-7.2.jar
-- tools/droid/lib/poi-3.13.jar
-- tools/droid/lib/saaj-api-1.3.jar
-- tools/droid/lib/saaj-impl-1.3.2.jar
-- tools/droid/lib/slf4j-api-1.4.2.jar
-- tools/droid/lib/slf4j-log4j12-1.4.2.jar
-- tools/droid/lib/spring-aop-4.0.3.RELEASE.jar
-- tools/droid/lib/spring-beans-4.0.3.RELEASE.jar
-- tools/droid/lib/spring-context-4.0.3.RELEASE.jar
-- tools/droid/lib/spring-core-4.0.3.RELEASE.jar
-- tools/droid/lib/spring-expression-4.0.3.RELEASE.jar
-- tools/droid/lib/spring-jdbc-4.0.3.RELEASE.jar
-- tools/droid/lib/spring-orm-4.0.3.RELEASE.jar
-- tools/droid/lib/spring-tx-4.0.3.RELEASE.jar
-- tools/droid/lib/spring-web-2.5.6.jar
-- tools/droid/lib/stax-api-1.0-2.jar
-- tools/droid/lib/trove4j-3.0.3.jar
-- tools/droid/lib/truezip-6.8.4.jar
-- tools/droid/lib/validation-api-1.1.0.Final.jar
-- tools/droid/lib/wsdl4j-1.6.2.jar
-- tools/droid/lib/wstx-asl-3.2.9.jar
-- tools/droid/lib/xercesImpl-2.9.1.jar
-- tools/droid/lib/xml-apis-1.3.04.jar
-- tools/droid/lib/xml-resolver-1.2.jar
-- tools/droid/lib/xz-1.0.jar
-- tools/fido/__init__.py
-- tools/fido/conf/DROID_SignatureFile-v90.xml
-- tools/fido/conf/container-signature-20170330.xml
-- tools/fido/conf/dc.xsd
-- tools/fido/conf/dcmitype.xsd
-- tools/fido/conf/dcterms.xsd
-- tools/fido/conf/fido-formats.xsd
-- tools/fido/conf/format_extension_template.xml
-- tools/fido/conf/format_extensions.xml
-- tools/fido/conf/formats-v90.xml
-- tools/fido/conf/pronom-xml-v90.zip
-- tools/fido/conf/versions.xml
-- tools/fido/fido.bat
-- tools/fido/fido.py
-- tools/fido/fido.sh
-- tools/fido/package.py
-- tools/fido/prepare.py
-- tools/fido/pronomutils.py
-- tools/fido/toxml.py
-- tools/fido/update_signatures.py
 - tools/pdfbox/pdfbox-app-1.8.10.jar
 - tools/pdfbox/preflight-app-1.8.10.jar
 homepage: ''