libis-tools 1.0.12 → 1.0.16

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 810ac9ed3a3b03f0ae8cc0d1deea5b1c6a0f70b0d34bce8eb752fba9da5583f9
-  data.tar.gz: 5b51eb0c78b0928e1183f53bd270ec148f35b47f42a066432705a657279ccfa9
+  metadata.gz: 8e11e0bac24cf562c5834d59510d7899d61e1295e3128fa66f549f40549d6c3e
+  data.tar.gz: 834820ad3715e8321c34af9099a7ff01278f15641ef99fb677eea3df0a0bcef4
 SHA512:
-  metadata.gz: a0a53c49318401d632db7a3e20a70a1378a50ce101c2e8f26ac0216aee3ffcb28c81287663b6daccb3693d2a962337c472f36254e6ea065aed2135e65cfbf876
-  data.tar.gz: 207d258c97de31f803d71dd0f39c6cc37b00fbcfe0120bf37a36c3669c3ac43d4d1253b2c1cfed0004a24ed8215e03e9990af4853389c40958f6c5300497068f
+  metadata.gz: b5c898f23a098f8c592f9cb0a672e8d5248ffc71b4c93062c759daff7341a0cbdd4e95af9dd1017244afd31627c9c4e8d26df8698d433669f919db338eb83a7f
+  data.tar.gz: 9c90ddab8ff5e79f0c9478eea359baa70117570a5cafad8fc8454050af4b3f7709ba54a1c314b8aa81ea6cbba791cb64e759ccec2e6c5b97360271aa04cbc388

data/lib/libis/tools/extend/empty.rb

@@ -19,10 +19,10 @@ class FalseClass
 end
 
 class String
-  BLANK_RE = /\A[[:space:]]^\z/.freeze
+  BLANK_REGEX = /\A[[:space:]]^\z/.freeze
   def blank?
-    empty? || BLANK_RE.match?(self)
-  end
+    empty? || BLANK_REGEX.match?(self)
+  end unless method_defined? :blank?
 end
 
 class Object

data/lib/libis/tools/extend/roo.rb

@@ -88,4 +88,16 @@ module Roo
     end
 
   end
+
+  class CSV
+    def each_row(options, &block)
+      if uri?(filename)
+        each_row_using_tempdir(options, &block)
+      elsif is_stream?(filename_or_stream)
+        ::CSV.new(filename_or_stream, **options).each(&block)
+      else
+        ::CSV.foreach(filename, **options, &block)
+      end
+    end
+  end
 end

data/lib/libis/tools/mets_file.rb

@@ -91,14 +91,18 @@ module Libis
                   end
 
         dmd_sec = xml_doc.root.xpath('mets:dmdSec', NS).inject({}) do |hash_dmd, dmd|
-          hash_dmd[dmd[:ID]] = dmd.xpath('.//dc:record', NS).first.children.inject({}) do |h, c|
+          id = dmd[:ID]
+          id = id.slice(0..(id.index('-dmd') + 3))
+          hash_dmd[id] = dmd.xpath('.//dc:record', NS).first.children.inject({}) do |h, c|
             h[c.name] = c.content if c.name != 'text'
             h
           end
           hash_dmd
         end
         amd_sec = xml_doc.root.xpath('mets:amdSec', NS).inject({}) do |hash_amd, amd|
-          hash_amd[amd[:ID]] = [:tech, :rights, :source, :digiprov].inject({}) do |hash_sec, sec|
+          id = amd[:ID]
+          id = id.slice(0..(id.index('-amd') + 3))
+          hash_amd[id] = [:tech, :rights, :source, :digiprov].inject({}) do |hash_sec, sec|
             md = amd.xpath("mets:#{sec}MD", NS).first
             return hash_sec unless md
             # hash_sec[sec] = md.xpath('mets:mdWrap/dnx:dnx/dnx:section', NS).inject({}) do |hash_md, dnx_sec|

data/lib/libis/tools/mets_objects.rb

@@ -289,7 +289,7 @@ module Libis
               compositionLevel: composition_level,
               # fileLocationType: 'FILE',
               fileLocation: location,
-              fileOriginalName: CGI.escape(original || target_name).gsub('+', '%20'),
+              fileOriginalName: original || target_name,
               fileOriginalPath: target_path,
               # fileOriginalID: URI.encode(location),
               # fileExtension: ::File.extname(orig_name),

data/lib/libis/tools/version.rb

@@ -1,5 +1,5 @@
 module Libis
   module Tools
-    VERSION = '1.0.12'
+    VERSION = '1.0.16'
   end
 end

data/lib/libis/tools/xml_document.rb

@@ -88,7 +88,7 @@ module Libis
 
       # Export the XML Document to an XML string.
       # @param [Hash] options options passed to the underlying Nokogiri::XML::Document#to_xml; default is:
-      #     !{indent: 2, encoding: 'utf-8'}
+      #     {indent: 2, encoding: 'utf-8'}
       # @return [String] a string
       def to_xml(options = {})
         options = {indent: 2, encoding: 'utf-8', save_with: Nokogiri::XML::Node::SaveOptions::DEFAULT_XML}.merge(options)

data/spec/mets_file_spec.rb

@@ -220,4 +220,426 @@ describe 'METS File' do
     end
   end
 
+  context 'builder' do
+
+    let(:dc_record) {
+      Libis::Tools::XmlDocument.parse <<-STR
+        <?xml version="1.0" encoding="UTF-8"?>
+        <dc:record xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
+                xmlns:dc="http://purl.org/dc/elements/1.1/"
+                xmlns:dcterms="http://purl.org/dc/terms/">
+          <dc:title>âbç d€f</dc:title>
+        </dc:record>
+      STR
+    }
+
+    let(:sample) {
+      {
+        dc_record: dc_record.root.to_xml,
+        amd: {
+          status: "ACTIVE",
+          access_right: "AR_EVERYONE"
+        },
+        representations: [
+          {
+            preservation_type: "PRESERVATION_MASTER",
+            usage_type: "VIEW",
+            label: "Archiefkopie"
+          },
+        ],
+        files: [
+          {
+            creation_date: '2021-08-27 14:14:26.367 UTC',
+            modification_date: '2018-04-06 05:12:27.077 UTC',
+            location: "/nas/upload/demo/test_diacritics/âbç d€f/aaa á aaa/b/abc.doc",
+            target_location: "aaa á aaa/b/abc.doc",
+            mimetype: "application/msword",
+            size: 56320,
+            puid: "fmt/39",
+            checksum_MD5: "16ddbd5ad80411d8ac1b6acf8348adfb",
+            label: "abc.doc",
+            representation: 1,
+          }, {
+            creation_date: '2021-08-27 14:14:26.367 UTC',
+            modification_date: '2018-04-06 05:13:04.635 UTC',
+            location: "/nas/upload/demo/test_diacritics/âbç d€f/aaa á aaa/b/def.dóc",
+            target_location: "aaa á aaa/b/def.dóc",
+            mimetype: "application/msword",
+            size: 45056,
+            puid: "fmt/39",
+            checksum_MD5: "b2907905a58f8599f07334cdc295c2f1",
+            label: "def.dóc",
+            representation: 1,
+          }, {
+            creation_date: '2021-08-27 14:14:26.366 UTC',
+            modification_date: '2018-04-06 05:12:17.935 UTC',
+            location: "/nas/upload/demo/test_diacritics/âbç d€f/aaa á aaa/b/ábc.doc",
+            target_location: "aaa á aaa/b/ábc.doc",
+            mimetype: "application/msword",
+            size: 61440,
+            puid: "fmt/39",
+            checksum_MD5: "312bba1438acdb17e3d69d19fe4cac4b",
+            label: "ábc.doc",
+            representation: 1,
+          }, {
+            creation_date: '2021-08-27 14:14:26.346 UTC',
+            modification_date: '2018-04-06 05:06:40.297 UTC',
+            location: "/nas/upload/demo/test_diacritics/âbç d€f/aaa á aaa/ccçcc/ã b ç d ë ¼.doc",
+            target_location: "aaa á aaa/ccçcc/ã b ç d ë ¼.doc",
+            mimetype: "application/msword",
+            size: 50688,
+            puid: "fmt/39",
+            checksum_MD5: "cd3e8cfa9cf5ada3485e205d706cee09",
+            label: "ã b ç d ë ¼.doc",
+            representation: 1,
+          }, {
+            creation_date: '2021-08-27 14:14:26.356 UTC',
+            modification_date: '2018-04-06 05:11:07.095 UTC',
+            location: "/nas/upload/demo/test_diacritics/âbç d€f/aaa á aaa/ss$ss/ã b ç d ë ¼.doc",
+            target_location: "aaa á aaa/ss$ss/ã b ç d ë ¼.doc",
+            mimetype: "application/msword",
+            size: 61440,
+            puid: "fmt/39",
+            checksum_MD5: "312bba1438acdb17e3d69d19fe4cac4b",
+            label: "ã b ç d ë ¼.doc",
+            representation: 1,
+          },
+        ],
+        divs: [
+          {
+            label: "âbç d€f"
+          }, {
+            parent: 1,
+            label: "aaa á aaa",
+          }, {
+            parent: 2,
+            label: "b",
+            files: [1,2,3]
+          }, {
+            parent: 2,
+            label: "ccçcc",
+            files: [4]
+          }, {
+            parent: 2,
+            label: "ss$ss",
+            files: [5]
+          },
+        ],
+        maps: [
+          {
+            representation: 1,
+            div: 1,
+            is_logical: false
+          }
+        ]
+      }
+    }
+
+    let(:target) {
+      ::Libis::Tools::XmlDocument.build do |xml|
+        xml[:mets].mets 'xmlns:mets' => 'http://www.loc.gov/METS/' do
+          xml[:mets].dmdSec ID: 'ie-dmd' do
+            xml[:mets].mdWrap MDTYPE: 'DC' do
+              xml[:mets].xmlData do
+                xml << dc_record.root.to_xml
+                # xml[:dc].record(
+                #   'xmlns:dc' => 'http://purl.org/dc/elements/1.1/',
+                #   'xmlns:xsi' => 'http://www.w3.org/2001/XMLSchema-instance',
+                #   'xmlns:dcterms' => 'http://purl.org/dc/terms/'
+                # ) do
+                #   xml[:dc].title 'âbç d€f'
+                # end
+              end
+            end
+          end
+          xml[:mets].amdSec ID: 'ie-amd' do
+            xml[:mets].techMD ID: 'ie-amd-tech' do
+              xml[:mets].mdWrap MDTYPE: 'OTHER', OTHERMDTYPE: 'dnx' do
+                xml[:mets].xmlData do
+                  xml.dnx xmlns: 'http://www.exlibrisgroup.com/dps/dnx' do
+                    xml.section.generalIECharacteristics! do
+                      xml.record do
+                        xml.key.status! 'ACTIVE'
+                      end
+                    end
+                  end
+                end
+              end
+            end
+            xml[:mets].rightsMD ID: 'ie-amd-rights' do
+              xml[:mets].mdWrap MDTYPE: 'OTHER', OTHERMDTYPE: 'dnx' do
+                xml[:mets].xmlData do
+                  xml.dnx xmlns: 'http://www.exlibrisgroup.com/dps/dnx' do
+                    xml.section.accessRightsPolicy! do
+                      xml.record do
+                        xml.key.policyId! 'AR_EVERYONE'
+                      end
+                    end
+                  end
+                end
+              end
+            end
+          end
+          xml[:mets].amdSec ID: 'rep1-amd' do
+            xml[:mets].techMD ID: 'rep1-amd-tech' do
+              xml[:mets].mdWrap MDTYPE: 'OTHER', OTHERMDTYPE: 'dnx' do
+                xml[:mets].xmlData do
+                  xml.dnx xmlns: 'http://www.exlibrisgroup.com/dps/dnx' do
+                    xml.section.generalRepCharacteristics! do
+                      xml.record do
+                        xml.key.preservationType! 'PRESERVATION_MASTER'
+                        xml.key.usageType! 'VIEW'
+                        xml.key.label! 'Archiefkopie'
+                      end
+                    end
+                  end
+                end
+              end
+            end
+          end
+          xml[:mets].amdSec ID: 'fid1-amd' do
+            xml[:mets].techMD ID: 'fid1-amd-tech' do
+              xml[:mets].mdWrap MDTYPE: 'OTHER', OTHERMDTYPE: 'dnx' do
+                xml[:mets].xmlData do
+                  xml.dnx xmlns: 'http://www.exlibrisgroup.com/dps/dnx' do
+                    xml.section.generalFileCharacteristics! do
+                      xml.record do
+                        xml.key.label! 'abc.doc'
+                        xml.key.fileCreationDate! '2021-08-27 14:14:26.367 UTC'
+                        xml.key.fileModificationDate! '2018-04-06 05:12:27.077 UTC'
+                        xml.key.fileLocation! '/nas/upload/demo/test_diacritics/âbç d€f/aaa á aaa/b/abc.doc'
+                        xml.key.fileOriginalName! 'abc.doc'
+                        xml.key.fileOriginalPath! 'aaa á aaa/b'
+                        xml.key.fileMIMEType! 'application/msword'
+                        xml.key.fileSizeBytes! '56320'
+                      end
+                    end
+                    xml.section.fileFixity! do
+                      xml.record do
+                        xml.key.fixityType! 'MD5'
+                        xml.key.fixityValue! '16ddbd5ad80411d8ac1b6acf8348adfb'
+                      end
+                    end
+                    xml.section.objectCharacteristics! do
+                      xml.record do
+                        xml.key.groupID! 'grp'
+                      end
+                    end
+                  end
+                end
+              end
+            end
+          end
+          xml[:mets].amdSec ID: 'fid2-amd' do
+            xml[:mets].techMD ID: 'fid2-amd-tech' do
+              xml[:mets].mdWrap MDTYPE: 'OTHER', OTHERMDTYPE: 'dnx' do
+                xml[:mets].xmlData do
+                  xml.dnx xmlns: 'http://www.exlibrisgroup.com/dps/dnx' do
+                    xml.section.generalFileCharacteristics! do
+                      xml.record do
+                        xml.key.label! 'def.dóc'
+                        xml.key.fileCreationDate! '2021-08-27 14:14:26.367 UTC'
+                        xml.key.fileModificationDate! '2018-04-06 05:13:04.635 UTC'
+                        xml.key.fileLocation! '/nas/upload/demo/test_diacritics/âbç d€f/aaa á aaa/b/def.dóc'
+                        xml.key.fileOriginalName! 'def.dóc'
+                        xml.key.fileOriginalPath! 'aaa á aaa/b'
+                        xml.key.fileMIMEType! 'application/msword'
+                        xml.key.fileSizeBytes! '45056'
+                      end
+                    end
+                    xml.section.fileFixity! do
+                      xml.record do
+                        xml.key.fixityType! 'MD5'
+                        xml.key.fixityValue! 'b2907905a58f8599f07334cdc295c2f1'
+                      end
+                    end
+                    xml.section.objectCharacteristics! do
+                      xml.record do
+                        xml.key.groupID! 'grp'
+                      end
+                    end
+                  end
+                end
+              end
+            end
+          end
+          xml[:mets].amdSec ID: 'fid3-amd' do
+            xml[:mets].techMD ID: 'fid3-amd-tech' do
+              xml[:mets].mdWrap MDTYPE: 'OTHER', OTHERMDTYPE: 'dnx' do
+                xml[:mets].xmlData do
+                  xml.dnx xmlns: 'http://www.exlibrisgroup.com/dps/dnx' do
+                    xml.section.generalFileCharacteristics! do
+                      xml.record do
+                        xml.key.label! 'ábc.doc'
+                        xml.key.fileCreationDate! '2021-08-27 14:14:26.366 UTC'
+                        xml.key.fileModificationDate! '2018-04-06 05:12:17.935 UTC'
+                        xml.key.fileLocation! '/nas/upload/demo/test_diacritics/âbç d€f/aaa á aaa/b/ábc.doc'
+                        xml.key.fileOriginalName! 'ábc.doc'
+                        xml.key.fileOriginalPath! 'aaa á aaa/b'
+                        xml.key.fileMIMEType! 'application/msword'
+                        xml.key.fileSizeBytes! '61440'
+                      end
+                    end
+                    xml.section.fileFixity! do
+                      xml.record do
+                        xml.key.fixityType! 'MD5'
+                        xml.key.fixityValue! '312bba1438acdb17e3d69d19fe4cac4b'
+                      end
+                    end
+                    xml.section.objectCharacteristics! do
+                      xml.record do
+                        xml.key.groupID! 'grp'
+                      end
+                    end
+                  end
+                end
+              end
+            end
+          end
+          xml[:mets].amdSec ID: 'fid4-amd' do
+            xml[:mets].techMD ID: 'fid4-amd-tech' do
+              xml[:mets].mdWrap MDTYPE: 'OTHER', OTHERMDTYPE: 'dnx' do
+                xml[:mets].xmlData do
+                  xml.dnx xmlns: 'http://www.exlibrisgroup.com/dps/dnx' do
+                    xml.section.generalFileCharacteristics! do
+                      xml.record do
+                        xml.key.label! 'ã b ç d ë ¼.doc'
+                        xml.key.fileCreationDate! '2021-08-27 14:14:26.346 UTC'
+                        xml.key.fileModificationDate! '2018-04-06 05:06:40.297 UTC'
+                        xml.key.fileLocation! '/nas/upload/demo/test_diacritics/âbç d€f/aaa á aaa/ccçcc/ã b ç d ë ¼.doc'
+                        xml.key.fileOriginalName! 'ã b ç d ë ¼.doc'
+                        xml.key.fileOriginalPath! 'aaa á aaa/ccçcc'
+                        xml.key.fileMIMEType! 'application/msword'
+                        xml.key.fileSizeBytes! '50688'
+                      end
+                    end
+                     xml.section.fileFixity! do
+                      xml.record do
+                        xml.key.fixityType! 'MD5'
+                        xml.key.fixityValue! 'cd3e8cfa9cf5ada3485e205d706cee09'
+                      end
+                    end
+                    xml.section.objectCharacteristics! do
+                      xml.record do
+                        xml.key.groupID! 'grp'
+                      end
+                    end
+                  end
+                end
+              end
+            end
+          end
+          xml[:mets].amdSec ID: 'fid5-amd' do
+            xml[:mets].techMD ID: 'fid5-amd-tech' do
+              xml[:mets].mdWrap MDTYPE: 'OTHER', OTHERMDTYPE: 'dnx' do
+                xml[:mets].xmlData do
+                  xml.dnx xmlns: 'http://www.exlibrisgroup.com/dps/dnx' do
+                    xml.section.generalFileCharacteristics! do
+                      xml.record do
+                        xml.key.label! 'ã b ç d ë ¼.doc'
+                        xml.key.fileCreationDate! '2021-08-27 14:14:26.356 UTC'
+                        xml.key.fileModificationDate! '2018-04-06 05:11:07.095 UTC'
+                        xml.key.fileLocation! '/nas/upload/demo/test_diacritics/âbç d€f/aaa á aaa/ss$ss/ã b ç d ë ¼.doc'
+                        xml.key.fileOriginalName! 'ã b ç d ë ¼.doc'
+                        xml.key.fileOriginalPath! 'aaa á aaa/ss$ss'
+                        xml.key.fileMIMEType! 'application/msword'
+                        xml.key.fileSizeBytes! '61440'
+                      end
+                    end
+                     xml.section.fileFixity! do
+                      xml.record do
+                        xml.key.fixityType! 'MD5'
+                        xml.key.fixityValue! '312bba1438acdb17e3d69d19fe4cac4b'
+                      end
+                    end
+                    xml.section.objectCharacteristics! do
+                      xml.record do
+                        xml.key.groupID! 'grp'
+                      end
+                    end
+                  end
+                end
+              end
+            end
+          end
+          xml[:mets].fileSec do
+            xml[:mets].fileGrp ID: 'rep1', USE: 'VIEW', ADMID: 'rep1-amd' do
+              xml[:mets].file ID: 'fid1', MIMETYPE: 'application/msword', ADMID: 'fid1-amd', GROUPID: 'grp' do
+                xml[:mets].FLocat 'xmlns:xlin' => 'http://www.w3.org/1999/xlink', LOCTYPE: 'URL', 'xlin:href' => 'aaa á aaa/b/abc.doc'
+              end
+              xml[:mets].file ID: 'fid2', MIMETYPE: 'application/msword', ADMID: 'fid2-amd', GROUPID: 'grp' do
+                xml[:mets].FLocat 'xmlns:xlin' => 'http://www.w3.org/1999/xlink', LOCTYPE: 'URL', 'xlin:href' => 'aaa á aaa/b/def.dóc'
+              end
+              xml[:mets].file ID: 'fid3', MIMETYPE: 'application/msword', ADMID: 'fid3-amd', GROUPID: 'grp' do
+                xml[:mets].FLocat 'xmlns:xlin' => 'http://www.w3.org/1999/xlink', LOCTYPE: 'URL', 'xlin:href' => 'aaa á aaa/b/ábc.doc'
+              end
+              xml[:mets].file ID: 'fid4', MIMETYPE: 'application/msword', ADMID: 'fid4-amd', GROUPID: 'grp' do
+                xml[:mets].FLocat 'xmlns:xlin' => 'http://www.w3.org/1999/xlink', LOCTYPE: 'URL', 'xlin:href' => 'aaa á aaa/ccçcc/ã b ç d ë ¼.doc'
+              end
+              xml[:mets].file ID: 'fid5', MIMETYPE: 'application/msword', ADMID: 'fid5-amd', GROUPID: 'grp' do
+                xml[:mets].FLocat 'xmlns:xlin' => 'http://www.w3.org/1999/xlink', LOCTYPE: 'URL', 'xlin:href' => 'aaa á aaa/ss$ss/ã b ç d ë ¼.doc'
+              end
+            end
+          end
+          xml[:mets].structMap ID: 'rep1-1', TYPE: 'PHYSICAL' do
+            xml[:mets].div LABEL: 'Archiefkopie' do
+              xml[:mets].div LABEL: 'âbç d€f' do
+                xml[:mets].div LABEL: 'aaa á aaa' do
+                  xml[:mets].div LABEL: 'b' do
+                    xml[:mets].div LABEL: 'abc.doc', TYPE: 'FILE' do
+                      xml[:mets].fptr FILEID: 'fid1'
+                    end
+                    xml[:mets].div LABEL: 'def.dóc', TYPE: 'FILE' do
+                      xml[:mets].fptr FILEID: 'fid2'
+                    end
+                    xml[:mets].div LABEL: 'ábc.doc', TYPE: 'FILE' do
+                      xml[:mets].fptr FILEID: 'fid3'
+                    end
+                  end
+                  xml[:mets].div LABEL: 'ccçcc' do
+                    xml[:mets].div LABEL: 'ã b ç d ë ¼.doc', TYPE: 'FILE' do
+                      xml[:mets].fptr FILEID: 'fid4'
+                    end
+                  end
+                  xml[:mets].div LABEL: 'ss$ss' do
+                    xml[:mets].div LABEL: 'ã b ç d ë ¼.doc', TYPE: 'FILE' do
+                      xml[:mets].fptr FILEID: 'fid5'
+                    end
+                  end
+                end
+              end
+            end
+          end
+        end
+      end.document
+    }
+
+    it 'produces METS XML' do
+      mets_file.dc_record = sample[:dc_record]
+      mets_file.amd_info = sample[:amd]
+      sample[:representations].each do |rep|
+        mets_file.representation(rep)
+      end
+      sample[:files].each do |file|
+        rep_id = file.delete(:representation)
+        rep = mets_file.representations[rep_id]
+        mets_file.file(file).representation = rep
+      end
+      sample[:divs].each do |div|
+        parent = div.delete(:parent)
+        files = div.delete(:files)
+        d = mets_file.div(div)
+        files.each { |f| d << mets_file.files[f] } if files
+        mets_file.divs[parent] << d if parent
+      end
+      sample[:maps].each do |map|
+        rep = mets_file.representations[map[:representation]]
+        div = mets_file.divs[map[:div]]
+        mets_file.map(rep, div)
+      end
+
+      expect(mets_file.xml_doc.root).to be_equivalent_to target.root
+    end
+  end
+
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: libis-tools
 version: !ruby/object:Gem::Version
-  version: 1.0.12
+  version: 1.0.16
 platform: ruby
 authors:
 - Kris Dekeyser
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2021-08-25 00:00:00.000000000 Z
+date: 2021-08-31 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler